Orbit Post Sitemap
US Stock Optical Module Surge: It's Not More Orders, but the "Market Access Certificate" Suddenly Became Valuable
Recently, the US stock optical module sector has collectively surged, with Applied Optoelectronics rising nearly 17%, Coherent and Lumentum up about 10%, and FabriNet increasing nearly 5%. Meanwhile, the A-share optical module sector has faced increased pressure due to a rumor. This divergence is not because of a sudden global increase in optical module demand, but because the market is re-pricing the value of the "US market access certificate."
1. Competition in AI Data Centers Has Entered a New Stage
Competition in AI data centers has evolved from "who has more GPUs" to "who can control data flow between GPUs." Optical modules, as the "highways" in computing clusters, with 800G, 1.6T, and future 3.2T modules, determine whether tens of thousands of chips can truly form a supercomputer.
Domestic manufacturers in the high-speed pluggable optical module field not only have price advantages but have also achieved comprehensive leadership in scale, yield, and delivery capabilities. What the US truly fears is not buying a few more domestic modules today, but that as AI data centers proliferate, domestic suppliers become indispensable infrastructure.
2. Industry Rules May Change
The core of this issue is not just security but also potential changes in industry rules. In the past, competition in the optical module industry mainly focused on technology, cost, and delivery capabilities. In the future, it may also depend on who can obtain the "market access certificate" for the US market.
This explains why several US-listed companies suddenly surged. The market is not trading on a sudden increase in global optical module demand but on the possibility that if domestic suppliers are restricted, order bargaining power and valuations may shift to non-domestic suppliers.
3. Capacity Replacement Is Not Easy
One easily overlooked point is that stock prices can be replaced in ten minutes, but production capacity cannot. Domestic manufacturers have long held a large-scale delivery advantage in high-speed pluggable modules. Although US manufacturers have capabilities in lasers, optical chips, coherent optics, and modules, they may not be able to fill all gaps in a short time.
The final result may not be US companies easily taking all orders but rather module price increases, longer certification cycles, and higher cloud provider construction costs. The benefits to several US-listed companies vary:
- Applied Optoelectronics: Closest to direct replacement logic, as it is expanding 800G capacity in the US and Taiwan and has already secured 1.6T bulk orders.
- Coherent and Lumentum: More like upstream technology platforms, benefiting from demand for laser devices, modules, and optical switching.
- FabriNet: May benefit from US manufacturers shifting more production to Southeast Asia.
4. Risks and Opportunities for A-Share Optical Modules
Looking at A-shares, Zhongji Xuchuang and New Easysun are the most sensitive, not because orders will disappear tomorrow, but because both companies have very high overseas revenue proportions—Zhongji Xuchuang about 91%, New Easysun about 96%. Note that overseas revenue does not equal US revenue but indicates their sensitivity to overseas cloud customers and trade rules.
If future restrictions only target new models, the real evaluation may not be on products already shipped in large volumes but on the growing 1.6T and upcoming 3.2T. In other words, short-term revenue cliffs may not occur, but long-term market reassessment of US business growth potential is possible.
Upstream component companies like Tianfu Communication and Guangku Technology are more complex: if restrictions only target complete modules, they may still supply US manufacturers or Southeast Asian manufacturing systems; if the scope expands to corporate identity, key components, and production sources, upstream will be more directly affected.
5. Conclusion: The Market Is Re-Pricing Risk
Therefore, what A-share optical modules truly face is not the disappearance of industry demand but that after the recent surge in US optical modules, the market suddenly needs to re-price the "US market access risk." This may bring high volatility and clear differentiation, but it is not yet appropriate to conclude that the Chinese optical module industry has lost the US market based on a plan not yet implemented.
The three key questions to watch next are:
1. Will the final determination be based on corporate identity or production location?
2. Can existing models continue to be sold?
3. Can Southeast Asian capacity obtain exemptions?
Before these questions are answered, US stocks trade on expectations, and A-shares trade on risk, not conclusions.AMD Earnings Preview (Earnings After Market Close on 8/4)
$AMD
From the options flow perspective, the 600 Call options on 8/7 and 8/5 are locking in the upper limit of the post-earnings rise, which is $600, similar to a high volatility play.
The Put Wall support is formed around $500, so the stop loss can be set near $495 to play for an overall post-earnings upward trend.
From the GEX perspective, the resistance above is very structured at 520, 530, 540, 550, 560, and 570, all showing relatively sizable GEX resistance. Therefore, AMD's rebound will not be very smooth and may be a multi-platform rebound pattern.
AMD's overall GEX is also positive, indicating the market is generally bullish on AMD. An earnings beat would also be an important milestone for the semiconductor sector.
The related $INTC will similarly benefit.
——————————
Options GEX only reflects the current options market's judgment on AMD's earnings; this is not 100% accurate and is for reference only. DYOR
#财报观察员：AMD与SpaceX交卷在即，Circle压轴 Market sentiment for SPCX has undergone a subtle shift. According to TradingBeats data, the bold long positions this morning are still holding their ground. Seven accounts worth millions of dollars hold nearly 240,000 long orders. While the book value gains aren't exaggerated, the attitude is clear—no one is retreating. The largest "0xb37" whale has an average price of $114.96 and a floating profit of only about $150,000, yet it has stubbornly held on, seemingly gambling on a $105 liquidation threshold as solid as a fortress. The morning session did give some respect, with the SPCX climbing from $114.68 to $116 and even touching a high of $117. But in the afternoon, the mood suddenly changed. Funds poured in like sharks smelling fish, but both sides split the new $75.67 million in "ammunition" evenly. Long positions increased by 38.51 million, and short positions by 37.15 million, with the overall battlefield suddenly expanding to $229 million. This wasn't a one-sided charge, but more like a probing and formation before a major battle. Subtle signals have appeared in the past hour. At this critical moment when prices rebound and earnings reports approach, 6 bulls chose to exit, while 34 shorts quietly opened new positions. Although the daily incremental volume was flat, marginal funds at the close clearly bet on the downside, or at least bought a "downside insurance." $SPCX It's like Texas Hold'em—before the flop, everyone pushes the chips, but the river hasn't been dealt yet. In the morning, bulls were betting that "the worst is over," while new shorts in the afternoon said, "The rebound is for short selling."SanDisk Extreme Reversal: Short Covering and AI Storage Narrative Reconstruction
The recent extreme stock price volatility of SanDisk essentially results from the combined effects of short covering, macro environment recovery, and AI storage narrative reconstruction. After an opening plunge of over 7%, the stock quickly recovered all losses, rebounded nearly 10% intraday at its peak, and finally closed up more than 6%, with a daily intraday range exceeding 18% from the lowest to the highest point. This extreme movement from a 7% drop to a 6% rise is not a simple reflection of the overall storage industry’s health but a market revaluation of SanDisk’s value logic.
1. Macro Environment Recovery: Direct Driver of Risk Appetite Shift
The first reason for SanDisk’s price reversal is the shift in the US stock trading environment from panic to risk appetite recovery. Brent crude oil fell nearly 5%, the US 10-year Treasury yield dropped from 4.75% to 4.68%, inflation concerns eased, and the interest rate pressure suppressing growth stock valuations relaxed, leading to a broad recovery in US stocks. The S&P 500 rose nearly 1.5%, the Nasdaq gained over 2%, and the Dow Jones increased about 1.3%, hitting a record closing high. As the market continued to rise and the Philadelphia Semiconductor Index turned from a sharp intraday drop to gains, investors who shorted SanDisk early in the session faced immense pressure—they were no longer dealing with a single declining stock but a highly volatile stock deeply hit by panic sentiment while the index was broadly rising.
2. Short Covering: The Core Driver of the Stampede-Like Rebound
SanDisk’s rise was not a slow accumulation but a simultaneous low-level support and short covering, resulting in a stampede-like rebound. When the stock price stopped making new lows, shorts had to start covering; as shorts covered, the price rose; the rising price forced more shorts to retreat. The key to this process is SanDisk’s high volatility and previous sharp pullbacks, which rapidly worsened the risk-reward ratio for shorts continuing to short before earnings.
3. Earnings Expectations: The "Damocles Sword" for Shorts
SanDisk’s earnings report is imminent, scheduled for after the US market close on August 5, covering Q4 and full-year results, along with an investor event. This means short sellers face not only intraday weakness but also the risk of deciding whether to carry their short positions into earnings. The market’s real concerns are not just how much SanDisk earned last quarter but three critical questions: Can NAND prices continue to rise? How strong is the demand for data center and enterprise SSDs? Can management’s future guidance continue to exceed market expectations? If any of these questions receive positive answers, shorts may face further counterattacks.
4. AI Storage Narrative Reconstruction: From a Cyclical Product to a Core AI Architecture Participant
After the close, SanDisk and SK Hynix announced advancing the standardization of High Bandwidth Flash (HBF), aiming to bring NAND into AI inference systems, closer to compute units. Although this news was released late and not the direct cause of the intraday reversal, it reveals the new story SanDisk wants to tell. Previously, the market generally believed AI mainly benefited HBM, with NAND still seen as a cyclical product for traditional servers, phones, PCs, and enterprise storage. But in the AI inference era, models require not only speed but massive data capacity; not all data is worth placing in expensive, capacity-limited HBM. Large weight caches and datasets need a storage medium with larger capacity, lower cost, and closer proximity to compute units than traditional SSDs—this is the position HBF aims to fill.
If this technical path ultimately materializes, SanDisk will face not just the traditional NAND price cycle but an opportunity to re-enter the core AI architecture.
5. Key Follow-Up Observations: Earnings and Fundamental Validation
SanDisk’s extreme reversal does not mean all risks in the storage industry have disappeared. Questions remain about whether NAND prices can sustain their rise, whether the industry will expand production again, whether profit margins can be maintained, and whether SanDisk’s high valuation can be supported by earnings. But shorts must face a harsh fact: when a stock falls more than 7% intraday but ends up rising over 6%, it shows selling pressure can no longer create a one-sided crush. This does not mean bulls have completely won but that the forces of bulls and bears are approaching a new balance.
More importantly, SanDisk clearly outperformed the broader market and most semiconductor stocks last night, indicating that capital does not simply view SanDisk as a cyclical stock to be discarded. At least some capital is betting on three things: first, that the NAND upcycle is not over; second, that earnings and guidance may not be as bad as the market fears; third, that new technologies like HBF may grant SanDisk an AI entry ticket that did not exist before.
Of course, one strong rally cannot directly prove a trend reversal. A true reversal is not just a single day’s shift from down to up but that the next negative event does not produce new lows, the next earnings report continues upward, and fundamentals can support the rising valuation. So the real determinant of SanDisk’s direction ahead is not last night’s reversal candle but the August 5 earnings report—whether it can prove that high prices, high profits, and AI demand still exist. If earnings and guidance remain strong, last night may mark the first gunshot signaling short sellers’ retreat; but if price gains slow, demand disappoints, or management signals caution, last night’s violent counterattack may be just the fiercest flare before earnings.
Last night does not mean bulls have won but that the market has clearly warned shorts for the first time: storage can continue to fall, SanDisk can continue to fluctuate, but it is probably no longer so easy to make money by blindly shorting as before. $SNDK Pre-market surge in optical module stocks: The US may directly rewrite global optical module orders
Summary:
This sudden breakout in optical module stocks is no longer just a typical oversold rebound.
The real catalyst igniting the pre-market rally is that the Trump administration is drafting new regulations to restrict Chinese optical modules from entering US data centers.
As of around 6:30 AM Eastern Time on August 4, $AAOI was up about 13.9% pre-market, $LITE up about 9.7%, and $COHR up about 14.8%.
Even more striking, on the previous trading day AAOI had already risen about 16.9%, with LITE and COHR up approximately 9.2% and 9.6%, respectively.
Since last Friday’s close, AAOI has gained about 33% cumulatively over two rounds, LITE about 20%, and COHR about 26%.
Yesterday, the market was buying into AI hardware oversold recovery; today, it’s buying into US optical module substitution.
The US may restrict Chinese optical modules from entering AI data centers
Latest news indicates the US Federal Communications Commission is drafting a new measure to restrict new types of optical transceiver modules produced by Chinese companies from entering the US market.
Optical modules are responsible for data transmission between servers, switches, and GPU clusters, making them one of the most critical connection devices inside AI data centers.
The US government cites cybersecurity risks including data theft, malware implantation, and data center service disruptions.
If the new regulation is implemented, the US may first restrict imports of all new model optical modules, then grant exemptions to non-Chinese suppliers.
This means the policy targets not just short-term orders of any single company but the qualification for future 800G, 1.6T, and even higher-speed optical modules entering US data centers.
Zhongji Xuchuang may become the most direct target
The restriction is most likely to impact Zhongji Xuchuang.
Zhongji Xuchuang currently holds about 27% of the global data center optical module market share and is an important supplier to North American cloud providers like Amazon, Google, and Meta.
The company has previously been listed by the US Department of Defense as a so-called "Chinese military-associated enterprise," which the market usually views as a warning signal for subsequent restrictions.
What excites the market is Zhongji Xuchuang’s massive global share.
If its new products can no longer enter the US market, even partial order transfers could bring US optical module manufacturers a multi-billion-dollar incremental market.
Therefore, the market is trading not just on AI capital expenditure growth but on the potential forced global redistribution of optical module orders.
Why did COHR and LITE surge first?
COHR and LITE are the most direct potential beneficiaries.
Both companies can offer data center optical communication products competitive with Chinese manufacturers and possess technologies including lasers, optical chips, optical engines, and high-speed optical modules.
More importantly, Nvidia has previously invested $2 billion each in COHR and LITE to support next-generation AI optical communication products and domestic US capacity expansion.
The market’s past concern was that although COHR and LITE have strong technology, they struggled to compete with Chinese optical module manufacturers on cost and mass production scale.
Now, the policy may directly change the competitive rules.
If cloud providers can no longer purchase some new products from Chinese suppliers, COHR and LITE won’t need to rely on price wars to win orders; simply being compliant with US policy requirements could earn them a larger market share.
This explains why both companies, after already surging the previous day, saw nearly double-digit gains again pre-market.
Why does AAOI have the greatest elasticity?
AAOI’s logic is more straightforward.
It is a US optical module manufacturer expanding its Pearland, Texas production base, adding nearly 400,000 square feet of manufacturing space focused on boosting 800G and 1.6T optical module capacity.
Under the new policy, US domestic manufacturing capability suddenly shifts from a normal supply chain advantage to a core asset that may determine order allocation.
AAOI is smaller than LITE and COHR, with higher stock price elasticity and short interest, so any order transfer expectations are quickly amplified.
Additionally, AAOI will release its earnings report on August 6. The policy tailwind combined with earnings expectations and short covering ahead of the report jointly propelled this pre-market surge.
What does this policy mean for the industry?
Previously, market analysis of optical modules mainly focused on three variables:
Cloud providers’ capital expenditure, 800G and 1.6T demand, and manufacturers’ capacity expansion speed.
Now a fourth variable must be added: supplier nationality and production location.
If the restriction is truly implemented, the market may see three changes.
First, US manufacturers will gain a higher share of orders.
Second, optical module prices and gross margins may be supported because cloud providers have fewer supplier options.
Third, US data center construction costs may rise because COHR, LITE, and AAOI currently lack sufficient capacity to fully replace Chinese manufacturers.
In the short term, this benefits US optical module companies’ valuations and order expectations; in the long term, whether revenue can be realized depends on capacity expansion and customer certification speed.
What is the biggest risk now?
It should be noted that this measure is still in the drafting stage and is not yet a formal ban.
The policy may still be modified, delayed, or even canceled, and currently mainly targets "new model" products, not immediately removing all deployed Chinese optical modules.
Also, Chinese manufacturers have already built capacity in places like Thailand. It remains unclear whether the final policy will judge based on company nationality, production location, or supply chain control.
Therefore, today’s gains are primarily based on order transfer expectations, not realized revenue and profits.
My judgment
This round of gains and yesterday’s rebound are driven by different logics.
