Orbit Post Sitemap
The SEC is meeting on Friday to discuss proposing Crypto regulation.
This means opening a customized issuance channel for certain investment contracts involving crypto assets. Projects can raise funds according to regulations without having to rigidly follow the full securities registration process, and when the team no longer manages deeply later, there is an exit path.
The Senate pushed the CLARITY bill to September 15 before referencing, temporarily stalling congressional legislation. Regulators stopped waiting and acted first.
This doesn't solve all the problems, but at least it gives primary market issuance a bit more certainty. $BTC is still hovering around 63,000, and this policy move is more worth watching than short-term prices.
For reference only, not investment advice.Safe-haven discussions about gold rose by 457%, while BTC returned to around $63,818, and ETH was around $1,885. The market is showing two directions simultaneously: traditional safe-haven demand is increasing, while crypto still lacks clear support. The divergence between the two means BTC cannot now be judged solely by the "digital gold" narrative. I will verify three points: whether BTC can hold above 63,800 and reclaim 64,000; whether ETH can hold above 1,875; and whether spot trading volume in the crypto market will increase if gold continues to strengthen. If prices and transactions remain out of sync, the divergence of funds will continue. Do you think safe-haven funds will eventually flow into BTC, or will they remain in gold? $BTC $ETH 🚨 CPI could be the market’s next big reset — and crypto traders should be paying attention.
One inflation number could completely change the Fed narrative.
The U.S. July CPI report drops today at 8:30 AM ET, and after weaker-than-expected jobs data, markets have already started leaning toward a more dovish Fed.
Now CPI has to confirm that story.
Economists expect Headline CPI around 3.4% YoY and Core CPI near 2.5%. If inflation comes in softer than expected, the market could quickly price in a more accommodative Fed.
That could mean: 📉 Treasury yields
📉 U.S. dollar
📈 Risk appetite
📈 BTC & ETH
📈 Potentially stronger flows into quality altcoins like $SOL, $BNB and $OKB
But there’s another side.
If CPI comes in hotter than expected, the “higher for longer” narrative could come roaring back. Yields and the dollar could rise, putting pressure on equities and crypto.
And we all know how quickly crypto can react when macro expectations change.
So today isn't just about the CPI number. It's about what that number does to Fed expectations.
The real question is:
Will CPI confirm the dovish narrative—or completely reset it? 👀
I’ll be watching yields, DXY, BTC reaction, and ETF flows more closely than the headline number itself.
One report could set the tone for the next major move.
#CPIToResetFedBets
#BTCETHETFFlowsDiverge
#AIInfraFundingDiverges
$BTC $ETH
#DailyOrbit Shein has finally reached the threshold of its Hong Kong IPO, but what it brings is not a sequel to the billion-yuan valuation legend, but a clearly discounted new price tag.
Valuation at $98.2 billion in 2022, now reportedly seeking $30 to $40 billion. The company still has scale and brand, with growth, profits, tariffs, and regulatory costs all rewriting valuations.
The most interesting thing about these IPOs is not how lively the first day is, but what capital is willing to pay for a company that once grew rapidly but now faces real costs.#Senate delays clarity bill vote until September, turning a clear positive into a black swan: 14% chance of passing, why should I pay for you?
The bill has been postponed. No vote in August, earliest on September 15.
Polymarket's passing probability dropped from 70% to 14%, TD Cowen directly says there's a 75% chance it won't pass.
Even more painful is the data—XRP ETF had a net inflow of 130 million in May, but only 27 million left in July, a 79% drop. Money is fleeing faster than anyone.
Think about it—
Where exactly is this bill stuck?
Democrats are holding firm on official stock ownership restrictions. The plan is: if holdings exceed 1 million or 10%, you have to sell. Sounds reasonable? But enforcement is entirely handed over to the Department of Justice, effectively letting "insiders investigate insiders." Who would accept these conditions?
Even harsher, even if it gets 60 votes to pass, there are still many amendments waiting. The time window after the September session is pitifully short, and once election season starts, the bill will be cannon fodder.