Yesterday, the market bought back optical modules because Mag 7 rose, AI capital expenditure concerns eased, and funds began to cover the most oversold AI hardware stocks.
Today’s pre-market second surge reflects the market starting to trade on US supply chain substitution.
The three most important upcoming signals are: the FCC’s official rule release, whether cloud providers reallocate orders, and whether US manufacturers can quickly ramp up 800G and 1.6T capacity. According to Reuters, the US is preparing to take action against China's AI industry chain again.
This time, the focus is on optical modules. The US is drafting new regulations to ban new optical modules produced in China from entering US data centers.
It is still just a draft, and the official document has not been released yet, but the market has already reacted: US optical module companies are rising, while Chinese manufacturers like InnoLight are under pressure.
Currently, Chinese manufacturers already hold a significant share of the global high-speed optical module market, with InnoLight alone accounting for about 27%. With such a large production gap, what can US manufacturers use to fill it in the short term?
So this issue cannot be seen simply as "bad for China, good for the US."
In the short term, it is indeed negative for InnoLight and positive for Coherent and Lumentum. But if the ban really takes effect, Amazon, Microsoft, and Google will also have to spend more money to buy optical modules, and the construction speed of AI data centers may be slowed down.
The US has a rosy idea: kick out Chinese manufacturers and keep the orders for themselves.
But production capacity is not something that can be conjured out of thin air by signing a document.
In the end, whether it hits Chinese optical modules or ends up hindering the US's own AI development first, it's really hard to say.
#从降息到加息，联储分歧全公开 HBF: A New Variable in the Reconstruction of AI Storage Systems
At the 2026 FMS conference, SK Hynix and SanDisk jointly released the first HBF (High Bandwidth Flash) open technical specification, marking a paradigm shift in the AI storage industry from a "single high-speed route" to "layered collaboration." HBF is not a replacement for HBM but adds a critical capacity tier to the AI storage system. Its core logic is to achieve a single stack capacity of 512GB and 3TB/s bandwidth with high parallel read capability through NAND stacking and UCle direct connection, mainly suited for "large volume, low modification" cold data scenarios such as large model weights and vector databases.
1. Technical Essence: Breakthrough from "Hardware Stacking" to "System Efficiency"
The technical path of HBF breaks the past inertia of "adding HBM when bandwidth is insufficient, and adding more HBM when capacity is insufficient." At its base, it is still NAND; a single NAND chip’s latency, random access, and write endurance cannot compare to DRAM. However, by vertically stacking 8 or 16 layers of bare NAND chips, combined with a logic base, TSV micro-bumps, and UCle interface, it achieves "high parallel reads by activating a group of NANDs simultaneously." This design shifts HBF’s core metric from "speed" to "capacity," with a single stack capacity reaching 512GB, theoretically more than ten times that of some HBM4, while achieving bandwidth levels of 3TB/s under high parallelism.
2. Industry Logic: Restructuring AI Storage Value Distribution
The emergence of HBF is changing the value distribution system of AI storage. Previously, the NAND industry competed on layer count, cost, and price, often falling into price wars when supply increased. HBF requires not only bare NAND chips but also logic chips, advanced TSV packaging, wafer bonding, UCle interface, ECC, and memory management software. Once commercialized successfully, NAND sales will no longer be just about capacity but about "how close this bit is to the GPU."
In the long term, HBF may suppress the route of "the bigger the model, the more HBM must be infinitely increased." HBM will not disappear but may revert from an expensive large warehouse to a truly high-speed working area dedicated to handling hot data, thus improving utilization efficiency.
3. Market Impact: Not a Short-Term Negative, but a Long-Term Competitive Restructuring
For HBM, HBF may not even be a short-term negative. With HBF handling cold data, HBM can specialize in hot data, improving utilization efficiency. However, in the long run, HBF may change the competitive landscape of AI storage. Previously, the market only focused on HBM, but from the release of HBF, the real competition may not just be about who has the most HBM but who can simultaneously control HBM’s speed and HBF’s capacity.
Hynix is already at the forefront of HBM and now, together with SanDisk, is betting on HBF. They are not trying to overthrow HBM but are telling the entire industry in advance: the future of AI storage cannot be a single path anymore.
4. Risks and Challenges: Commercialization Still Faces Multiple Obstacles
However, HBF cannot yet be hailed as a mature industry; several key questions remain unanswered: How high is the random latency? What about write endurance? How is the yield for 16-layer stacking? Can power consumption and heat dissipation be controlled? Is the cost truly significantly lower than HBM? And most importantly, who will be the first major customer?
Therefore, today’s HBF is not an order that can be immediately fulfilled; it is more like a new map just released by the AI storage industry. Its true significance is to lead AI storage from a "single high-speed route" to "layered collaboration," providing a more economical capacity solution for model scaling and opening new value space for the NAND industry. $SKHYNIX $SNDK The meaning of Gate is: At the same time as we paid 100,000 USDT and 800,000 ALD to the "scammer's" wallet according to the contract, Gate's alpha automatically grabbed the ALD tokens, then it couldn't be disclosed who connected the token listing process, and finally the scammer's wallet transferred the tokens into Gate alpha for an airdrop. Is that correct?
The hash is here, the answer is here
When a project has paid, got listed, and then is told "the person communicating with you is not our staff, and the project is listed on Gate" — this is already a credibility issue for Gate.Negative factors gradually digested! $BTC rebounds above $64,000, ETF funds diverge, ADA leads altcoins strongly
On Tuesday, Bitcoin fluctuated higher, rising 1.6% in 24 hours, reaching a high of $64,160, marking a new stage high since July 31. The two major negative factors suppressing the market over the weekend continued to fade, market sentiment slightly warmed, but the overall market remains in extreme fear territory with clear divergence between bulls and bears.
1. Two major negative factors settled, selling pressure basically absorbed by the market
1) Emotional impact of Coldcard cold wallet theft incident fades
Since the vulnerability was exposed on July 30, hackers have stolen 1,816 BTC from over 5,200 victim addresses, equivalent to about $114 million in market value. Initially, security panic triggered short-term risk-off selling, but after two days of digestion, the market rationally views this single-point hardware firmware vulnerability, with no sustained panic selling, basically ending the drag on the overall market.
2) Institutional Strategy reduction is a realization, not a bearish exit
From July 27 to August 2, Strategy sold 1,638 BTC at an average price of $63,957. This is the third reduction this year, with the selling price below its overall holding cost of $75,419. The cash-out funds were used for preferred stock STRC dividends and share buybacks, which is an operational capital allocation rather than a bearish liquidation of Bitcoin holdings, representing a one-time short-term selling pressure clearance.
2. Capital side: ETF funds sharply diverge, overall market sentiment cautious
1. Crypto Fear & Greed Index dropped to 25, locking in extreme fear territory, retail buying willingness is weak;
2. BTC spot ETFs had a net outflow of $61.5 million last week, then reversed to a net inflow of $170 million on Monday, showing significant short-term divergence among institutional funds;
3. ETH ETFs had a net inflow of $27.4 million last week, then a net outflow of $11.4 million on Monday, indicating weak Ethereum fund support.
Macro linkage: USD/JPY exchange rate disturbance indirectly affects the market
After US and Japan confirmed joint intervention, the yen plunged nearly 4% in one day, reigniting speculation on dollar carry trades; BTC and USD/JPY have a strong negative correlation of -0.9 for the year. A strong dollar remains a key external macro variable to watch.
3. Altcoins polarize: ADA leads strongly, ATOM’s rise hides bearish hedging
1) ADA shows independent rally, stabilizing above $0.195, a 1-month high
Weekly gain up to 24%, highlights:
- On-chain data: Cardano non-zero addresses decreased by 7,070 in the past two months, indicating this rally is not driven by retail inflows but led by major capital;
- Derivatives: ADA futures open interest hit a record high of 2.79 billion tokens, funding rate slightly positive, showing stronger long initiative, making it the most capital-focused mainstream altcoin currently.
2) ATOM up nearly 9%, but market risks evident
Spot price rises, but futures open interest nears highs with funding rate and trading spread still negative, accompanied by large short hedging positions, casting doubt on the sustainability of the upward move.
4. Derivatives data: volatility falls to low levels, options lock in key range $60,000–$72,000
1. BTC 30-day implied volatility dropped to 36%, a new low since late May, indicating short-term narrow range consolidation and strong market wait-and-see sentiment; however, volatility at historical lows makes directional breakout likely in the future;
2. Deribit options trading focus:
- Puts concentrated around $60,000, a market consensus support level;
- Calls clustered at $70,000 and $72,000, two key resistance levels;
Ethereum’s most active contract is the $1,900 call option.
Overall, funds await a range breakout, with these two key price levels determining the near-term market direction.
5. Overall market summary
Short-term negative factors cleared, bringing technical rebound, but extreme fear sentiment remains unbroken and ETF funds fluctuate back and forth. The market currently lacks the foundation for a strong one-sided rally. BTC maintains a large box range between $60,000 and $72,000; altcoin sector shows structural divergence, with ADA strongly supported by capital while most other coins passively follow the market. Priority should be on position control, avoiding blind chasing of short-term rebounds. #从降息到加息，联储分歧全公开 #财报观察员：AMD与SpaceX交卷在即，Circle压轴 #Palantir营收增93%，盘后涨13%
$SOL$OKB$BNB$XRP$DOGE$TRX$HYPE$ZEC$BCH$TAO$LINK$SUI$SHIB$PEPE$WLD$PUMP$MU$BEAT$AXTI$SNDK$SAMSUNG$KORU$SOXL$SKHYNIX$ETH$BTCComplete market review of the U.S. stock market one hour after opening
Overview of major indices
After the opening, all three major stock indices rose across the board. The Dow Jones and S&P 500 successively hit new all-time highs, and the Nasdaq steadily climbed. As of 10:30 AM Eastern Time, the Nasdaq's gain reached 1.06%, with overall market risk appetite rising.
Market divergence is very clear, with funds collectively flowing into the semiconductor hardware sector, while the established AI heavyweight stocks show mixed internal gains and losses.
Performance of the seven U.S. tech giants
NVIDIA: Maintained a strong rise of about 2%, with capital focused on computing power themes continuing to hold, showing stable support;
Tesla: Slightly oscillated upward with small volatility;
Apple: Price remained almost flat, with no capital entering consumer hardware;
Microsoft and Google slightly pressured and retreated, with early session profit-taking by short-term funds;
Meta and Amazon led the declines, with a large amount of previously accumulated profits being cashed out concentratedly within the first hour after opening.
Semiconductor & Storage sector (today's main theme)
The Philadelphia Semiconductor Index surged 5%, with AI hardware experiencing a concentrated explosive rally.
SanDisk: Opened strong, hitting an intraday high of $1415.37, then short-term funds took profits and pulled back to $1392.71, an intraday gain of 8.13%, with a turnover of 7.178 billion;
ARM and Marvell Technology surged over 11%, Intel rose 7%;
Micron and SK Hynix steadily rose around 6%, with the entire storage sector seeing short-term rotation driven by earnings expectations.
Other hot sectors
Optical communication stocks exploded, with Coherent up 16%, and multiple optical module companies rising over 7%; Palantir, beating earnings expectations, surged 20%.
The social sector weakened, with funds withdrawing from AI software and social stocks that had completed a round of strong gains, switching to chip hardware.
Current market nature and risks
This is a typical sector rotation market. NVIDIA remains the core market theme, while storage is just a short-term capital play on earnings.
Storage stocks like SanDisk have seen long positions saturate after the morning surge, with strong resistance at $1415; as long as earnings merely meet expectations, it is easy to trigger a sell-off after the positive news is priced in.
Cryptocurrency market linkage
The hot U.S. hardware sector has not driven the crypto market. Bitcoin and Ethereum remain range-bound, with all incremental liquidity flowing into the U.S. tech sector. The crypto market lacks new capital inflows. #从降息到加息，联储分歧全公开 #财报观察员：AMD与SpaceX交卷在即，Circle压轴 #Palantir营收增93%，盘后涨13% 📊 Coinbase Premium has stayed negative for 90 straight days—but Bitcoin ETFs kept buying in July.
🔹 The Coinbase Premium has remained below zero since May 6, marking its longest negative streak in the past year.
💡 However, the data tells a more nuanced story:
✅ After adjusting for USDT's slight discount, nearly half of those days turn positive, with 20 out of 31 days in July showing a positive adjusted premium.
📈 At the same time, spot Bitcoin ETFs resumed net buying in July, led by IBIT, despite the raw premium never turning positive.
⚠️ The takeaway: A negative Coinbase Premium doesn't necessarily mean weak U.S. demand. Part of the signal appears to be driven by USDT pricing, while ETF flows suggest institutional interest remains intact.
#Bitcoin #BTC #Coinbase #ETF #IBIT #USDT #OnChain #CryptoNews#韩国杠杆ETF成交额降九成，波幅收窄
🔥The trading volume of South Korea's leveraged ETFs evaporated by 90% overnight, and this "artificial bull market" has finally hit the brakes.
Brothers, the "volatility amplifier" of the South Korean stock market has finally been unplugged.
Data from the Korea Exchange on August 3 shows that the combined trading volume of 16 single-stock leveraged/inverse ETFs linked to Samsung Electronics and SK Hynix dropped to only 1.2 trillion KRW. The day before the new regulations took effect (July 30), this figure was 12.4 trillion KRW. In just two days, the trading volume was cut to one-tenth.
The largest, KODEX SK Hynix Leveraged ETF, saw its trading volume shrink from 3.6 trillion KRW to 422.6 billion KRW, only one-eighth of the original. Another, the SOL SK Hynix 2x Inverse ETF, fell from 5 trillion KRW to 180.7 billion KRW.
This is not a natural market cooldown; regulators have directly cut off leveraged funds.
A 30 million KRW threshold keeps retail investors out
The new regulation by South Korean financial regulators, effective July 31, centers on one key point: the base margin for single-stock leveraged ETFs was raised from 10 million KRW to 30 million KRW, while the qualification to offset with securities assets was canceled, and a T+2 cash payment system was introduced.
From 10 million to 30 million, a threefold threshold. Moreover, stock assets can no longer be used as margin; cash must be provided. This is fatal for many retail investors in South Korea—many accounts simply do not have 30 million KRW in cash. Previously, they leveraged through revolving financing and T+0 turnover, but now that path is blocked.
The effect was immediate. On the first day of the new rule, the trading volume of 16 products dropped from 12.4 trillion to 3.3 trillion KRW, a 75% decrease. By August 3, it further shrank to 1.2 trillion KRW.
Individual investors net sold for two consecutive days; the previously hottest KODEX SK Hynix Leveraged ETF had a single-day net sell of 23.5 billion KRW.
In short, this "semiconductor frenzy" fueled by retail leveraged funds was doused cold by regulators.
Behind the narrowing volatility: a "tail-wagging dog" monster locked in a cage
How crazy has South Korea's leveraged ETFs been in the past six months? The Korean stock market has triggered 5 full-market circuit breakers this year, while historically there have only been 11. Program trading halts have been triggered over 30 times this year.
The mastermind behind this is these single-stock leveraged ETFs.
Their mechanism is "perverse": they do not directly hold the underlying stocks but synthesize 2x leverage through swap contracts. To maintain a fixed leverage multiple, issuers must "buy high and sell low" daily before market close—buy more when the underlying rises, mechanically sell when it falls.
This "daily rebalancing" amplifies volatility in a choppy market. On March 3, when SK Hynix plunged over 10%, in the last hour before close, rebalancing-related trading volume accounted for 60% of the stock's total volume.
Wall Street calls this a "tail wagging the dog"—derivatives should follow the underlying stock, but when leveraged products grow large enough, their mechanical rebalancing dominates the underlying stock's pricing. In Korea, due to a high retail investor ratio and concentrated funds in Samsung and Hynix, this effect was magnified to the extreme.
Now with trading volume down 90%, this "tail-wagging dog" is locked in a cage. Narrower volatility is positive for the overall Korean stock market—at least it won't see crazy days of circuit breakers and wild surges.
Where did the money go? From "gamblers" back to "investors"
An interesting phenomenon: as leveraged ETFs cooled down, net purchases of regular stock ETFs expanded from 220 billion KRW to 470 billion KRW.
What does this mean? It means money hasn't left the market; it has just shifted from "high-leverage gambling" to "normal investing." Some retail investors forced to deleverage started buying regular broad-based ETFs or underlying stocks. This is positive for the long-term health of the Korean stock market—the capital structure is returning from extreme speculation to relative rationality.
But note, this shift is "passive." It's not because retail suddenly became rational, but because the threshold made leverage unaffordable. If regulations loosen in the future, leveraged funds could easily return.
Implications for crypto and global markets
This South Korean leveraged ETF drama sounds a warning bell for all high-leverage markets.
First, leverage amplifies volatility, not returns. South Korean retail investors have been crushed on leveraged ETFs in the past six months. Of 14 single-stock leveraged ETFs, 13 fell below their issue price, some dropping over 60% from their highs. 2x leverage feels great in a one-way up market but is a "volatility tax" in choppy markets—even if the underlying eventually recovers, the leveraged product's decay causes heavy losses.
Second, regulators can change market structure overnight. South Korean regulators took less than six months from opening the gate to shutting it. Announced permission in January, launched in May, raised thresholds and halted new products by July. Policy risk is always the biggest black swan for high-leverage products.
Third, direct reflection on crypto. South Korea is one of the most active crypto trading markets globally, with a strong retail speculative atmosphere. After stock market leverage is suppressed, will some speculative funds flow into crypto? In the short term, there might be a small spillover effect, but not much—because crypto is also deleveraging recently. The deeper impact is that South Korea's tightening stance on leverage may extend to crypto derivatives. If the Financial Services Commission targets crypto leverage next, crypto volatility will also be suppressed.
Summary:
South Korea's leveraged ETF trading volume dropped 90%, not because the market died, but because the "mad bull" was tamed. Narrower volatility is a medium- to long-term positive for the Korean stock market, but for short-term traders used to high volatility, it means fewer opportunities to make money.
For crypto, South Korea's regulatory tightening on leverage is a cautionary example. Globally, from stocks to crypto, leveraged funds are retreating. This is not an isolated market phenomenon but a turning point of an era.
The above is purely personal opinion and does not constitute investment advice. What do you think will happen to funds after Korea deleverages its stock market? Let's discuss in the comments.NVIDIA HBM Optimization: Demand Restructuring Under Efficiency Improvement
NVIDIA Rubin GPU's HBM technology optimization has sparked market concerns, but from a technical and industry logic perspective, this adjustment will not weaken long-term HBM demand. Instead, it may amplify market space through Jevons Paradox and restructure the AI storage value distribution system.
1. Essence of Technical Optimization: From "Hardware Stacking" to "System Efficiency"
NVIDIA's HBM optimization is not simply about reducing hardware usage but improving memory utilization efficiency through hardware-software collaboration:
1. Layered Computing Precision: Using FP8/FP4 low-precision computing during inference to reduce intermediate result memory usage, while dynamically switching precision to control precision loss.
2. Data Scheduling Optimization: Offloading dynamic data such as KV Cache and activation values from HBM to lower-cost media like SOCAMM and HBF, retaining only static data like model weights in HBM.
3. Memory Hierarchical Architecture: Building a three-tier system of "HBM-SOCAMM-SSD," automatically allocating storage media based on data access frequency to maximize HBM bandwidth utilization.
This optimization directly reduces HBM resource consumption per unit of computing power. More importantly, it breaks the linear logic of "larger model → more HBM demand" and establishes a nonlinear growth curve of "efficiency improvement → cost reduction → demand explosion."
2. Jevons Paradox Validation: The Dialectical Relationship Between Efficiency Improvement and Demand Growth
Jevons Paradox points out that improving resource use efficiency may increase total consumption because cost reduction stimulates more demand. This theory has been validated in the AI storage field:
- Kimi K3 Model: Inference efficiency improvement reduced single Token cost by 50%, but call volume increased tenfold, ultimately increasing total server DDR5 and eSSD demand.
- Google Gemini 3: Core AI response cost dropped by 30%, but single-minute Token processing rose from 10 billion to 16 billion, directly driving data center storage demand.
- NVIDIA GB300: Single Token cost dropped 35 times, promoting large-scale deployment of AI inference clusters by cloud providers, with HBM total procurement growing 200% year-over-year.
These cases show that the core driver of AI storage demand is not unit resource consumption but total Token generation. NVIDIA's HBM optimization essentially lowers AI inference marginal cost, unlocking many application scenarios previously unfeasible due to high costs, such as AI Agents, long-context conversations, and real-time video generation. The explosive growth of these scenarios will fully offset the impact of reduced unit resource consumption and even bring greater total demand increments.
3. Real Structure of HBM Demand: From "Capacity Competition" to "Value Stratification"
Market concerns about HBM demand essentially stem from a single-dimensional understanding of "capacity growth." But according to Rubin GPU official parameters, NVIDIA has not reduced HBM hardware configuration:
- Single Card Capacity: Rubin GPU supports up to 288GB HBM4, a 104% increase over Blackwell's 141GB.
- Bandwidth Improvement: HBM4 bandwidth reaches 22TB/s, a 175% increase over Blackwell's 8TB/s.
- System Integration: Vera Rubin NVL72 system carries 20.7TB HBM4, 2.8 times the previous generation.
This "optimization + expansion" combined strategy reveals the real structural change in HBM demand:
1. Rigid Demand for High-Value Data: Core data like model weights and high-frequency KV Cache still need to be stored in HBM, and this demand will continue to grow rigidly with model parameters.
2. Diversion of Low-Value Data: Cold data and intermediate results are diverted to media like SOCAMM and SSD, reducing this demand but improving HBM resource utilization efficiency.
3. Exponential Growth in Bandwidth Demand: Scenarios like long context and high-concurrency inference require memory bandwidth far more than capacity, making HBM4's bandwidth improvement a core competitive advantage.
This structural change means HBM's value logic has shifted from "capacity scarcity" to "bandwidth scarcity," and NVIDIA's technical optimization, by improving bandwidth utilization, further strengthens HBM's irreplaceability in the AI storage system.
4. Industry Landscape Restructuring: Opportunities and Challenges for Storage Vendors
NVIDIA's HBM optimization will have a profound impact on the storage industry landscape:
(a) Opportunities for HBM Vendors
1. Continuous Growth in Total Demand: According to SemiAnalysis forecasts, HBM supply will be about 6% below demand in 2026, expanding to a 9% gap in 2027. NVIDIA's technical optimization will further amplify this gap because demand growth driven by efficiency improvements will exceed the decline in unit resource consumption.
2. Product Structure Upgrade: HBM4's bandwidth advantage will become a core competitive edge. Leading vendors like SK Hynix, Samsung, and Micron will benefit from product structure upgrades, with gross margins expected to remain above 50%.
3. Long-Term Contract Lock-In: Cloud providers will sign long-term contracts with storage vendors to secure HBM capacity for AI inference capabilities, further strengthening storage vendors' bargaining power.
(b) Challenges for the Storage Industry
1. Urgency of Capacity Expansion: HBM capacity is expected to grow from 123K wafers/month at the end of 2023 to 331K wafers/month by the end of 2025, a 2.7x increase in two years, but still insufficient to meet demand. By the end of 2026, it is expected to reach 473K wafers/month, and 668K wafers/month by the end of 2027, a 5x increase in four years. However, bit conversion rate is only 1/3 of general DRAM, so capacity expansion speed still lags demand growth.
2. Pressure of Technology Iteration: The validation progress of HBM4 has uncertainties. NVIDIA is concurrently evaluating multiple HBM specifications for Rubin Ultra, including HBM4e 8hi and HBM4 12hi, posing higher requirements on storage vendors' technology iteration capabilities.
3. Cost Control Challenges: HBM manufacturing costs are high, including advanced packaging and high-bandwidth interfaces. Storage vendors need to reduce unit costs through economies of scale and technological innovation to cope with downstream customers' price pressure. $SNDK $NVDA Four earnings reports next week, Circle is the finale
There are four earnings reports next week:
Palantir, AMD, SpaceX, Circle.
But honestly, the information the first three can provide has basically been given in the last round of tech giant earnings, which is to verify whether AI demand is genuinely strong or just hype. This expectation has been repeatedly traded in the market, so even if there is deviation, the marginal impact won't be too large.
Instead, the last one, Circle, is the most worth watching.
Coinbase revenue dropped 18.5%, Robinhood crypto revenue fell nearly 40%, and USDT growth has stopped. Three reports from the crypto side have already been submitted, all pointing in the same direction—the market is shrinking. But one thing hasn't been confirmed yet: is the capital really leaving, or just changing containers?
Circle's answer can confirm this. If USDC circulation is rising, it means the money hasn't gone far, just waiting for a compliant entry point. If it also falls, then it really is a net outflow, and the stablecoin market is shrinking overall.
This concerns a more fundamental question:
whether institutional funds are still waiting at the door. The scale of compliant stablecoins is a leading indicator of this issue.
As for the rest, there's nothing much to focus on, just do your own well.
$BTC $ETH
#DailyOrbit #FedSplitGoesPublic #BigTechEarningsWatch #PalantirBeatAndRaise SPCX First Earnings Report Options GEX Preview
$SPCX
This is an options performance dominated by negative GEX
Meaning SPCX will amplify volatility regardless of whether the earnings report is good or bad
From the options flow, there are a large number of sell puts executed around $90-$110
This implies that SPCX's lower bound will have considerable buying support between $90-$110
————————
Key Information:
Recently, SPCX has experienced two consecutive months of negative growth in Starlink, which is SpaceX's largest revenue segment. This may impact SPCX's revenue expectations, and everyone should be well aware of this. (See Figure 3)
#SPCX首份财报将公布，千亿美元解禁在即 🐋 Long-Term Holders may be signaling the start of Bitcoin's next major rally.
📊 Historically, $BTC bull markets unfold in two phases:
🔹 First rally: Long-Term Holders (LTHs) gradually take profits.
🔹 Pullback: LTHs aggressively accumulate more $BTC.
🔹 Second rally: LTH supply declines again as the strongest leg of the bull market begins.
💡 In the current cycle, LTH accumulation has peaked and is now starting to decline—a pattern that has previously preceded Bitcoin's second major rally.
⚠️ This cycle has already lasted 31 months, much longer than previous ones, likely due to the impact of Spot ETFs and sustained institutional demand.
👀 One key takeaway remains: Long-Term Holders still control a record amount of BTC, and their supply has begun to decrease once again.
#Bitcoin #BTC #OnChain #Whales #LongTermHolders #Crypto #BullMarket #ETFAmerican Bitcoin, the mining company under the Trump family, released its Q2 financial report.
A net loss of $57.2 million, marking the third consecutive quarter of losses.
Logically, such a financial report should cause the stock price to crash immediately.
So what happened? The stock price rose by over 5% after the report was released.
Is the market crazy?
No. The market is clearer than anyone.
Let's look at the numbers:
In Q2, the company's Bitcoin reserves increased from 7,021 to 8,002 coins, a 14% growth in a single quarter.
During the same period, it mined 932 Bitcoins, setting a company record.
Mining revenue was $67 million, an 8% quarter-over-quarter increase.
Where did the $57.2 million loss come from? A $71.2 million impairment of Bitcoin holdings' fair value.
Translation:
BTC price dropped, so the book value of holdings was revalued at market price, decreasing by $71.2 million. This is not a cash outflow; it's an accounting game.
The company didn't sell a single Bitcoin.
All 932 mined coins were added to the reserves.
The loss is a book figure; the increased holdings are real money.
What did Eric Trump say?
"Bitcoin never moves in a straight line, and when we established this company, we never assumed it would."
"Our advantage is that we don't buy at market price; our mining cost is about half the market price."
Mining cost per BTC is about $36,500.
On that day, BTC price was $63,150. Mining one yields a profit of $26,650.
Would you do this business?
Even more interesting—
In the same week, Trump Media transferred $165 million worth of BTC (about 2,628 coins) to Crypto.com.
The company said, "No sale, just custody transfer."
But on-chain data clearly shows BTC holdings dropped to about 4,261 coins, exactly equal to the amount pledged for convertible bonds.
One company is desperately hoarding, another is quietly leveraging.
Same family, same coin price, completely opposite allocation choices.
American Bitcoin's CEO said, "Our view of the world is simple: Bitcoin is a growing capital asset, and we believe its long-term compound growth will exceed our cost of capital."
What about Trump Media? They use BTC as collateral to borrow money.
One treats it as a son to nurture, the other as a tool to use. Which do you believe?
Finally, a couple of practical points:
The stock price has dropped 95% from its peak and was forced to do a 1-for-15 reverse split to maintain Nasdaq listing eligibility.
But the market still responded with a 5% increase.
Because everyone understands:
This is not a money-losing company.
This is a machine that continuously hoards BTC using mining cash flow.
Accounting losses are on paper; 8,002 BTC is real.
At the current price of about $63,500 per coin, the reserves alone are worth $500 million.
"Mining company losses under fair value accounting"—is this a worrying signal or a hurdle that must be overcome?
My answer is the latter.
If you believe BTC will rise long-term, then the current book loss is just the holding cost.
If you don't believe, then you shouldn't even look at this company. The market has already answered this question with a 5% price increase. 🚨💸 Coldcard Wallet Hack Update: Official Damage Exceeds $100 Million
According to the latest report from Galaxy Research, the damage from the firmware vulnerability in the Coldcard hardware wallet has exceeded $100 million.
Approximately 1,596 $BTC were stolen from around 7,300 different wallet addresses.
The attackers carried out three major waves of attacks along with about 14 other smaller incidents exploited by various parties.
Approximately 90% of the stolen Bitcoin remains in the attackers' wallets and has not yet been transferred to exchanges. Furthermore, researchers are investigating a suspected fourth attack, which could bring the total damage to $130 million.📊 Morning Market Snapshot on August 4: BTC found support near $63,300 and started to rebound, with overall sideways consolidation over the past 24 hours. The key resistance level for the day is at $64,000; after touching this price, fluctuations and reversals began, showing weak short-term directional strength.
🧭 Technically, BTC is still above the EMA line, but upward momentum has clearly weakened. Market sentiment remains defensive, with no strong bullish signals released. From the chart structure, it looks more like a tug-of-war between bulls and bears in a critical zone.
🎯 Today's main focus is simple: whether BTC can continue to hold above $63,000. If this level can be repeatedly confirmed as stable, there may be short-term buying opportunities on pullbacks. The trading approach could consider looking for dip-buying points in the $63,000–$63,400 range, targeting the $64,100–$65,200 area.
⚠️ It is important to note that the current environment is a defensive consolidation, with high risks in chasing highs. Patience in waiting for the right position is more important than rushing into the market. Market trends change rapidly; discipline should always come first.A publicly listed company has locked nearly 5 million $ETH. Will retail holders really end up with fewer coins?
#BitMine成全球最大ETH质押方
As of July 26, BitMine holds about 5.787 million ETH, accounting for 4.8% of the total supply; among them, 4.917 million have already been staked, meaning it has locked 85% of its ETH into the network. According to the company's disclosure, it is currently the largest single entity staking ETH globally.
It is clearly not satisfied with just being a large coin holder; its goal is to acquire 5% of the total ETH supply and earn staking rewards through its MAVAN validator network. If all holdings are staked, the company estimates it can receive nearly $300 million in rewards annually.
This is certainly a medium- to long-term positive for ETH.
Millions of ETH moving from market circulation into staking means fewer coins available for short-term sale; BitMine is no longer just betting on coin price but starting to generate cash flow from ETH itself. Previously, listed companies hoarding coins was like putting gold in a safe; now this "gold" can continuously earn interest.
This news supports a long-term bullish outlook but is unlikely to be the sole reason for a short-term surge in ETH price. What to watch next is whether BitMine will continue buying, whether staking nodes are sufficiently decentralized, and whether Ethereum chain transactions and fees can recover in tandem.