Back to the market moves—
BTC: 64,000 lost. The positive catalyst is gone, CPI data still looming. Institutions are buying gold as a safe haven; GLD saw $1.4 billion inflow in the first week of August. Cryptocurrency? Figure it out yourself.
Gold: Already surged to a 9-week high of 4435.
Time for a soul-searching question—
A "positive" with only a 14% chance of passing, and the coin has already dropped 25%, is betting on it worth the numbers in your account? Is the nonsense of "regulation is coming" really worth that much real money premium?
The market now faces not the expectation of legislation passing, but the chain reaction of that expectation failing. 14% chance, want to go long BTC on this? $BTC $ETH $XAU 🧵 Something interesting is happening in crypto ETF flows — BTC and ETH are starting to tell two very different stories. 👀
At first glance, the market looks strong.
But look a little closer, and the money is starting to move differently.
$BTC: Spot ETFs have maintained a strong inflow streak throughout August, with no single-day net outflows so far. On August 3 alone, BlackRock bought $111M, Fidelity added $33M, and Franklin Templeton returned with a $9M purchase after more than 30 days.
$ETH: Completely different picture.
ETH ETFs saw $12.3M in single-day outflows and $30.4M in 7-day outflows during the same period.
And then there's something even more interesting. 👇
Italy's largest bank, Intesa Sanpaolo, reportedly cut its BlackRock IBIT position by 94% in Q2 while tripling its ETH ETF exposure.
That doesn't necessarily mean institutions are abandoning BTC.
It could simply be capital rotating from BTC into ETH.
And that's why I'm watching this closely.
When ETF flows between the two largest crypto assets start diverging, it can be an early sign that institutional positioning is changing.
The big question now:
Is this the beginning of a BTC → ETH rotation, or just temporary portfolio rebalancing?
Either way, the next few weeks could tell us a lot. 👀
Personal analysis, not financial advice.
#BTC #ETH #ETF #Crypto #InstitutionalInvestors #BTCETHETFFlowsDiverge
#DailyOrbit Friends, I found a quasi-demon coin: HOLO coin.
I examined its complete life cycle structure and other data and discovered a gene highly similar to TUT.
1. Complete Lifecycle: 270 days of extended sideways trading power—perfect!
HOLO was launched on September 11, 2025. After speculation on the listing, it plummeted from 0.86 to 0.043, then entered a bottoming consolidation period lasting nine months.
From November 2025 to July 2026, prices will fluctuate narrowly within the 0.047-0.092 range, with monthly trading volume drying up from 3057m to 126m, less than 1/20th of the IPO level. The land volume trades sideways, with chips fully concentrated in the hands of the dealer.
In August 2026, with increased volume breaking through 0.099, a signal for initiation appeared.
This structure is almost identical to before TUT started. TUT has been sideways for 300 days, HOLO has been sideways for 270 days, both belonging to ultra-long sideways movement accumulation (> 200 days). We have backtested, and an ultra-long sideways movement is the most fully charged sign of strong players, with the strongest explosive power after a breakout. Summary: 270 Tiandi Volume is trading sideways, the chips have been completely washed out.
2. Chip structure: More concentrated than TUT, terrifying!
The top 10 addresses controlled 96.7%, the largest address 53%, and TGE circulation rate was only 16.96%. Before TUT launched, the top 10 controlled the market at 91.2%, the largest at 22%, and HOLO's concentration even surpassed TUT.
Binance Futures offers 75x leverage, extremely low circulation + high leverage, with complete conditions for pulling the market. Summary: Counting📊 $ETH Contract Liquidation Express (August 12)
According to liquidation data, ETH shows a pattern of short-term bear crushing and medium- to long-term long positions sharply expanding long liquidations, marking a fierce shift in direction:
· Short Cycle (1H/4H): 1-hour short liquidation $308,200, long position only $94.08, short crush long position 3,276 times, short squeeze intensity at nuclear explosion level, extremely rare; 4-hour short $413,300, long $50,000, short is 8.27 times longer than long position. Short-term short-chasers were targeted and blown up, with extremely intense short squeezes.
· Medium- to long-term cycle (12H/24H): 12-hour long liquidation $20.8065 million, short $6.6639 million, bulls are 3.12 times bearish, completely reversing direction; 24-hour long $22.88 million, short $10.1268 million, bulls are 2.26 times bears. The scale of medium and long-term long position liquidations on long positions has expanded exponentially, with cumulative liquidations in 24 hours exceeding $33.0106 million, with long positions accounting for nearly 70%. Bulls are in a bloody stream, and the bullish selling momentum is unstoppable.
· Total liquidations exceeded $33.01 million**, ranking second in volume among cryptocurrencies covered today, just behind BTC, with long positions liquidated at $22.88 million, accounting for over 69%.
⚠️ Risk warning: ETH has a 1-hour short squeeze intensity as high as 3276 times, extremely rare; The medium- to long-term cycle direction reverses, with bears controlling the market and a very high risk of both bulls and bears being killed. Leverage is recommended to be compressed to within 3 times; do not chase rallies or sell on declines. Strictly control positions and wait for a clear direction.
🔥 Market Barometer | August 12
Today's three hot topics point to the same theme: the market is fully transitioning from "storytelling" to a stage of "handing in the answer sheet"—the capital feast of AI infrastructure is entering its first round of return validation.
🏗️ Cloud vendors submit financial reports: AI investment enters return validation phase
During the Q2 earnings season, the four major cloud providers delivered their first "report card" on AI investments. Amazon AWS revenue reached $42.2 billion, +37% year-over-year, marking the fastest growth in 18 quarters; Microsoft Azure rose +43% year-on-year, with full-year Azure revenue surpassing $100 billion for the first time; Google Cloud's revenue was $24.8 billion, an 82% year-on-year increase. The combined cloud business revenue of the three companies was approximately $116.2 billion, a year-on-year increase of about 43%.
More importantly, order reserves are key. AWS order backlog reached $496 billion, a triple-digit year-over-year increase; Google Cloud backlog of $514 billion; Microsoft's commercial RPO increased 84% year-over-year to $678 billion—the visibility of future revenue is improving.
But the cost is just as real. Amazon's free cash flow over the past 12 months turned from positive $18.2 billion to negative $7.6 billion; Google's free cash flow is under short-term pressure. The combined quarterly capital expenditure of the four companies has soared to $151.4 billion.
The market is voting with its feet: rewarding companies that can turn hash power into real cloud revenue, and punishing narratives that only invest without return.
📊 CPI released tonight: The scale for a rate hike in September hangs in the balance
At 20:30 Beijing time on August 12, the US July CPI will be released. The market expects overall CPI year-on-year to fall from 3.5% to 3.4%. Before the data release, CME data showed the probability of a rate hike in September remained at 51.2%.
Deutsche Bank expects the CPI to be 0.15% month-on-month, and the core CPI to be 0.26% month-on-month. The Cleveland Fed forecasts a slight month-on-month increase of 0.09% in overall CPI and 0.21% in core CPI. If tonight's data exceeds expectations, the hawkish camp will rapidly expand; If moderate, rate hike expectations may fade further.
💰 Nvidia's $500 billion vs. Intel's $20 billion: divergence between two paths
On August 10, two chip giants simultaneously announced financing plans.
NVIDIA has partnered with six institutions—including Apollo, BlackRock, BlackRock, Goldman Sachs, and KKR—to establish an independent computing power financing platform, aiming to leverage over $500 billion in third-party capital. Jensen Huang stated: "Computing power has now become infrastructure like electricity and the internet. "The essence is to turn GPUs from consumables into financing infrastructure assets. After the announcement, Nvidia's stock closed down 2.86%.
Intel announced a $20 billion issuance of common stock, marking the largest single equity financing since its IPO in 1971. On the day of the announcement, the stock price closed down 4.06%. Both paths point to the same conclusion: the competition in AI chips has escalated from a technological race to a capital race.