$ETH 1. Fundamentals (Underlying SpaceX entity fundamentals, indirectly affecting SPCX sentiment)
1) Positive supporting factors
1. On July 24, the 13th Starship test flight successfully completed the full flight process, successfully deploying Starlink V3 satellites, soft landing, verifying the basic flight framework. This provides marginal short-term aerospace narrative benefits and can bring pulse-like emotional buying;
2. Starlink's global satellite network is steadily being established, generating stable and continuous operational cash flow, which is the company's only stable cash-generating business; meanwhile, it has signed space computing power cooperation orders with AI companies, presenting a potential second growth story in the long term;
3. The US stock IPO has been completed, raising $75 billion in ample cash, so there is no short-term risk of cash flow disruption or sudden default.
2) Core suppressing negatives (leading to a moderately weak mid-term pattern)
1. Aerospace launches and Starship R&D continue to burn large amounts of cash, with overall significant losses. After the IPO, the market shifts from "speculating on expectations" to focusing on commercialization realization, reducing patience; Starship still requires months or even years to verify stable reusable launches and commercial manned/cargo operations. A single successful test flight cannot completely reverse the pessimistic valuation expectations;
2. Early August will see concentrated unlocking of low-cost primary shares from the IPO, causing significant selling pressure on the US stock common shares, which will directly drag down the crypto-side SPCX linkage;
3. Elon Musk's attention is divided (Tesla, X platform, AI business), and the market worries that SpaceX's execution progress may not meet earlier aggressive expectations; US FAA aerospace launch approvals still face regulatory uncertainties.
Summary: No fundamental qualitative positive changes, only sporadic event-driven catalysts, no hard performance to support a trend of large gains.
2. News (short-term direct market drivers)
Positive catalysts (only pulse rallies)
1. New round of official Starship launch/test flight success announcements;
2. Official announcements of large commercial launch orders, Starlink overseas major client renewals, large space AI computing power orders;
3. Rising expectations of Federal Reserve rate cuts, overall recovery of Nasdaq tech stocks boosting high-growth sentiment; Elon Musk publicly expressing confidence in SpaceX's long-term value.
Negative suppressors (high probability of near-term realization)
1. Concentrated selling of unlocked shares, pressure on US stock common shares, derivatives passively following down;
2. Starship test flight faults causing aborts, FAA issuing restriction and rectification notices;
3. Overall pullback in US tech sector, weakening BTC in crypto market, high-volatility themes prioritized for sell-offs;
4. Exchanges adjusting SPCX contract fees, leverage, and position limits, causing on-exchange funds to flee for safety.
Current immediate news environment
No major sudden positive news recently; the market is in an emotional vacuum after positive realizations, with funds waiting for SpaceX's first official financial report post-listing. The news environment is overall neutral to weak.
3. Technicals (Current SPCXUSDT 118 price structure)
Daily level
1. Clear overall bearish trend: 5/10/20-day moving averages all downward, price running below moving averages long-term, each rebound is on low volume followed by gradual decline, with a large amount of high-level trapped positions piled up above (intense trapped selling pressure in the 130~160 range);
2. Current support and resistance:
- First support: 115 USDT (recent local low, short-term psychological support); strong support at 108~110 USDT, breaking below this range opens a new downward space;
- First resistance: 125 USDT, then 132~135 (US stock issue price benchmark, strong resistance, difficult to break through in one go short-term);
3. Volume characteristics: 24-hour volume continuously shrinking, on-exchange trading activity declining, mainly stock game, exhausted incremental chasing funds, very poor sustainability of volume-less rebounds.
Hourly level (currently near 118 in night session)
Narrow range consolidation bottoming, bulls and bears temporarily balanced, bulls lack active attack momentum, bears' selling pressure temporarily exhausted, short-term likely narrow sideways waiting for news/US stock linkage guidance.
4. Capital flow
1. Crypto on-exchange: early-stage speculative hot money gradually leaving high levels, only retail short-term speculative chips remain, large long positions proportion decreasing, short positions slowly rising;
2. US stock linkage funds: IPO unlocking imminent, primary profit-taking funds strongly willing to realize gains, institutional funds tend to be cautious, no large-scale active bottom-fishing;
3. This derivative's liquidity is far lower than mainstream BTC/ETH, large orders easily cause spikes and slippage, weak liquidity amplifies price volatility.
5. Short-term cycle trend forecast (next 3~15 trading days)
Scenario 1: Baseline neutral trend (65% probability)
Range: 110 ~ 125 USDT wide range oscillation
No major positive news currently, unlocking selling pressure suppresses + funds cautious, night session first consolidates slightly between 116~122;
- Faces resistance near 125 and falls back, returning to around 115 to consolidate bottom;
- Dips near 110 will attract short-term bottom-fishing funds for small recovery rebounds, difficult to form a one-sided trend, mainly back-and-forth shakeouts.
Scenario 2: Optimistic pulse rebound (25% probability)
Trigger conditions: sudden Starship launch positive news, Nasdaq strong recovery, BTC strong rally driving market sentiment;
Trend: rapid surge to 128~132 strong resistance, hitting issue price benchmark area, trapped positions concentrate selling, surge then fall back, a short-term emotional arbitrage rally, hard to stabilize above 135 to form reversal.
Scenario 3: Breakdown downtrend (10% probability)
Trigger conditions: US stock unlocking with heavy volume drop, sudden Starship negative news, crypto market systemic bear;
Trend: effectively breaks 115 support, quickly drops to 108~110, extreme sentiment tests 100 round number.
6. Key trading risk highlights
1. Product structure risk: This is only an exchange synthetic derivative, not real stock. In extreme markets, exchange premium/discount divergence, trading suspension, contract rule adjustments may occur, with no reliable protection of position equity;
2. Trend risk: Mid-term bearish structure unchanged, currently only suitable for short-term light position range trading, heavy bottom-fishing for reversal has very low cost-effectiveness;
3. Time window risk: Early August unlocking is the biggest potential negative window, avoid heavy long positions during this period;
4. Crypto derivatives inherently have leverage; narrow range spikes easily trigger liquidation, strictly control leverage and position size.Italy's largest bank reduces $BTC holdings but shifts funds to staked ETH ETF
Italy's largest bank Intesa Sanpaolo made significant adjustments to its crypto ETF portfolio in Q2. Its holdings of BlackRock Bitcoin Spot ETF IBIT common shares dropped from 646,809 shares to 40,723 shares, a decrease of 606,086 shares, or 93.7%; the IBIT call options also shrank from 2,496,500 shares to 18,000 shares, a 99.3% reduction.
However, it has not completely exited the crypto market.
The latest holdings show a new put option position corresponding to 500,000 IBIT shares; meanwhile, BlackRock's staked Ethereum ETF holdings increased from 116,200 shares to 349,600 shares, a net increase of 233,400 shares, tripling the total size. On the other hand, Bitwise Solana Staked ETF holdings dropped from 2,817 shares to 7 shares, shrinking about 99.75%; XRP Trust holdings remained steady at 712,319 shares.
This data looks more like a portfolio reallocation rather than a simple "crypto liquidation."
The spot exposure and upside leverage on the $BTC side were both reduced, while put options were added for protection, indicating the bank is clearly strengthening downside risk management. However, the 13F filing does not disclose option strike prices, expiration dates, premiums, or Delta, so based on the 500,000 shares scale alone, it is unclear whether this is an active bearish bet or a protective hedge for the remaining IBIT position.
The real noteworthy change is in $ETH. Traditional IBIT holdings were reduced, but staked Ethereum ETF holdings increased against the trend, suggesting funds are shifting from pure price-tracking $BTC products to $ETH products that offer both price exposure and staking yield expectations. If this trend continues, institutional allocation logic may shift from "whose narrative is strongest" to "who can provide more stable on-chain cash flow."
As for $SOL being mostly cleared out and $XRP remaining unchanged, this indicates the bank is not broadly betting on altcoins but is making very specific asset selections.
Therefore, this adjustment is cautious on $BTC, relatively positive on $ETH, and clearly cooling on $SOL. It is not an exit from crypto but a repositioning into structures that better align with the bank's risk preferences.
This is purely personal market observation and does not constitute investment advice. DYOR.
#交易之声：你的经验值得被听到 After Bitcoin dropped to 80,000 last year and then rebounded sharply, New York Fed President Williams came out to release information. The Fed has canceled the original forward guidance and gradually evolved into using voting distributions and regional Fed presidents' hints as alternative guidance. The advantage of this approach is distributed responsibility; if any unexpected news shocks the market, no one can be held accountable. However, this news flash is actually rehashing old news, with the source being an interview from last Friday.
Based on his past record, he seems more like a forerunner for signaling, and Friday's speech was essentially explaining the reason why the FOMC held steady this time. It follows the typical approach of shifting responsibility or decision-making power to the data, thereby distancing the Fed from responsibility in case of market turmoil.
From Trump's perspective, ensuring a steady rise in the stock market before the midterm elections is definitely the best plan. Theoretically, what he can do is to halt actions on Iran and tariffs to reduce inflation. Given that he recently secured quite a bit of money from Japan and South Korea, temporarily easing off on oil and taxes also has an economic basis. But practically, tariffs are one of his governing foundations and cannot be removed, and Iran won't allow easing on oil. So ultimately, it comes down to using the money harvested overseas to provide welfare and buy votes.#FedSplitGoesPublic #BigTechEarningsWatch #PalantirBeatAndRaise All three earnings reports might disappoint, but the scariest thing is: after the disappointment, the stock still doesn't rise
This week, AMD, Western Digital, and SanDisk will release their earnings reports, and the market is hoping they can pull tech and storage stocks out of the hole. Many believe that as long as the results beat expectations, the stock price will reverse. But I want to ask a harsher question first: what if all three reports are good?
If AMD's data center continues high growth, Western Digital's cloud drives remain in short supply, and SanDisk's gross margin stays ridiculously high, yet the stock price only spikes after hours and then gets crushed the next day, what does that mean?
It means the market has never lacked good news; what it lacks are people willing to keep paying for that good news.
So what really matters this week is not whether these three companies deliver a knockout blow, but whether the players at the table are still willing to bet after the knockout is thrown. Because today's market no longer rises just because of good earnings; good earnings alone are not enough, even a big beat might not suffice—you have to prove next quarter will be better, next year will be better, and growth can't slow down.
Stocks are not bought for profits, but for the illusion of profits continuing to expand. Once that illusion stops expanding, even the most beautiful profits might just be flowers at a funeral.
1. AMD: More than just a cheaper alternative to Nvidia
Many people interpret AMD's earnings as whether the "cheaper alternative to Nvidia" can turn things around, but that's too shallow. What AMD really needs to answer is: is AI capital expenditure still expanding, or has it shifted from "buying everything" to "choosing selectively"?
If AMD's data center continues to grow and AI chip revenue keeps rising, the market will believe cloud providers haven't hit the brakes yet, and at least a second computing power system beyond Nvidia still has buyers.
But here's the problem: if AMD's results are good only because Nvidia's deliveries are delayed and customers temporarily bought some AMD chips, is that growth? No, that's just picking up leftovers.
AMD must prove customers buy from them not because Nvidia is out of stock, but because AMD has truly entered their long-term architecture. Being a substitute who can score doesn't mean the coach has decided to start you.
So the most important thing in AMD's earnings is not how much they earned this quarter, but whether MI450 Helios and subsequent system-level products are just PPT slides from the launch or real cash on customer purchase orders.
If AMD only delivers a "decent" report, the stock price might actually be at risk. Because against high expectations, meeting expectations is underperforming, and slightly beating expectations means growth is slowing.
The most expensive stocks fear not bad news, but good news that no longer excites.
2. Western Digital: The data graveyard of the AI era
Many ask, with AI so hot, why look at a company selling mechanical hard drives?
Because AI models can deceive, orders can be packaged, capital expenditures can circulate, but data doesn't disappear out of thin air. Training a model leaves training data; running inference generates logs, caches, videos, voice, and enterprise data.
GPUs create the data, hard drives are responsible for burying it. So Western Digital isn't selling hard drives; it's selling the data graveyard of the AI era.
Computing power is fireworks; data is the trash left scattered after the fireworks fall.
Western Digital's earnings need to verify: is AI just frantically buying chips, or has it started truly generating massive amounts of data?
If Western Digital's cloud shipments, per-drive capacity, prices, and long-term orders all rise simultaneously, it means AI capital expenditure is moving from "buying shovels" to "digging out the mine."
But if Western Digital's profit growth mainly relies on price hikes while shipment capacity slows, be cautious. Because price increases don't necessarily mean strong demand; they might mean suppliers are shrinking supply and wildly raising prices during the last shortage.
Price hikes can be evidence of prosperity or the last harvest before a cycle peak.
3. SanDisk: Excessive profits alone are not good news
SanDisk might have the most exaggerated numbers and the wildest stock reaction among the three. NAND price increases, enterprise SSD shortages, long-term contract lock-ins—almost all good news is concentrated here.
Sounds stable, right? Quite the opposite.
SanDisk's biggest risk is the market treating it like a money-printing machine that never stops.
When gross margins are high, the market won't praise you; it will ask: how much longer can this last?
When contract amounts are large, the market won't be satisfied; it will ask: have future years' demands already been signed off in advance?
When everyone knows you're making excessive profits, the profits themselves are no longer good news. Only if the excessive profits continue accelerating can they feed valuation.
SanDisk is trying to turn a cyclical stock into a contract stock, turning today's price hikes into cash flow for years to come. But there's a bold question: do these long-term contracts put a safety belt on SanDisk, or put a ceiling on future prices?
If NAND prices continue rising but SanDisk has locked in prices in advance, then what looks like the safest contract today might become a profit cap tomorrow. Contracts can lock in downside but also lock in upside.
4. The real reversal is more than just beating expectations
So can these three earnings reports trigger a reversal?
My answer is: just looking at "beat expectations" is far from enough.
AMD must prove AI computing power purchases are still accelerating, Western Digital must prove computing power is truly turning into a flood of data, and SanDisk must prove storage price hikes are not a one-time excessive profit but a sustainable profit mechanism.
If any one of the three chains is missing, the market might define any rise as just a rebound. Only if all three conditions are met simultaneously—computing power keeps buying, data keeps growing, and storage profits keep locking in—will the market acknowledge that the AI industry is not overdrawing the future but creating a new future.
But there's an even more dangerous indicator this week: if earnings greatly beat expectations but the stock price can't move up, that might be scarier than an earnings bomb.
An earnings bomb only means the company has problems. Good earnings with no price rise means the entire market's pricing system has problems. It means bulls have already bought all the good news they can think of, and even if more good news comes, it can only help those who bought earlier break even.
So the real knockout this week is not how pretty the numbers these three companies deliver are, but whether the funds, after seeing the pretty numbers, continue to rush in or take the opportunity to escape. I am a member of the ALD community, and we are co-builders of web3. Gate has seriously affected industry compliance and has significant internal vulnerabilities, which greatly impact the industry. As a leading figure in the industry, we need to jointly maintain industry transparency and credibility. We ask the official management to pay attention to: "Regarding the listing matters between Gate and the ALD community."Why do many people think altcoins are doomed as soon as $BTC market dominance rises?
I actually see it differently.
When $BTC market dominance rises, it often doesn't mean altcoins have no chance at all; rather, the market is re-confirming its risk anchor. Funds return to BTC first, indicating that everyone wants to survive first and prefers to put their positions back into the assets most easily accepted by the market. After BTC stabilizes market sentiment, then funds will look for higher volatility targets.
So the rise in $BTC market dominance shouldn't be seen only as altcoins being suppressed. Sometimes it's just a phase within a bull market: first concentration, then diffusion; first defense, then offense.
I now prefer to understand it as BTC holding up the market first, so altcoins have room to grow afterward. #从降息到加息，联储分歧全公开 #财报观察员：AMD与SpaceX交卷在即，Circle压轴 #Palantir营收增93%，盘后涨13% SpaceX will release its first earnings report since going public tonight, but before the financial figures are finalized, the market has already ramped up volatility.