💎 Summary
Cloud vendors proved the real AI demand with 43% revenue growth, but the $151.4 billion quarterly capital expenditure also reminded the market that the pace of burning cash has never slowed; Every basis point in tonight's CPI could determine which way the scales tip for a rate hike in September; Meanwhile, the 500 billion and 20 billion yuan financing plans announced by Nvidia and Intel on the same day mark the official entry into a new stage of "capital-intensive" AI competition. When industry logic, macro narrative, and capital strategy converge on the same day, the market is moving from "storytelling" to a stage of "delivering the answer sheet." #今晚CPI公布, will the pricing for a rate hike in September be rewritten?
#财报观察员: AI infrastructure earnings report debuts one after another
#黄金站上4400美元, demand for risk avoidance is heating up #财报观察员: AI infrastructure earnings report debuts one after another
In the past, when talking about AI, everyone would compete over who was smarter in the model; Now, the topic is gradually shifting: who can find enough cheap and long-term funding for data centers?
#海力士推进NAND扩产, storage supply expectations are rising
$NVDA Nvidia, together with several major financial institutions, is driving over $500 billion in computing power financing plans, indicating that AI has moved from "software imagination" to a heavy-asset stage.
Chip supply is only the first hurdle; beyond that come power supply, construction cycles, utilization rates, and debt repayment pressure. The story of computing power hasn't disappeared; it's just that the bill has finally caught up.CPI hits tonight. $BTC Not a news preview, but a countdown. At 8:30 a.m. U.S. time today (20:30 Beijing time tonight), the July CPI was officially released. The expected year-on-year growth is about 3.4%, but last month it just hit 3.5%. Month-on-month and core CPI—every decimal point will directly rewrite the pricing of US stocks, futures, and crypto markets from tonight to tomorrow. There is only one main theme: whether the inflation figures are still on a "controlled decline" trajectory. If the numbers are low, the market will rush toward rate cuts and a rebound in risk appetite; If the numbers are high or core issues are sticky, interest rate hike expectations, duration pressure, and growth stocks and cryptocurrencies will all be hit hard. So on the same topic, you'll see stock index futures, Nasdaq, gold, Bitcoin, and the US dollar index all named — they're not all stewed together, but different tentacles of the same interest rate expectation. When CPI sounds, interest rate expectations move first; When interest rates move, valuations and leveraged assets move accordingly. Who is benefiting? The first to benefit are those who short volatility and go long on the "soft landing scenario." Funds whose rate cut trades haven't fully died out, leveraged long positions, and duration investors betting on falling U.S. Treasury yields. If the numbers are significantly lower than expected, Nasdaq components, growth stocks, and high-valuation tech stocks are the most elastic. Who is under pressure? When interest rate hike trading surges, growth stocks, small and mid-cap stocks, high-valuation, and financing-driven expansion companies are the first to suffer. In the stronger dollar range, emerging market assets and risk-appetite assets also struggled. If energy is the main culprit driving up inflation, that's another story—but core CPI is what truly determines the Fed's stance. Who is the most anxious?#财报观察员: AI infrastructure earnings report debuts one after another
Currently, two major market themes are brewing simultaneously: on one hand, the SpaceX sales restriction window continues to advance; on the other, AI infrastructure companies are intensively disclosing financial reports, becoming the core barometer for short-term asset pricing.
After the first batch of restricted shares was unlocked, SpaceX's stock price briefly regained its IPO offering price, temporarily stabilizing market sentiment. However, risks have not yet been lifted. The next batch of restricted shares accounting for about 7% will be unlocked on August 20, and potential selling pressure will still require ongoing monitoring.
Meanwhile, the AI industry chain has been released in a flurry of financial reports, with impressive data confirming the continued explosive demand for computing power.
Lumentum released its FY2026 Q4 financial report, with revenue of $1.01 billion, a year-on-year surge of 109.3%. Adjusted EPS reached $3.23, with two core indicators exceeding market expectations across the board.