SPCX is currently priced around $116.40, about 13.8% below the $135 issue price, and down approximately 42% from the post-IPO high of $201.80. The options market expects the stock price to fluctuate about 15% around the earnings release. Based on the current market cap of roughly $1.5 trillion, this expected volatility corresponds to a market cap change of about $225 billion.
A bigger supply variable is coming on August 6.
On that date, approximately 911.5 million shares held by insiders and early investors will become eligible for sale. At $116.40 per share, this corresponds to a market value of about $106.1 billion. At the time of SpaceX’s IPO, only about 5% of shares were in circulation. Roughly estimated, this unlock represents about 1.5 times the initial float.
Unlocking does not mean all 911.5 million shares will be sold simultaneously, but it will alter the pricing structure originally supported by scarce shares. Currently, about 63% of the float is lent out, indicating very high short interest. Earnings beating expectations could trigger short covering, while the unlock might increase sustained selling pressure during any rebound. These two forces will simultaneously impact the price.
This is especially important for those trading SPCX on crypto platforms.
SPCX perpetual contracts trade 24/7, with one contract representing exposure to one share price, and funding rates settled every 8 hours. Tokenized stocks like XSPCX also offer 24-hour price trading but do not represent ownership of actual SpaceX shares nor confer shareholder voting rights.
Therefore, what happens on August 6 is not an "on-chain token unlock." The real increase is in the underlying shares available for sale in the US stock market. This impact will first be reflected in SpaceX’s spot price, then transmitted to indexes, tokenized stocks, and perpetual contracts. If the on-chain market remains open after the US market closes for earnings, there may be short-term price jumps, liquidity thinning, and widening contract premiums.
The market expects SpaceX’s Q2 revenue to approach $7 billion, but this time, simply exceeding revenue expectations is not enough. The stock price must also absorb the 15% expected volatility, the roughly $106.1 billion potential unlock size, and already crowded short positions.
The earnings report will determine the initial gap move, but the true trading volume after the unlock will decide if the rebound holds. For SPCX, this is more like a stress test jointly conducted by traditional equity supply and crypto-style around-the-clock derivatives trading.
#SPCX首份财报将公布，千亿美元解禁在即 Although spaceX caused Elon Musk's net worth to plummet, it does not affect my continued bullish outlook
Just understand these four points:
1. Starlink: From a money-burning monster to a cash flow cow
Self-sustaining ability: Explosive global subscriber growth, securing massive B2B orders from Boeing, aviation, navy, and emergency communications.
Breaking the fate: Traditional aerospace relies on government budgets, while Starlink provides extremely rich EBITDA and free cash flow, offering continuous self-sustaining ability for deep space exploration.
2. Deep integration with xAI: New narrative of space + AI computing power
Reconstructing valuation: Valuation logic leaps from hard tech manufacturing directly to global AI infrastructure.
Orbital data center: Combining ground computing power with near-Earth orbit Starlink network opens up imagination space for global supercomputing and cloud computing.
3. Starship maturity: Absolute monopoly on launch costs
Cost dimension reduction strike: Falcon 9 handles high-frequency reuse, while Starship reduces unit payload orbit cost by several orders of magnitude.
Flywheel effect: Extremely low cost makes deploying large-size Starlink satellites more efficient, further strengthening network effect barriers.
4. National defense security and extremely high moat
Starshield: Government and defense-grade private network contracts create a solid revenue floor.
Near-Earth orbit overlord: Holds de facto dominance in heavy-lift capacity and global low Earth orbit satellite chain.
Starlink's cash flow + Starship's monopoly cost + AI infrastructure valuation premium, the triple logic overlay jointly supports SpaceX's value leap
$SPCX
#SPCX首份财报将公布，千亿美元解禁在即 The scarcer the storage, the more the stock price falls: The market has already started trading the shortage's funeral
In August 2026, the storage industry suddenly dropped a bombshell — the DRAM and HBM full-year 2027 capacities of the three giants Samsung, Micron, and SK Hynix have already been sold out in advance, and NAND Flash capacity is also about to be fully booked. This means that the storage capacity for the next year and a half has been completely locked in, with cloud providers and AI companies prioritized for supply, while smartphone, PC, and other consumer electronics manufacturers may face an extreme situation of "even with money, they can't buy."
Logically, such "capacity sold out" news should have sent storage stocks soaring collectively, but the reality is absurd like a joke: SK Hynix delivered record profits, yet its stock price plummeted; Samsung's chip business profits surged, but its stock price rose then fell back; enterprise SSDs are insufficient, PC DRAM is extremely tight, consumer memory prices keep rising, yet storage stocks not only didn't rise but frequently plunged.
So is the industry lying, or are the stocks crazy? The answer is: the industry is not lying, the stocks are not crazy, what really changed is what the market is trading.
1. From "Will there be a shortage?" to "Can it get even worse?": The game of expectations is over
In the past, the market traded on "whether storage will be in shortage," now it trades on "it is already this short, can it get even worse?" This sentence is very important — supply and demand determine profits, but expectations determine stock prices.
Now HBM is tight, server DRAM is tight, enterprise SSDs are tight, which has almost become market consensus, even 2027 capacity is being fought over in advance. When everyone knows the same thing, it no longer counts as an "expectation gap," but a "script already written into the stock price." The market will not pay twice for the same shortage logic.
More crucially, stocks never trade on "whether storage prices rise," but on "the speed of price increases." Suppose prices rose 90% in Q1, 60% in Q2, 20% in Q3 — prices are still rising, companies are still making money, but stocks may have already started falling. Because the market sees "prices still going up, but the acceleration of the rise is slowing down."
The most dangerous time for cyclical stocks is often not when products start to drop in price, but when products are still rising in price, yet stock prices refuse to rise. Spot prices look in the rearview mirror; stock prices look through a telescope.
2. "Comprehensive shortage" is a lie: Structural shortage is the truth
The so-called "comprehensive storage shortage" actually mixes several completely different markets into one sentence. The real tightest are HBM, server DRAM, high-capacity enterprise SSDs, and some PC memory squeezed by capacity shifts. But consumer NAND is not simultaneously short in all models, enterprise SSDs are fighting for capacity, while the consumer market is already being suppressed by high prices.
Some SSD retail prices have surged, possibly due to channel inventory, product discontinuation, and supply structure adjustments, which does not mean all NAND wafers are severely short. So the real industry status now is: high-profit AI storage is extremely tight, low-profit consumer storage is squeezed in supply, but consumer demand itself is being slowly strangled by high prices.
3. Price hikes turn from "positive" to "poison": The start of demand suicide
Qualcomm has begun warning that memory costs erode profits; smartphone, computer, and server manufacturers all face rising costs. For every extra dollar storage makers earn, downstream must pay an extra dollar. This transmission chain is brutal: storage price hikes → terminal cost rises → terminal manufacturers raise prices or reduce specs → consumers delay upgrades → sales decline → ultimately storage demand growth slows.
So the initial storage price hike is called "bargaining power," but when prices rise to the point customers can't sell products, it is called "demand suicide." Shortage can raise unit prices, but if unit prices ultimately kill sales, profit growth shifts from "volume and price both rising" to "only price holding up."
4. High profits are a "suicide ad": Capital is on its way
When storage makers' gross margin and operating profit margin hit historic extremes, the market doesn't just think "how much more can they earn," but "will Samsung expand production? Will Hynix expand? Will Micron expand? Will Chinese manufacturers accelerate catching up?"
High profits themselves are the strongest advertisement for future new supply; the closer profits get to heaven, the faster capital rushes to build highways to heaven. Today customers fight for 2027 capacity, giving manufacturers more reason to expand 2028 supply. So today's sellout is both a short-term positive and possibly proof of financing for the next round of capacity oversupply.
5. Sold-out capacity ≠ real demand: The fog of panic inventory
If a manufacturer can only deliver 70% of requested volume, customers wanting 70 units might apply for 100 units first; if one manufacturer is not trusted, they might apply for quotas from multiple suppliers simultaneously. Fear of shortage leads to early orders, early orders create more severe shortage, shortage stimulates more hoarding.
Orders increase, but how much is real consumption and how much is panic inventory, no one can say for sure. Sellout proves "capacity is being fought over," not that every chip has found its final consumer.
6. The collapse of crowded trades: expectation peak, valuation compression, clearing
Previously storage stocks rose too much, positions were too crowded. When everyone believes the same story, as long as earnings reports don't exceed the wildest whispers, it triggers realization. Earnings trigger declines, leveraged funds, quant trading, and ETF rebalancing amplify the drop.
So the recent plunge does not mean storage fundamentals suddenly collapsed, but three things happened simultaneously: expectation peaked, valuation compressed, crowded trades cleared.
What to really watch next is not how many more chips are short in the news, but four things:
1. Can price increases accelerate again?
2. Can spot and contract prices strengthen in sync?
3. Can smartphone and PC sales bear high costs?
4. After the next super earnings report, will stock prices still "die on sight"?
Remember the last two sentences: a good company is not a good stock, and a perfect earnings report can't save a stock priced at 120 points. The real reversal signal for storage is not another exaggerated shortage news, but that after good news, the stock price finally stops falling. $SKHYNIX $MU $SAMSUNG Palantir's revenue surged 93%, after-hours up 13% — but the more impressive the numbers, the more suspicious I feel
Last night Palantir's earnings report exploded — revenue $1.94 billion, up 93% year-over-year; net profit $1.06 billion, compared to only $329 million last year; adjusted EPS $0.41, market expectation was only $0.35. After-hours it soared more than 13%.
The circle of friends is cheering "AI applications have finally exploded," but the more I look at some numbers, the more something feels off.
Let's start with the brightest spot: U.S. commercial revenue surged 149%.
$764 million, marking the fourth consecutive quarter with growth over 100%. CEO Karp himself said it was "extraordinary" and "unstoppable." The company also raised its full-year revenue guidance sharply from $7.65 billion to $8.15 billion.
Sounds all positive, right?
But I noticed one detail — the company's stock price has dropped 29% this year.
Closing at $125.65 before the earnings report, down nearly 30% since the start of the year. Why? The market has been worried that big model companies like OpenAI and Anthropic would squeeze Palantir out of the market. Can one earnings report reverse all those concerns? I don't believe it.
Now let's look at the quality of growth.
The 149% growth in U.S. commercial revenue is indeed scary, but the base is small. A year ago, U.S. commercial revenue was just over $300 million, now $764 million. Doubling is easy, but what happens after doubling? Karp says "it can last at least 18 more months," which sounds familiar — the last person who said that has been long gone.
More worth pondering is the "sovereign AI" narrative.
Karp talks extensively about the "AI sovereignty revolution" in the shareholder letter, saying customers "refuse to be vassals of AI labs." Translated: customers don't want to feed their data to OpenAI, so they come to Palantir for private deployment.
Does this logic hold? In the short term, yes, some companies fear data leaks. But in the long term, will giants like OpenAI, Google, and Microsoft just stand by? They have plenty of money to invest in R&D and sales. Is Palantir's moat really as deep as Karp claims?
Citigroup said this earnings report "further weakens the bearish logic of intensified AI competition." But I think one earnings report can't change the competitive landscape of an industry.
What really makes me cautious is — the market is using the new story of "AI application explosion" to replace the old story of "AI compute peak." Palantir is up, Microsoft is up, but what about Nvidia? AMD? Capital is shifting from hardware to software. This kind of rotation usually happens when the old story is losing steam and the new story is just starting.
Palantir's earnings are indeed impressive, but the 155% "40 rule" score and 62% adjusted operating margin — these numbers are too perfect, so perfect they remind me of Snowflake in 2021.
I'm not saying Palantir is bad. I'm saying, when a stock has dropped 29% since the start of the year and one earnings report can make it jump 13% after hours, are you buying fundamentals or the emotion of "finally a good news"?
#Palantir营收增93%，盘后涨13%
Purely my personal speculation, not any adviceThe US-Iran situation is seeing a turning point; can the Strait really be reopened?
Starting tonight at 7 PM Beijing time, three sources—the Saudi Arabian TV station, the Qatari Foreign Ministry spokesperson, and US Treasury Secretary Janet Yellen—have verified that arrangements for reopening the Strait of Hormuz will be announced within the next few hours or by tomorrow.
Although the news does not come from official reports by the US, Iran, or Oman, it includes statements from Gulf countries and senior US officials, and cross-verification greatly increases the accuracy of the information.
Next, we will watch for official reports from the US, Iran, and Oman to further confirm the accuracy of the news; for now, ignore Trump's remarks.
Once the new management plan for the Strait is confirmed, the key focus will be on the fee structure. If Iran remains the primary fee collector, the agreement will be difficult for the US to accept. Conversely, if fees are mainly service charges collected jointly, this could be the best co-management fee scheme for the Strait and acceptable to the US.
Once the Strait plan is finalized, it means the US and Iran can return to the negotiating table. The next focus will be on when the US and Iran officially set the negotiation location and time. This trend also suggests that the recent new US-Iran conflict of nearly a month is coming to an end.
Moreover, I boldly believe that this US-Iran peace window plus the resolution of the Strait issue can bring at least a three-month geopolitical optimistic buffer window. At least before the midterm election results are announced, the US and Iran will maintain some restraint, focusing on bilateral technical negotiations.
Of course, don’t expect the final US-Iran negotiations to be resolved in the short term; this is a century-old problem. What we will likely see is a process of negotiation mixed with conflict and tug-of-war.
As for crude oil, taking US oil as an example, once the new Strait management plan is confirmed, watch for prices around 70-75. Once the US and Iran set the negotiation location and time, prices near 65 can be expected. The situation is becoming optimistic step by step. Congratulations to everyone for finally shedding a geopolitical pressure point! #美伊重回谈判桌，油价回吐 [LINK: The real narrative is not about 'old coins,' but about data entry points] Many people, when they mention LINK, think it's a classic project from the previous cycle—not as exciting as MEMEs, nor as storytelling as good at new public chains. But what the market most easily overlooks is often the most critical infrastructure. The crypto world needs to connect to the real world; on-chain contracts need to read prices, interest rates, reserves, settlements, and institutional data. The unavoidable keyword is oracles, and LINK remains the most prominent name in this sector. First, LINK's advantage lies in its strong positioning. It doesn't live on a single graph, a slogan, or a wave of community charges, but stands at the gateway of "on-chain world calling external data." As long as DeFi, RWA, stablecoins, and derivatives still require trustworthy data, oracles will not disappear. When the market is cold, people complain about its slowness; When the market is hot, funds suddenly remember: it turns out infrastructure is the real hydropower, electricity, and coal. Second, LINK's imagination comes from institutionalization. Nowadays, the market increasingly talks about RWA, fund tokenization, on-chain settlement, and traditional financial entry. Here's the question: After institutional assets are listed on-chain, where do price data, risk control data, proof of reserves, and cross-chain messages come from? This is not something that can be solved with a simple "Long live decentralization." If LINK continues to capture the incremental gains from institutional data and cross-chain communications, its valuation logic will be no longer just a rebound in ordinary altcoins, but rather a revaluation of infrastructure. Third, the risks are very real. LINK is not that kind of dayThe damp, hot soil seeped through the Geely suit, my right eye pressed tightly against the eyepiece of the 32x optical scope, breathing slowed to three times per minute, fingertips touching the cold metal trigger.
The first rule of field stealth: never be disturbed by disguised noise, only watch which direction the heaviest cannon on the position is calibrated.
An 8-K battle report was decrypted from frontline encrypted communications—Strategy pulled the shipping trigger between July 27 and August 2, releasing 1,638 BTC, cashing out $104.7 million. The average transaction price was $63,957, this shot directly pierced through their $75,419 cost basis. Total holdings were reduced to 842,138 BTC. Discounted liquidation in their own ammunition depot is not a blind retreat but a position commander forcibly supplementing a high-pressure lifeline for fixed 12% preferred stock dividends and stock buybacks.
Zooming in the scope’s scale, carefully observing this trajectory: compared to the fierce firing from July 1 to 5 that unloaded 3,588 BTC in one go, blasting out $216 million, after a full four weeks of absolute silent stealth, this time the firing volume was cut by more than half. Between August 2 and 3, on-chain radar detected another secret transfer of about 299.843 BTC—that was the sentry silently changing defense, awaiting secondary confirmation of next week’s battle information.
When will this heavy weapon reload and buy ammunition again? The conditions are extremely harsh and cold: only when the preferred stock price, currently discounted nearly 10%, climbs back near the issue price and the logistics supply line is restored will their add-position trigger be unlocked.
At this moment, fine-tune the crosshair to the damage zone of the US stock token-linked target $XCH. The pause and retreat of the main firepower point is generating a sidewind effect. The linkage between the US stock derivative token market and the underlying asset is like a bullet slightly yawing when penetrating air of different densities. The slight tremor of $XCH on the market exposes the real resistance level of the derivative target’s pressure capacity during the tactical convergence period of large funds. Before the heavy artilleryman repositions, $XCH’s fluctuations are the best dashboard to measure the overall liquidity wind speed.
Wind speed 2.4 m/s, humidity 85%, correction angle two clicks left.
Those rookies who frequently expose their positions and blindly pull the trigger without a perfect risk-reward ratio have long been shattered by cross-period arbitrage stray bullets. Without absolute confidence in the odds, my bullet will never be chambered.
The crosshair in my right eye’s scope has already locked onto the next defensive position.
#MSTRSells1638BTC Technical trend breakdown (monthly + daily charts)
1. Monthly Moving Average Level (Medium-term Trend)
The current price is trading above the May monthly moving average of $202.48, with the short-term moving average flattening. The October monthly moving average is about to turn flat, with upward momentum exhausting and entering a high-level box consolidation phase; The monthly MACD turned positive for the first time, and the mid-term uptrend shifted from a one-sided bullish trend to a consolidating market. The major rally has temporarily ended, and the next 3~6 months will mainly focus on range-bound consolidation to digest valuations. The long-term moving averages for 20 and 60 months have continued to rise, and the foundation of the super long-term bull market has not been broken, so there are no conditions for a bear market reversal.
Core box: lower boundary at $190 (strong support), upper boundary at $230 (strong resistance).
2. Short-term daily chart (before financial report, 8.4~8.26)
Support: 205→200→190; Pressure: 215→222→230;
Before the earnings report, expectations are likely to fluctuate sideways in a narrow range of $205~215, while funds are waiting for earnings to materialize.
The two core policy variables (the biggest sources of uncertainty)
1. U.S. chip controls on China (long-term suppression has led to growth)
In June 2026, BIS will block loopholes in overseas subsidiaries' detours in procurement, and Blackwell will completely ban the sale of all high-end chips to China; Nvidia has completely lost the high-end AI computing power market in China, maintaining only fragmented revenue from low-end edge chips;
Potential risks: Further restrictions on Nvidia's chip re-exports to Southeast Asia and the Middle East will indirectly reduce global shipments; Positive news for hedges: Nvidia is increasing its investment in emerging computing power markets in the Middle East, Southeast Asia, Japan, and South Korea to fill gaps.
2. U.S. domestic antitrust and export control spillover risks
The U.S. Congress continues to review NVIDIA's monopoly position in the AI market. If it forcibly splits the hardware-software binding model and forcibly opens up the CUDA ecosystem, it will directly weaken core barriers and put significant pressure on valuations; At present, it remains only at the research stage, with a very low probability of short-term implementation.
3. Geopolitical incidental disturbances
Escalation of Middle East geopolitical tensions will push global risk appetite downward. The Nasdaq is under pressure and has driven Nvidia to passively pull back, which is a short-term sentiment shock and does not change fundamentals.
3. Path to valuation system repair and restructuring
1. Current valuation positioning
TTM PE is 32x, Qianzhen FY2027 PE is about 27~29x, PEG ≈ 0.68, compared to global hardcore tech leaders: higher than Apple and Microsoft, but significantly below the 2024 peak bubble valuation of over 50x;
Historical Pivot of the Chip Industry: Mature chip giants have a PE ratio of 15~20 times, while high-growth AI chips have a reasonable PE of 25~30 times. Current valuations are in a reasonably above range with no severe bubbles, but also not extreme undervaluation.
2. Three-stage valuation restructuring
1) Short-term (1~2 months, financial reporting cycle): Performance validation period
If the performance meets the target, the valuation will fluctuate within the 28~32 times range; Performance fell short of expectations, valuations were passively compressed to 25~27 times, and the stock price fell back to $190.
2) Mid-stage (3~12 months): Growth convergence period
As industry supply eases and revenue growth falls from 80%+ to a steady 40%~50%, the valuation center is gradually being revised downward to 24~28 times; Relying on net profit growth to absorb valuations, while stock prices are slowly rising driven by earnings, the room for valuation cuts is limited.
3) Long-term (1~3 years): Growth logic switch
After AI computing infrastructure saturation, NVIDIA shifted from a high-growth stock to a stable value growth target, with valuations returning to 20~25 times the industry's mature valuations, relying on automotive chips, AI software subscriptions, and industrial AI businesses to open up a second growth curve to support valuations.
3. Core valuation catalysts and conditions triggered by valuation reduction
✅ Valuation upward revision catalyst: Rubin architecture mass production exceeded expectations, major clients added annual orders worth tens of billions, the Federal Reserve began rate cuts, CUDA software subscription revenue surged rapidly;
❌ Valuation decline triggered: U.S. antitrust measures, a sharp reduction in global AI capital spending, and HBM oversupply leading to chip price wars, with gross margins falling below 70% for two consecutive quarters.
4. Stock Price Scenario Forecast by Cycle
1. Short-term (around the financial report from 8.4~8.26)
- Benchmark Trend: 205~215 USD ranges sideways; The earnings report was neutral, with a slight after-hours rally and pullback, closing at $210~$218;
- Optimistic: Earnings exceed expectations + guidance raised to break $220, testing $225~$230;
- Pessimism: Earnings falling short of expectations, quickly pulling back to $200, and extremely testing the $190 support level.
2. Mid-term (3~6 months, until the end of 2026)
Mid-range range: $190~$240
Optimal path: steady performance realization, moderate valuation rise, and a fluctuating upward trend to reach $235~240;
- Weak path: Liquidity fluctuations + cooling industry sentiment, repeatedly oscillating within the $195~$220 range.
3. Long-term (12 months)
The institutional consensus target price is $269, corresponding to a valuation of about 28 times, provided annual revenue growth stays above 40% and gross margin stays above 70%; If growth falls short of expectations, the target price will be revised down to $230~245.
5. Core Risk Summary
1. Policy Risks: U.S. regulations tighten again and antitrust sanctions; China's comprehensive domestic substitution of computing power is a long-term loss of incremental loss;
2. Supply risk: TSMC's CoWoS and HBM capacity are being rapidly released, leading to a reversal in industry supply and demand, and chip price cuts eroding gross margins;
3. Competitive risk: AMD, Broadcom, and Google TPUs are aggressively capturing data center market share;
4. Macro risks: Delayed Fed rate cuts, systemic U.S. stock market pullbacks, global tech companies cutting AI capital spending;
5. Risk of technological iteration: The new computing architecture disrupts the traditional advantages of GPUs.
6. Summary
Nvidia's fundamentals remain the strongest in the global chip industry, with ultra-high profit margins, a monopoly-level ecosystem, and long-term contract orders locking in short-term performance floors, so there is no logic of fundamental collapse; The current stock price of $210 is a phase of volatility and bottoming out under reasonable valuation. The large-scale rally has ended, and future earnings have shifted from earning valuation premiums to earnings from earnings growth.
Operation perspective: Around $190 is an excellent medium- to long-term positioning range; above $230, the cost-performance ratio is low, suitable for taking profits and reducing positions; Before the financial report, avoid heavy positions and gambling; focus on position control and range-bound trading.
The above is merely the fundamental logic analysis of Nvidia by "When Water Overflows," and does not constitute any trading or investment advice. U.S. stocks are highly volatile, and high-valuation tech stocks share the same profit and loss source. Be sure to manage your position risk carefully.Both traditional DRAM and NAND have simultaneously hit multi-year highs. My first reaction is not to chase storage stocks, but to first distinguish: is this a demand surge, or a structural shortage caused by AI squeezing out capacity? HBM and AI servers occupy advanced capacity, Samsung, SK Hynix, and Micron reduce supply of older DDR4 and traditional NAND, and PC manufacturers are pre-stocking, naturally pushing prices higher. This is indeed favorable for Hynix: HBM earns high gross margins, and traditional storage enjoys price increases, effectively profiting from both ends. But stock prices trade on the future. Although SK Hynix's Q2 profits hit a record, they still fell short of market expectations; this year's capital expenditure plan exceeds 40 trillion KRW, and Changxin Memory's capacity expansion also raises supply pressure after 2027. The current contradiction is: spot prices are at cycle highs, yet the market is beginning to worry about the next wave of capacity releases. Previously, a 2x long position on Hynix products dropped 16.65%. I will watch whether DRAM contract prices can rise for two consecutive quarters, rather than just looking at a one-month high. Do you think the storage price increase cycle can continue until 2027, or has the stock price already peaked in advance? $SKHYNIX $MU #韩股重挫5%，存储多空信号对峙 GPUs are still selling shovels, but Palantir has already started collecting taxes
Over the past three years, the world has poured trillions of dollars into the AI field. GPUs have sold out, data centers have sprung up, and tech giants have even started researching nuclear power plants to compete for electricity. But when business owners open their financial reports, AI is indeed very smart, but where exactly is the money being made? OpenAI is still raising funds, Anthropic is still burning cash, cloud providers’ capital expenditures are like bottomless pits, and many companies have bought a bunch of AI tools, with the biggest result being just that employees write weekly reports five minutes faster. Just as everyone began to doubt whether the so-called AI revolution might just be an expensive PPT contest, Palantir suddenly delivered a nearly outrageous financial report.
In Q2 2025, Palantir’s revenue was about $1.94 billion, a 93% year-over-year increase; U.S. commercial revenue surged 149%; the value of new contracts signed in a single quarter was $3.37 billion; free cash flow exceeded $1 billion for the first time; and the full-year revenue guidance was raised by nearly $500 million at once. After hours, the stock price rose more than 12%. The biggest significance of this report is not that Palantir exceeded expectations again, but that the AI industry finally has a real player who turns computing power into cash flow. Ironically, this company neither manufactures GPUs nor trains the strongest models. While OpenAI, Google, and Anthropic are fighting fiercely over whose model is smarter, Palantir stands aside and says: “Keep fighting, it doesn’t matter who wins, because in the end, everyone has to connect to my system.”
This is what makes Palantir so formidable. Nvidia sells computing power, model companies sell intelligence, Palantir sells control. Enterprises certainly want to use AI, but they dare not hand over customer lists, production data, supply chain information, and internal decisions to an external model company, because the stronger the AI, the more afraid enterprises become. They worry about data leaks, loss of control, models spouting nonsense, and most of all, that one day their core knowledge and experience will all become someone else’s training material. So what Palantir sells is not a chatbot, but a set of AI cages: who can see the data, who can call the model, what the model can do, which actions must be manually approved, how to assign responsibility if something goes wrong — all locked inside the enterprise’s own system. Management calls this AI sovereignty. To put it more bluntly, enterprises don’t refuse to use AI, they just don’t want to hand over the reins to AI companies. Palantir’s job is to bring AI through the door for enterprises and then put a collar on it.
This also explains a very exaggerated data point: the number of U.S. commercial customers grew about 35%, but U.S. commercial revenue grew 149%. Customers didn’t grow fourfold, but revenue nearly grew 1.5 times, indicating that the real driver of performance is not constantly acquiring new customers, but that once old customers start using it, they buy more and more. First let AI handle customer service, then manage inventory, analyze data, enter approvals, and finally even directly intervene in decisions in factories, banks, hospitals, and military systems. Palantir is no longer selling software licenses but laying a nervous system inside the enterprise’s veins. Once it’s in, it’s not so easy to pull out.
Even more counterintuitive is that the cheaper, stronger, and more homogeneous models become in the future, the more valuable Palantir may become. Because models will gradually become commodities — today you use OpenAI, tomorrow you can switch to Anthropic, the day after you can connect to open-source models — but the enterprise’s own data, permissions, and business processes won’t be easily replaced. Models are just engines; Palantir controls the steering wheel, brakes, and traffic rules. So the stronger OpenAI gets, it doesn’t necessarily eliminate Palantir. On the contrary, the stronger the model’s capabilities, the more enterprises need a system to prevent it from running wild. This is why Palantir may become the most unique type of company in the AI era: it’s not responsible for creating the smartest AI, but for deciding what these AIs can and cannot do, and ultimately how to make money for clients.
But the problem lies exactly here: Palantir’s business is materializing, but its valuation has already celebrated many years in advance. Based on the current market value and full-year revenue guidance, its expected price-to-sales ratio is close to or even exceeds 40 times, meaning the market expects not just continued excellence but nearly abnormal growth for many consecutive years. As soon as U.S. government orders slow down, enterprise AI investment cools, or OpenAI, Microsoft, and Google start filling in data governance, permission control, and workflow capabilities themselves, this valuation could quickly lose support.
So this financial report does not prove that Palantir is already more profitable than Nvidia; it proves another thing: the first phase of AI is that whoever owns the GPU owns the discourse power; the second phase of AI is that whoever can integrate these expensive GPUs into enterprise processes and truly save money, make money, and make decisions for clients will take the profits. Nvidia is still selling shovels at the mine entrance; Palantir has already entered the mine and started charging tolls based on mining results. And in any gold rush, selling shovels is profitable, but the real windfall often belongs to the one who sets the rules, controls the roads, and charges everyone. $GOOGL The preliminary results of Jieli Technology's lottery are out, witnessing history. Subscription funds amounted to about 1.41 trillion yuan, a historic high and far exceeding expectations. The threshold for 100 shares is about 4.75 million, for 2+1 about 9.6 million, and for 3+1 about 14.25 million. It was a bit close, but I made it ashore safely. Senhe Hi-Tech will go public tomorrow, so let's analyze it. The company's concept is average, but its performance is decent. However, there are many old shares in circulation, totaling 16.9642 million shares, with a market capitalization of 993 million, which is quite substantial. Stir-frying it is quite difficult. The expected increase is 50-100%, corresponding to a price of 43.6-58.1. If it doesn't fall short of expectations, I don't have a broad vision. There will be another subscription for Hengxing Co., Ltd. tomorrow—let's analyze this. Subscription funds are expected to be between 11,500 and 12,000 yuan, which should not exceed expectations. If it's 120,000 yuan, the threshold for 100 shares is about 6.4 million. Let me explain the subscription plan. Give up: For accounts under 6 million, the chances of winning the lottery are slim, so you can give up. 100 shares: 6.5 million is very promising. 2+1: 13 million to get 200 shares is very hopeful; 13.5 million to get 2+1 shares is very promising. Top slot: 15.1165 million is most likely 2+1, so hitting the top slot isn't very meaningful. If you're really anxious, you can consider top slot. I plan to get 13.5 million for a 2+1 ticket. -- Today is a rare broad-based rally. Technology collectively stood up: storage +6%, CPO +6.7%, chips +5.3%. Recently, the tech market has been booming—you can't play without a big heart. Just after leaving the ICU, you're taken to a KTV, and before you even warm up, again...Tonight's market seemed to be paused then suddenly released; ETH released a bunch of news within just over ten minutes, while the price seemed to be waiting for a bigger answer. Have you ever felt for a moment that you're not staring at the candlestick chart, but rather the remote control in the hands of some big shot? To be honest, my impression tonight is: the market isn't without direction; it's that its sense of direction has been temporarily "taken over" by external events. With the US and Iran returning to the negotiating table, oil prices immediately gave back. This should have been a signal of rising risk appetite, but ETH's reaction was somewhat "ambiguous"—no aggressive rally, no panic sell-offs, more like repricing the "uncertainty decline" itself. Here's a point that's easy to overlook: the real impact of a drop in oil prices isn't crude oil itself, but the market's expectations for the inflation path. Once inflation expectations ease, interest rate pressure eases, which is potentially beneficial for long-duration assets (such as high-beta crypto counterfeits and yield-bearing ETH assets). But the question is, has this positive factor already been priced in advance? Judging from ETH's relative strength over the past few days, it seems funds have partially bet on the "geopolitical cooling" scenario. The cross-market linkage logic is as follows: - Event: US-Iran negotiations → oil prices fall → inflation expectations fall → pressure on US Treasury yields eases → risk assets gain breathing room. - But within the crypto market, funds have not become aggressive because of this; instead, it feels more like "taking advantage of the situation to rebalance." BTC remains steady, ETH attempts to catch up, but it's still a hill🔥The harsh truth of the market: those who speak the truth don't last until the bubble bursts.