CoreWeave announced its Q2 results, with revenue of $2.58 billion, up 112% year-on-year, and backlog of orders climbing to $104 billion; At the same time, the company estimates full-year capital expenditure for 2026 to reach $35 billion to $39 billion.
Next, companies such as Coherent, Applied Materials, and Cisco will successively disclose their earnings. The market is testing a core question: can the booming demand for AI computing power, optical communications, and semiconductor equipment truly translate into sustained revenue and profit?
At the same time, there is a contradiction that cannot be ignored: the industry chain continues to spend heavily on capital, which in the long run is likely to keep suppressing the sector's valuation ceiling.
Sector differentiation will intensify going forward. Stocks with strong earnings reports that deliver strong expectations and healthy capital expenditure structures are more likely to emerge independently.#海力士推进NAND扩产, storage supply expectations are rising
📉 Storage Sector Undergoes a Shift Inventory: SK Hynix Expands Production and Implements, How Much Longer Can the Boom Cycle Last?
Today, I came across some new developments in Hynix's Dalian Phase II expansion, so I'd like to share a few personal thoughts.
Key point: The Dalian plant plans to start equipment in the second half of 2026, and add 30,000 to 50,000 units per month of NAND capacity in the first half of 2027. This scale is not small, and note, this time both China + South Korea are expanding simultaneously, not a single factory filling a weak spot.
In the short term, AI data centers are indeed using up enterprise SSD inventory, NAND supply is tight, and prices are supported. But the market is not currently speculating about "current shortages," but "whether demand can meet supply in 2027." There is a lag in expanding production to mass production, and capital market pricing often reflects expectations 6-12 months in advance. In other words, the second half of next year may begin as early as the supply and demand turning point.
The market also confirmed this dilemma—today, the tech sector opened lower, following a typical pattern of structural differentiation. Funds were not blindly rushing in, but rather controlling their positions. SKHY Hynix's US stock XSKHY rose nearly 3%, but A-share related storage stocks lacked follow-up gains, indicating that domestic investors are more sensitive to the pace of expansion and the risk of price wars.
My judgment:
· In the short term (to 2026H1): demand for enterprise-grade SSDs remains high, manufacturers are eager to raise prices, and leading storage companies are confident in their performance.
· Mid-term (2026H2-2027): Expansion of production + uncertainty in consumption-side recovery, supply and demand may shift from "tight balance" to "loose balance," price elasticity declines.
· Long-term: AI storage demand is a structural increment, but NAND itself is highly homogeneous. If expansion cycles resonate, the price war history will repeat.
In terms of operations, I focus on two areas: first, stocks with a high proportion of enterprise-grade SSDs (less affected by expansion), and second, monitoring whether storage prices show signs of loosening around Q4 2026. Currently, it's not advisable to push the market trend linearly; it's safer to go all-in while walking and watching.#财报观察员: AI infrastructure earnings report debuts in succession. AI infrastructure-related companies are intensively releasing financial reports, with the core signal being a sustained explosion in computing power demand, capital expenditure guidance for hyperscale cloud providers being revised upward, and supply chain shipments and orders remaining highly prosperous. AI infrastructure remains one of the most certain technological themes in the coming years, but the pace will shift from highly elastic growth to a phase of "validation delivery + margin improvement."Fun fact: An IPO oversubscription of 8,000 times does not mean the company's value increased 8,000 times overnight; it just means too many people are simultaneously buying the same ticket.
This time, Unitree has directly pushed the chatbot popularity onto subscription data. But the more exaggerated the hype, the more the discussion should shift from "can you grab it?" to "how much future the market has already paid in advance."
Technical stories can be long, but valuations ultimately come down to commercialization, profit, and R&D efficiency. The subscription hall is noisy, and the ledger usually doesn't speak.📊 $DOGE Contract Overload Express (August 12)
According to liquidation data, DOGE shows a pattern of short-term bulls crushing and medium- to long-term short liquidations leading to overtakes, with a clear pattern of both bulls and bears being killed:
· Short-term (1H/4H): 1-hour long liquidation $2,186.62, short only $23.12, bulls crushing bears 94.6 times, with strong bull selling; 4-hour long $32,000, short at $356.75, bulls crushing short 89.7 times. Short-term bulls targeted blowdown, with intense long selling.
· Medium- to long-term cycle (12H/24H): 12-hour long liquidation $736,500, short $1,146,800, short overtaking long position by 1.56 times, completely reversing direction; 24-hour long $784,700, short at $1,512,800, bears at 1.93 times bulls. Medium- to long-term short liquidation scale continues to expand, with short squeeze dominating.
· Total liquidations surpassed $2.2975 million**, with short positions liquidated at $1.5128 million, accounting for nearly 66%. Short sellers were bleeding like rivers, and the short squeeze was unstoppable**.
⚠️ Risk warning: DOGE's short-term long sell-off is nearly 90 times stronger, with a mid- to long-term trend reversal, short liquidations and overtaking occurrence, extremely high risk of both long and short sells. Leverage is recommended to be compressed to within 3x; do not chase rallies or cut lows; strictly control positions and wait for clear direction.
🔥 Market Barometer | August 12
Today's three hot topics point to the same theme: the market is fully transitioning from "storytelling" to a stage of "handing in the answer sheet"—the capital feast of AI infrastructure is entering its first round of return validation.
🏗️ Cloud vendors submit financial reports: AI investment enters return validation phase
In Q2 earnings season, the four major cloud providers delivered their first "report card" of AI investment. Amazon AWS's revenue was $42.2 billion, +37% year-on-year, marking the fastest growth in 18 quarters; Microsoft Azure was +43% year-on-year, with annual Azure revenue surpassing $100 billion for the first time; Google Cloud's revenue was $24.8 billion, an 82% year-on-year increase. The combined cloud business revenue of the three companies was about $116.2 billion, up about 43% year-on-year.
More importantly, order backlog. AWS orders backlog reached $496 billion, triple-digit year-over-year growth; Google Cloud order backlog reached $514 billion; Microsoft's business RPO rose 84% year-over-year to $678 billion—visibility into future revenue is improving.
But the cost is just as real. Amazon's free cash flow turned from positive $18.2 billion to negative $7.6 billion over the past 12 months; Google's free cash flow is under short-term pressure. The combined quarterly capital expenditure of the four companies has soared to $151.4 billion.
The market is voting with its feet: rewarding companies that can turn hash power into real cloud revenue, and punishing narratives that only invest without return.
📊 CPI released tonight: The scale for a rate hike in September hangs in the balance
At 20:30 Beijing time on August 12, the US July CPI will be released. The market expects the overall year-on-year CPI to fall from 3.5% to 3.4%. Before the data release, CME data showed the probability of a rate hike in September remained at 51.2%.
Deutsche Bank expects CPI to rise 0.15% month-on-month and core CPI to 0.26% month-on-month. The Cleveland Fed forecasts overall CPI for July to rise slightly 0.09% month-on-month, with core CPI 0.21% month-on-month. If tonight's data exceeds expectations, the hawkish camp will quickly expand; If moderate, rate hike expectations may fade further.
💰 Nvidia's $500 billion vs. Intel's $20 billion: divergence between two paths
On August 10, two chip giants simultaneously announced financing plans.
NVIDIA has partnered with six institutions including Apollo, BlackRock, BlackRock, Goldman Sachs, and KKR to establish an independent computing power financing platform, aiming to leverage over $500 billion in third-party capital. Jensen Huang stated: "Computing power has now become infrastructure like electricity and the internet. "The essence is to turn GPUs from consumables into financing infrastructure assets." After the announcement, Nvidia's stock price closed down 2.86%.
Intel announced a $20 billion common stock issuance, marking the largest single equity financing since its IPO in 1971. On the day of the announcement, the stock price closed down 4.06%. Both paths point to the same conclusion: competition in AI chips has escalated from a technology race to a capital race.