In the early years, CZ predicted a crypto super cycle in 2026, which seemed exaggerated at the time.
Later, it became clear that as an "ecosystem builder selling shovels," his stance was inherently different.
A quote from Grantham reveals Wall Street's unspoken rules.
1929, 1972, 2000—countless bubbles in history.
Although the vast majority of analysts predicted a bear market, no institution dared to publicly call for an exit.
Why?
If you warn of risks early and the market keeps soaring, you lose your clients.
Grantham’s business was halved because he was bearish too early.
Keynes famously said: it’s better to be wrong with the crowd than right alone.
The same applies to today’s AI market.
Some say AI is an epic bubble; others say computing power demand is endless.
It’s not necessarily about who’s lying.
Many views stem not from a lack of understanding but from differing positions.AMD: Numbers Are Within Expectations, The Real Focus Is on the AI Pace in the Second Half
Consensus roughly locks in revenue at $11.1–11.3 billion (about +47% YoY), adjusted EPS around $1.61. The company's previous guidance midpoint was $11.2 billion ± $300 million, and consensus almost exactly matches the midpoint, indicating the market has already priced in "strength."
A deeper look:
Data center business remains core. This segment clearly accelerated last quarter; the market is now watching the actual shipment pace of Instinct AI accelerators, ASP changes, and whether EPYC server CPUs are accelerating share erosion against Intel.
Whether gross margin can hold steady near 56%. Last quarter showed a good recovery; if maintained, it indicates the product mix is shifting toward higher-value AI.
The guidance for the second half is the real catalyst. The mass production pace of the MI series and Helios platform, along with cloud vendor order visibility, will determine if the market is willing to continue granting AMD a higher valuation premium.
Historical pattern: AMD has often slightly beaten expectations in recent quarters, but the real stock price surge usually happens when management provides clearer, more optimistic statements about AI demand in the second half. If tonight’s report is just "in line with expectations + neutral guidance," the reaction may be muted; stronger AI visibility could trigger a re-rating.
2. SpaceX: The First Ever Public Quarterly Report, Core Conflict Between "Starlink Cash Flow" vs "AI + Starship Burn"
Consensus roughly at $6.8–6.9 billion revenue, loss per share about $0.23–0.26. This is the company’s first formal quarterly report since going public, and the market is highly sensitive to the numbers.
More important when broken down:
Connectivity (Starlink): Currently the only consistently profitable segment. The market watches user growth, whether ARPU stabilizes, and if profit margins can continue to improve. If Starlink’s margin beats expectations, it will significantly ease concerns about overall losses.
AI segment: Revenue expectations have jumped sharply (from about $800 million in Q1 to possibly near or above $2 billion), but capital expenditures are very high. The market’s real concern is whether this AI compute power has already converted into quality, renewable customer revenue, rather than just "building first, then finding customers."
Space segment: Launch services + Starship R&D are still burning cash. Progress on Starship (launch frequency, recovery success rate, commercial payloads) will directly impact the long-term narrative.
A deeper conflict lies in valuation versus reality mismatch. The company remains at a trillion-dollar market cap level and soon faces large-scale lockup expirations. If tonight’s report shows Starlink cash flow is strong enough and AI order visibility improves, the market may temporarily tolerate losses; otherwise, if AI burn rate clearly outpaces revenue realization combined with lockup pressure, stock volatility will significantly increase.
3. Circle: USDC Circulation Is the Real "Fundamental Thermometer"
To be released pre-market tomorrow. Consensus roughly at $710–740 million revenue, EPS $0.16–0.21. Surface numbers show little fluctuation, but underlying metrics are more critical.
Key observation points:
USDC end-of-period circulation: About $77 billion at Q1 end, recently fallen back to around $73 billion. Whether circulation stops declining and if average circulation still grows slightly will directly determine the sustainability of reserve income.
Reserve yield and distribution costs. Changes in interest rate environment + revenue-sharing arrangements with channels like Coinbase will affect actual retained profits.
Progress in non-crypto scenarios. If management can provide concrete data on payments, institutional settlements, on-chain applications, and other "non-speculative demand," sentiment will ease considerably; if still highly dependent on crypto trading activity, the market will reassess growth potential.
The stablecoin business is essentially a "leverage game of interest rates + circulation." If circulation continues to shrink, even if revenue barely meets targets, the market will shift focus to whether growth has peaked. #财报观察员：AMD与SpaceX交卷在即，Circle压轴 Solana is making two moves simultaneously: expanding capacity while accelerating deflation
Two recent Solana proposals are worth looking at together.
The first has already been implemented: on July 29, Solana increased the computational limit per block from 60 million CU to 100 million CU, a 66.7% increase. For ordinary transactions, this means a block can accommodate more parallel tasks. Assuming a popular account uses 12 million CU, its theoretical block share would drop from 20% to 12%, effectively widening overall throughput capacity.
But expansion hasn’t completely solved congestion.
The per-block write limit to the same account remains 12 million CU, and the account data growth limit is still 100MB. In other words, ordinary transfers may be smoother, but high-frequency DEXs, liquidation protocols, and popular tokens may still get stuck on the same account. The road is wider, but the toll booths haven’t increased; during peak times, queues are still inevitable.
The second proposal directly affects $SOL supply. The dual deflation proposal SGP-0002 has entered the support phase, planning to raise the annual deflation rate from 15% to 30%, allowing inflation to drop faster to a 1.5% endpoint. According to the proposal’s calculations, the time to reach terminal inflation will shorten from about 5.7 years to about 2.8 years, reducing approximately 18.9 million $SOL issuance over the next six years.
This can’t be viewed simply as "deflation." Roughly calculated at about $73 per $SOL, 18.9 million corresponds to about $1.38 billion in potential new supply. Reducing issuance helps lower long-term dilution, but validator staking rewards will also decline faster. Network security budgets, node operation revenues, and SOL staking attractiveness all need to be monitored simultaneously.
There’s another detail in the governance data: screenshots show about 27.19 million $SOL supporting votes, with a threshold of 43.27 million. Calculating directly from these, completion is about 62.8%, which doesn’t match the 41.9% reported in the news; the final reference should be the on-chain governance page.
So this round of changes isn’t simply positive. Capacity expansion improves demand-side experience, and accelerated deflation tightens supply-side increments. Both need to work together to potentially strengthen $SOL’s long-term value capture; if popular account bottlenecks persist or validator rewards drop too fast, the market will reprice accordingly.
Next, I’m more focused on four data points: failed transaction rate, median priority fee, DEX peak period account contention, and the actual staking support progress for SGP-0002.
This is only my personal market observation and does not constitute investment advice, DYOR.
$SOL #交易之声：你的经验值得被听到 #From Rate Cuts to Rate Hikes, Fed Disagreements Fully Exposed
According to TradingBeats monitoring, Hyperliquid currently tracks five popular targets: SKHX, MU, SNDK, SPCX, and CXMT. The single address with the highest order amount has placed a total of 6 large orders, with a nominal amount of approximately $30,442,000. The only new position order is a low-level buy order for SNDK. The specific layout is as follows:
SNDK: Currently at $1292, up 3.4% in 24 hours, Q4 and full-year earnings released on August 5
SNDK is currently at $1292, up 3.4% in 24 hours, with a trading volume of about $486 million and an open interest value of about $117 million; the current hourly funding rate is about -0.0013%, with shorts paying longs.
TOP 1 order address is 0x0ad9: currently holding 3890.8 SNDK long positions, valued at about $5.028 million, with an average entry price of $1287 and an unrealized profit of about $20,000.
This address placed a new buy order of $2.5 million at $1187 and a "reduce only" sell order of $5.435 million at $1397, which are 8.1% below and 8.1% above the current price respectively. It is arranging a two-way range trade around the current price before the earnings report.
SPCX: Currently at $115.98, up 6.5% in 24 hours, recently including the first post-IPO earnings report and lock-up expiration
TOP 1 order address 0x0871: long 50,000 SPCX with 20x full margin, position valued at about $5.797 million, average entry price $126.36, unrealized loss about $521,000, liquidation price about $73.5.
This address placed a "reduce only" sell order of $7.65 million at $153, covering the entire position. This price is 31.9% above the current price, 21.1% above cost, and also above the $135 IPO price, still betting on a stronger rebound after earnings or lock-up bearish factors are digested.
SKHX: Currently at $1072, down 3.6% in 24 hours
TOP 1 order address 0x364a: long 5488.2 SKHX with 2x full margin, position valued at about $5.888 million, average entry price $1288.6, unrealized loss about $1.184 million.
This address has placed a "reduce only" sell order at $1305, planning to close the entire position at once, with an order amount of about $7.162 million. This price is 21.7% above the current price but only 1.3% above the average entry price, waiting for SKHX to rebound above cost before exiting.
MU: Currently at $830.5, down 1.3% in 24 hours
TOP 1 order address is also 0x0ad9: long 11,100 MU with 2x full margin, position valued at about $9.186 million, average entry price $856.4, unrealized loss about $284,000.
It has placed a "reduce only" sell order of $4.5 million at $860, planning to sell 5232.6 units, reducing about 47.3% of the position. The order price is only 0.4% above cost, indicating this address is not waiting for MU to return to highs but is preparing to reduce positions near breakeven.
CXMT: Currently at $7.984, up 2.0% in 24 hours
TOP 1 order address 0x934d: long 255,600 CXMT with 4x isolated margin, position valued at about $2.043 million, average entry price $6.30, unrealized profit about $433,000, liquidation price about $4.30.
This address has placed a "reduce only" sell order of $3.195 million at $12.5, covering the entire holding. This price is 56.6% above the current price and 98.5% above the entry cost, planning to take profits by doubling.NVIDIA (NVDA) Comprehensive In-Depth Analysis
Benchmark Date: August 4, 2026, US Eastern Market Open at $210; Next Earnings Report: August 26, 2026, After Market Close for FY2027 Q2
Overall Conclusion Upfront: Currently in a high-level oscillation and bottoming consolidation range, fundamentals are very resilient but valuation lacks momentum for a unilateral surge; earnings report likely to slightly beat expectations, short-term will see oscillation and tactical play, mid-term growth driven by Blackwell architecture ramp-up and HBM shortage maintaining high profit margins; main constraints come from China-US policy decoupling, tightening US stock market liquidity, and marginal easing of peer supply; full-year valuation will gradually converge from extreme premium to a reasonable growth valuation center, stock price will oscillate upward within a range, no unilateral bull run or deep crash expected.
I. Current Market Basic Status ($210)
1. Core Basic Data
- TTM P/E ratio: 32.17x, total market cap about $5.1 trillion, 52-week range $95.04~$236, current price in mid-to-high annual range, about 11% pullback from the yearly high;
- In the past three months, stock price has narrowly declined 2.4%, capital rotated from pure AI mainline to storage, supporting hardware, and other low-level sectors, NVIDIA has entered a phase of stock capital competition;
- Shareholding structure: Institutional holdings 70.8%, Vanguard and BlackRock combined hold over 17%, passive base stable, selling pressure mainly from short-term profit-taking and active fund rebalancing, no systemic dumping basis; Jensen Huang holds only 3.33%, very low pledge and selling pressure.
2. Short-term Market Pricing Logic
The market has already priced in high growth expectations for Q2 earnings, current price $210 prices in: revenue of $91.8 billion, EPS $2.08 consensus; unless earnings report shows better-than-expected order guidance or gross margin rises again, positive news likely to trigger "buy the rumor, sell the fact"; negative scenarios only arise from order slowdown, sharp gross margin decline, or sudden policy tightening.
II. Six Core Dimension Breakdown Analysis
(1) Industry Supply and Demand Fundamentals (Determines Revenue Base)
1. Demand side: Global cloud providers have locked in orders to underpin the base
Microsoft, Google, Meta, and Amazon have locked most Blackwell (B100/B200, GB200) capacity for all of 2026 and first half of 2027; AI agents and trillion-parameter large model iterations continue to drive up computing demand, GPU rental contract prices have risen 40% since October 2025, spot supply remains tight, delivery cycles maintain 36~52 weeks.
Consumer GeForce gaming chip demand is stable, automotive chip business steadily growing, data center business accounts for 88% of total revenue, the absolute core base.
Negative constraints: China market revenue share has shrunk to 9% (from 26.4% pre-regulation), high-end chips cannot be directly sold to China, H200 special chips have almost no domestic procurement, permanently losing a market increment worth hundreds of billions, domestic Ascend computing power continues to erode local demand, this is a long-term growth ceiling constraint.
2. Supply side: Bottleneck shifts from GPU wafers to HBM memory + TSMC CoWoS packaging
TSMC advanced CoWoS packaging capacity is fully booked until mid-2027, HBM3e/HBM4 supplied by Samsung, SK Hynix, and Micron, 2026 full-year capacity already pre-sold; new Rubin architecture yield issues with HBM4 cause 25% annual production cut, supply tightness cannot be fundamentally relieved in 2026, NVIDIA retains continuous pricing power, core guarantee of high gross margin.
Marginal changes: Slight packaging capacity expansion after Q4 2026, supply tightness mildly eases, pricing power slowly weakens, slight downward pressure on gross margin.
(2) Raw Materials and Supply Chain Costs (Directly Affect Gross Margin)
1. Core cost composition
Currently, HBM high-speed memory accounts for over 45% of chip production cost, wafer foundry 30%, packaging, components, and labor combined 25%;
HBM contract prices in 2026 are stable with slight increases, Samsung and SK Hynix have no intention to significantly cut prices, no significant cost reduction benefits from raw materials; TSMC foundry pricing stable, no foundry price hikes.
2. Cost hedging advantages
NVIDIA has long-term agreements with upstream suppliers locking prices, bulk purchasing dilutes unit cost; new Blackwell chips greatly improve single-card computing density, unit computing amortized cost continues to decline; scale effects and integrated hardware-software bundled sales hedge slight raw material price increases, raw materials will not materially impact profits.
(3) Profit Margin Projection: Current Earnings and Mid-to-Long Term Trend
1. Historical Profit Baseline
- FY2026 full-year comprehensive gross margin 71.1%, Q4 single quarter peaked at 75%; FY2027 Q1 gross margin 74.9%, net margin stable between 55%~63%, hardware achieves software-level profitability, core moat very strong.
2. Upcoming Q2 (FY2027 Q2) Earnings Margin Forecast
1) Optimistic scenario (40% probability): Blackwell deliveries exceed expectations, high-end high-margin chip proportion rises, gross margin maintains 74.5%~75.5%, revenue breaks $93 billion, EPS > $2.1, slight gross margin increase, earnings guidance raised, stock price short-term surges to $225~$230;
2) Neutral baseline scenario (50% probability): revenue $91~92 billion, EPS $2.05~2.08, gross margin slightly falls to 73.5%~74.2%; due to passive increase in low-margin supporting servers and mid-range chip shipments, overall profitability stable, post-earnings report high-level oscillation, support at $200~205;
3) Pessimistic scenario (10% probability): major client order guidance conservative, gross margin falls below 73%, short-term capital outflow, stock price tests strong support near $190. In the market, Bitcoin still maintains a volatile trend. After the price dropped yesterday, it rebounded with some demand intervention, but I judge this demand to be temporary and sporadic, unlikely to be sustained. The bigger implication is the linkage with global capital markets. Yesterday, crude oil prices fell sharply, and U.S. stock indices rebounded, driven by capital replenishment following the U.S.-Iran ceasefire, leading Bitcoin to follow this correlated movement.
The minor resistance is around 65,000, with support still above 61,000. The weekly chart is likely to continue narrow-range oscillation, and building a bottom will still take time. Be patient and wait for a decline 📉$SNDK (SanDisk) Market and News Analysis
⚠️ Just casual market talk, not investment advice! Storage is a highly cyclical stock with high volatility, earnings reports approaching, huge single-day swings, strong sector correlation, and wild price surges and drops are normal.
Current Market Situation:
Current price 1375, directly hitting a strong short-term resistance zone. After a brutal drop in July, it has entered an oversold recovery phase, with significantly increased intraday volume—partly short covering, partly earnings speculation.
Highly correlated with Micron and SK Hynix; when SK Hynix moves in the Korean market, Sandisk resonates accordingly. There is a large amount of previous trapped positions above. Now funds are speculating on the August 5 after-market Q4 fiscal earnings, with full long-short divergence. Volatility will further increase approaching earnings. Having reached the resistance zone, a volume-less rally is prone to a quick pullback.
📰 News Breakdown
🟢 Bullish Logic
1. AI inference servers are driving real demand for enterprise SSDs; cloud providers have locked in long-term purchase orders; NAND spot prices remain on the rise; enterprise business proportion continues to increase; product mix is optimized. The market unanimously expects this earnings report’s revenue and gross margin to be very impressive.
2. Multiple top investment banks remain optimistic; Goldman Sachs sets a $2200 target price, bullish on the continuation of the AI storage super cycle. After deep July correction, a large amount of panic selling has been released; with US Treasury yields falling, oversold funds are willing to enter and play the recovery.
3. Inclusion in the Nasdaq 100 index brings passive index fund buying as a base allocation.
4. BiCS next-generation flash technology samples have landed, product competitiveness is strong, and major customer orders are sufficient.
🔴 Core Bearish Factors (Suppressing the Main Market Theme)
1. Collective capacity expansion by major manufacturers is the biggest concern. Samsung and SK Hynix are increasing NAND capital expenditures; the market is pricing in 2027 capacity release expectations, fearing that current ultra-high gross margins will be diluted in the future. This is the root cause of the July plunge. Even if this earnings report looks good, if next year’s guidance is conservative, the stock price is prone to be hammered down.
2. Market expectations are very high, leaving low tolerance for earnings misses. Just meeting revenue and gross margin is not enough; focus is on management’s outlook for next year. Falling short of expectations can easily trigger a sell-off after initial gains.
3. Strong sector correlation risk. SK Hynix Korean stock market, Micron’s market, and NAND spot prices all directly drive Sandisk’s price moves; new Korean stock market regulations tighten leverage, speculative hot money in the sector is retreating, amplifying overall volatility.
4. Cloud providers’ capital expenditures are uncertain; if Microsoft and Google reduce AI capital spending later, storage demand expectations will be directly downgraded.
5. High-valuation growth stocks are very sensitive to changes in US Treasury yields.
Key Price Levels
- First resistance above: 1430-1470, strong short-term barrier; needs volume to hold above for rebound continuation; volume-less rallies are prone to quick pullbacks.
- Strong resistance: 1540-1580, only after breaking 1470 is there a chance to test previous highs.
- Short-term lifeline support: 1290-1320, core defense zone for this rebound; holding this maintains box range oscillation for speculation.
- If it decisively breaks below 1230, this recovery rally fails and will deeply retest lower levels again.
Long-Short Realistic Logic
✅ Bulls: AI enterprise SSD is a rigid demand; NAND price increase; deep oversold in July with short covering; institutions optimistic on earnings; Nasdaq component stock brings allocation funds.
❌ Bears: Market worries about long-term capacity expansion squeezing margins; earnings expectations are maxed out with low tolerance; heavy trapped positions above; storage sector sentiment is highly volatile.
Three Scenario Simulations
1️⃣ Optimistic: Hold 1290-1320 support; earnings revenue, gross margin, and next year’s guidance all significantly beat expectations; storage sector collectively strengthens; volume increases and holds above 1470; challenges above 1540; strictly avoid chasing highs; earnings-driven spike, not suitable for chasing.
2️⃣ Base Case (Most Probable): Large 1230-1470 box range with intense oscillation. Earnings meet expectations but next year’s guidance is neutral; profits are taken in batches after good news; long and short forces tug back and forth; waiting for NAND spot prices to guide direction.
3️⃣ Pessimistic: Effectively breaks below 1290 support; earnings or new fiscal year guidance disappoints; combined with storage sector sentiment collapse; further deep retest below 1230.
Harsh Truth of Storage Cyclical Stocks: The market now trades not on how much was earned in the past but whether high gross margins can be maintained next year. Good earnings but conservative guidance still leads to big drops.
Current Positioning: Short-term defense reference near 1290; if it hits 1430-1470 and fails to break through, prioritize partial selling to avoid earnings black swan risk.
No Position: Prefer to watch; if testing and speculating, participate with small positions, strictly use stop loss, and prepare mentally for huge single-day volatility on earnings day.
Focus on Four Key Items: Next year’s gross margin & capital expenditure guidance in the August 5 after-market earnings, NAND flash spot prices, SK Hynix market, 1290-1320 support status, and US Treasury yield changes.
#从降息到加息，联储分歧全公开 The meaning of Gate is: At the same time as we paid 100,000 USDT and 800,000 ALD to the "scammer's" wallet according to the contract, Gate's alpha automatically grabbed the ALD tokens, then it couldn't be disclosed who connected the token listing process, and finally the scammer's wallet transferred the tokens into Gate alpha for an airdrop. Is that correct?
The hash is here, the answer is here
When a project has paid, got listed, and then is told "the person communicating with you is not our staff, and the project is listed on Gate" — this is already a credibility issue for Gate.Before tomorrow's morning session, Circle will report again. The three companies are respectively positioned in the three currently most sensitive tracks: AI computing power, space + AI infrastructure, and stablecoins. Market sentiment from tonight to tomorrow will be quite direct.
1. AMD: Q2 figures are basically "priced in," focus on guidance and data center details
AMD confirmed it will release its fiscal 2026 Q2 earnings after market close on August 4 Eastern Time, with a conference call at 5:00 p.m. ET.
Wall Street consensus roughly expects:
Revenue around $11.1–11.3 billion (about +47% YoY)
Adjusted EPS about $1.61–1.62 (over 230% YoY surge)
Last quarter, the data center business was already strong; this time the market is more concerned about:
Actual shipment pace and ASP of Instinct AI accelerators
Whether EPYC server CPU market share continues to erode Intel
Whether next quarter and full-year guidance will be raised (especially the AI GPU segment)
Whether gross margin can hold steady around 56%
Current stock price fluctuates between $480–500. If results just "meet expectations," the reaction may be muted; what can truly drive the stock price is management's comments on the MI series and Helios platform for the second half.
2. SpaceX: First-ever public quarterly report, can Starlink continue to "support" AI and Starship
SpaceX (SPCX) will release its first quarterly report as a public company after market close today.
Consensus roughly:
Revenue about $6.8–6.9 billion
Loss per share about $0.23–0.26
More importantly when broken down:
Connectivity (mainly Starlink): still the only consistently profitable segment; market watches user growth, ARPU, and profit margin
AI segment: revenue expectations sharply increased, but capital expenditure is very high, losses continue to widen
Space segment: launch services + Starship R&D still burning cash
Company valuation remains at the trillion-dollar level and soon faces a large-scale lockup expiration. The core of this report is not "how much loss," but whether Starlink's cash flow can support the long-term narrative of AI computing power and Starship. If Connectivity's profit margin exceeds expectations or management provides a clearer path, sentiment will improve noticeably; otherwise, market tolerance for "burn rate" will quickly decline.
3. Circle: Grand finale, USDC circulation is the real barometer
Circle (CRCL) will release Q2 earnings before market open on August 5, with a live broadcast at 8:00 a.m. ET.
Consensus roughly:
Revenue about $710–740 million
EPS about $0.16–0.21
Key observation points are only one and a half:
USDC end-of-period circulation (about $77 billion at Q1 end, recently fallen back to around $73 billion)
Average circulation, reserve yield, distribution costs (especially the revenue-sharing arrangement with Coinbase)
The profitability of the stablecoin business highly depends on interest rate environment and circulation volume. If circulation continues to shrink, even if revenue barely meets expectations, the market will focus on "whether growth has peaked." Conversely, if management can provide progress on USDC in non-crypto scenarios (payments, institutional settlement, on-chain applications), sentiment will be much more relaxed.
Tonight: AMD and SpaceX release simultaneously, AI computing power + space narrative resonance, volatility will be relatively high.
Tomorrow: Circle, as the representative of the stablecoin track, its results will directly affect sentiment in crypto-related sectors.#财报观察员：AMD与SpaceX交卷在即，Circle压轴 ISM manufacturing hits a four-year high, US Treasury yields fall, is the market betting on a rate cut?
An interesting signal appeared in the global market today.
The US ISM manufacturing index hit a nearly four-year high, indicating that the US economy is not as weak as the market previously feared.
At the same time, US Treasury yields have fallen.
This deviates somewhat from traditional logic.
Normally, the stronger the economic data, the more the market worries about inflation, the less room the Federal Reserve has to cut rates, and US Treasury yields should rise.
But this time, funds chose to buy US Treasuries.
What the market is really trading may not be the current economy, but future monetary policy.
The ISM manufacturing index has long been regarded as an important indicator of the US economy.
Previously, the market was worried that high interest rates would suppress business activity and might even lead to a rapid economic cooldown.
But the latest data shows that US manufacturing is recovering.
Corporate production is improving.
Market demand is rebounding.
The economy remains resilient.
This means the US economy has not entered a clear recession phase.
But problems have also emerged:
With such a strong economy, why would the Federal Reserve cut rates?
This is currently the biggest divergence in the market.
Some investors believe:
The economy remains stable, inflation continues to decline, and the Federal Reserve still has opportunities to gradually ease policy in the future.
Others believe:
An overheating economy may cause inflation to rebound, and the high interest rate environment may last longer.
The drop in US Treasury yields indicates the market currently leans toward trading future easing expectations.
Funds are positioning in advance:
If rates fall in the future, bond prices will rise, and risk assets may also gain new liquidity support.
For the crypto market, liquidity has always been a core variable.
Past BTC cycles have been highly correlated with changes in the global funding environment.
When the market expects rate cuts:
Dollar liquidity improves.
Funding costs decrease.
Investors are more willing to allocate to high-risk assets.
This is also why institutional funds began to continuously focus on BTC after Bitcoin ETFs were approved.
Currently, Bitcoin price is fluctuating around $62,000.
After BTC failed to break $65,000 previously, it entered a correction phase.
In the short term, bulls and bears are waiting for new catalysts.
On one hand, ETF funds and institutional allocations provide long-term support.
On the other hand, macro policy uncertainty limits upside potential.
BTC currently focuses on:
Support below:
$60,000-$62,000 support zone.
Resistance above:
$65,000 resistance level.
If US Treasury yields continue to fall and market risk appetite improves, BTC may challenge resistance again.
Ethereum is currently priced around $1,850.
ETH has recently underperformed BTC.
The reason is not that the ecosystem lost value, but that funds are more cautious.
RWA, stablecoins, DeFi, and Layer 2 remain important directions for Ethereum's long-term development.
But short-term performance still requires a supportive liquidity environment.
ETH currently focuses on:
$1,800 support.
SOL is currently fluctuating around $70.
Compared to BTC and ETH, SOL is more sensitive to market sentiment.
Over the past year, Solana has attracted a lot of funds with its Meme ecosystem, low fees, and high transaction activity.
But the characteristic of highly elastic assets is:
They rise more when the market is optimistic.
They adjust faster when the market is cautious.
If funds return to risk assets in the future, SOL may still become a rotation target.
This divergence between ISM manufacturing data and US Treasury yields sends an important signal:
The market is shifting focus from "how is the economy now" to "will future funding become more accommodative."
For the crypto market, what really matters next is not a single piece of news.
But:
Federal Reserve policy direction.
Dollar liquidity changes.
Global risk appetite.
BTC focuses on $62,000 support.
ETH focuses on $1,800 defense.
SOL focuses on the $70 area.
If liquidity expectations continue to improve, the crypto market may welcome new funding opportunities.
But if inflation heats up again and the Federal Reserve maintains high interest rates, risk assets will still face pressure.
In the next phase, the market competition is not about whose story is bigger.
But about who can attract real capital inflows. $BTC #ISM创四年新高，美债收益率反跌 Solana's tokenomics could be heading for a major shift.
Two governance proposals are now under discussion that may significantly reduce future $SOL issuance while increasing the amount permanently removed from circulation. Recent discussions around SIMD-0550 have also accelerated the debate over Solana's long-term inflation model.
To make it easier to understand, I built a simple simulator where you can instantly see:
• How much future SOL issuance could be avoided
• The projected circulating supply by 2032 compared with the current schedule
• The estimated point where daily token burns begin to offset a much larger share of new issuance
At the moment, the network creates roughly 61,800 SOL per day while burning around 1,700 SOL.
$SOL
#FedSplitGoesPublic #韩国杠杆ETF成交额降九成，波幅收窄
After South Korean regulators raised the cash margin threshold for leveraged ETFs, the effect was immediate: the trading volume of individual stock leveraged ETFs shrank by 90%, and extreme market volatility with sharp rises and falls noticeably narrowed. Previously, the wild swings in the Korean stock market were largely driven by retail investors speculating with leveraged ETFs—buying passively on the rise and selling passively on the fall, amplifying volatility infinitely. Now, with higher entry barriers, ordinary retail investors are kept out, speculative players have largely exited, and the KOSPI has returned from swings of over ten points to a relatively stable state.
Many think this is just about the local Korean stock market and unrelated to Bitcoin, but here are two layers of connection explained.
First layer: a barometer of global risk speculation sentiment.
Korean retail investors are a significant speculative force in global high-risk assets, trading both local leveraged ETFs and heavily participating in the crypto market. Regulatory crackdowns on high-leverage tools indicate overseas efforts to cool down retail speculation. The signal conveyed: the environment for high-leverage gambling is contracting, and overall risk appetite will trend more conservative. This is unfavorable for frenzied speculation in highly volatile assets in the short term, making violent continuous rallies in Bitcoin unlikely.
Second layer: semiconductor sector linkage transmission.
The core of the Korean stock market is Samsung and SK Hynix semiconductors, and the memory chip market is deeply tied to crypto mining power and the AI sector. After the leveraged tools retreat, semiconductor stock prices will no longer experience huge single-day swings, and the volatility of related crypto tokens tied to mining power and AI will also be suppressed. Without external catalysts causing sharp rises and falls, Bitcoin and altcoins will mostly move sideways within ranges.
However, a caution from me:
A plunge in trading volume ≠ all funds leaving the market. It just means no new positions are allowed; existing shares remain in the market. The hidden risk of forced selling pressure from leveraged ETFs rebalancing during big drops has not completely disappeared, only weakened. The risk cannot be ignored entirely.
On the practical Bitcoin front, my personal view:
1. This event will not directly change BTC’s long-term trend; it is an indirect emotional influence, not a decisive factor.
2. The global contraction of speculative leverage means don’t expect another crazy bull run in the short term; volatility will become the norm, and the cost-effectiveness of chasing highs will continue to decline.
3. Key to watch next: after Korean stocks stabilize, will foreign capital flow back into equities or continue to flee? If foreign capital keeps returning to risk assets, it will indirectly support the crypto market; if capital continues to seek safety, resistance above Bitcoin will increase further.