💎 Summary
Cloud vendors proved AI demand is real with 43% revenue growth, but the $151.4 billion quarterly capital expenditure also reminds the market that the pace of burning cash has never slowed; Every basis point in tonight's CPI could determine where the scales of September rate hikes will tip; Meanwhile, the 500 billion and 20 billion yuan financing plans announced by Nvidia and Intel on the same day mark the AI race officially entering a new phase of "capital-intensive." When industry logic, macro narratives, and capital strategies converge on the same day, the market is fully moving from "storytelling" to "handing in the answer sheet." #今晚CPI公布, will the pricing for September rate hikes be rewritten?
#财报观察员: AI infrastructure earnings report debuts one after another
#黄金站上4400美元, demand for risk avoidance is heating up When the guiding midpoint slipped down like a gently knocked central pawn, the sensitive funds immediately redirected the cart; But the grandmaster's gaze never fixed on that pawn; instead, it landed on the opponent's exposed king-wing gap—Sandisk's move was not a slip, it was bait.
The earnings report marks a completed start. Quarterly revenue and adjusted earnings per share that exceeded consensus indicate the start was not off track. And the loss of the median guidance for the next quarter was just a deliberate suspension. The market was eager to interpret this as a prelude to weak demand, but I saw an open line about to be opened—if management knew demand hadn't stalled, then this conservative forecast was a trapping tactic set for Investor Day on August 13. It forced all opponents to choose between "caution" and "weakness," while the real players were already prepared for a third response: this was merely to adjust the activity of their pieces.
The supply and demand tides of NAND flash memory have never been static compositions. Everyone knows that if you hesitate on the rear wing pawn chain, your opponent will immediately control the d5 central grid. The fundamental game now is: can the AI storage roadmap become a true dark horse advancing forward? When the entire tech chessboard revolves around computing power and reasoning, storage hoarding and release become a mid-term pawn chain structure. Will you choose to hold onto the current valuation, or trade that seemingly heavy defensive pawn early? Every public speech and every product route adjustment is changing the direction of this pawn chain.
And the $14 billion buyback is not a gesture, but a clear long-term shift—using the company's own funds to adjust the security of the royal city. It tells the market: even with short-term guidance, players remain confident they can survive until the endgame. But the real experts look at whether this huge sum will lock in future flexibility in NAND expansion and AI storage. Capital allocation is always a double-edged game: excessive conservativeness traps the elephant in the corner; excessive aggressiveness leaves the king exposed on the open line.
The interactions of the US stock market are more like mirrored moves on a chessboard. When Sandisk, a piece, shifts even a square and the derivative expectations represented by XLLY calculate equivalent changes on the same timeline. Many people chase fleeting ups and downs on intraday charts, like betting on a mistake in a quick kill at the start; But the players who truly make money have already rehearsed all branches of this path twenty moves in advance. They don't look at the crumpled chart before them, only whether the pivot of the entire board has been established.
The outcome of the entire chess game is not in the financial reports already dropped, nor in the guidance for next month. Instead, on August 13, will management open their calculations and show the best path to the AI storage endgame, or will they merely add a mediocre transitional move? Every number dropped is a new prompt; Every defensive tone is an alarm of sideline threats. The fog on the chessboard is thickening, but the ones truly worth calculating are always the hidden lines no one wants to look at even once.
At this moment, the blind spot is not in the center of the board #sandiskinvestorday#AI基建融资升温, Nvidia and Intel are diverging in their paths
AI infrastructure financing heats up, and NVIDIA and Intel have embarked on two completely different expansion paths
The AI computing power arms race continues to heat up, and the way capital competes is undergoing a major transformation. NVIDIA and Intel have chosen two completely opposite financing paths, signaling that the industry's future development pattern may undergo new changes.
Recently, NVIDIA partnered with top Wall Street institutions such as BlackRock, BlackRock, and Goldman Sachs to build a brand-new AI computing power financing platform, aiming to leverage over $500 billion in third-party capital. The core logic of this model is very clear: NVIDIA does not directly spend money to expand its own production capacity, but instead collaborates with financial giants to provide financial support to downstream customers, making it easier for them to purchase GPUs and build data centers. Simply put, NVIDIA is building channels to direct external capital to the computing power demand side, further consolidating its chip sales base. This major cooperation is still ongoing, and although the final agreement has not yet been finalized, it is enough to shake the entire AI industry.
On the other side, Intel has chosen to inject capital inward. Reports indicate that the fundraising scale for Intel's common stock issuance may be raised to $20 billion, with market enthusiasm exceeding expectations, and subscription demand surpassing $100 billion. All the funds raised will be used to expand capital expenditures, supplement operating cash flow, and increase investment in AI chip R&D and advanced chip manufacturing production lines. Intel hopes to use the additional share financing to independently expand its manufacturing and chip businesses and actively participate in the computing power market competition.
The two models are fundamentally different. Nvidia leverages financial capital to empower downstream customers and leverages market funds to drive GPU demand, which is a form of leveraging capital outward; Intel supplements its own funds through equity financing, focuses on self-developed and self-produced products, and expands capacity through its own strength.
As AI industry investment grows larger, financing capability will become a key benchmark for the capital market to judge tech giants. Both approaches have their pros and cons: Nvidia's model can quickly expand the computing power market pie, but continuously amplifying industry leverage means that if downstream computing returns decline, the chain reaction risk cannot be underestimated; Intel's share issuance model can firmly control its own supply chain but faces valuation pressure from equity dilution.
The era of deep intertwining between capital and technology has arrived; AI is no longer just a technological competition but a long-term contest of financial strength. The different choices of the two chip giants will not only affect companies' valuations but also change the global AI computing power supply landscape, while also affecting market expectations for global technology and risk assets. Future developments are worth continuous tracking.#海力士推进NAND扩产, storage supply expectations are rising
Recently, I have been continuously following news in the storage sector. SK Hynix's push for NAND capacity expansion is worth taking a moment to review carefully.
According to the currently disclosed plans, the expansion of SK Hynix's Dalian Phase II NAND production line is progressing steadily, with production equipment expected to be introduced gradually in the second half of 2026, and by the first half of 2027, 30,000 to 50,000 new wafer capacity per month will be gradually released. Not only domestic factories, but SK Hynix is also ramping up its NAND capacity in South Korea. The expansion is not a single regional move but a simultaneous overall capacity increase.
From today's perspective, AI data centers are developing rapidly, and demand for enterprise-grade SSDs continues to rise. Currently, overall NAND chip supply remains tight, and the logic of short-term shortages and firm prices still holds. This is the core reason why the storage sector has maintained its momentum recently.
But the market always anticipates trading in advance, and risks often slowly brew during hot market moments. As major manufacturers implement their expansion plans, capital attention is quietly shifting. Previously, the most discussed topic was the short-term supply shortage and price increase dividends; The next market debate will shift to 2027 demand, whether this concentrated supply release can be steadily absorbed, and how long the storage upward cycle will last.
To put it bluntly, short-term market sentiment still exists, and the positive news hasn't been fully realized immediately. But the medium- and long-term divergences are already clear before everyone. In storage, you can't just focus on immediate shortages and be optimistic; while enjoying the current cycle dividends, you also need to be alert in advance for future unleashed production capacity. Controlling the pace will become increasingly important, and the margin for error in chasing higher prices will gradually decrease. Maintaining rational observation is the more appropriate attitude at present.#财报观察员: AI infrastructure earnings report debuts one after another
Let's talk about the AI infrastructure earnings season is like a relay race, running one after another, running at full speed. Everyone is focused on the same core question: has the money spent burning through AI investments over the years really turned into real cash income?
Let's look at CoreWeave: Q2 revenue was 2.58 billion, doubling instantly, with an order backlog soaring to $104 billion, and it surged 9% after hours. The losses are still ongoing, but much less than expected. To put it bluntly, the losses are purely due to aggressive expansion, and demand is unmistakable. This company is basically a GPU rental giant, with customers lining up to pay for computing power, and orders are scheduled for years. This isn't telling a story—it's real demand not meeting demand.
Looking at Lumentum, demand for optical modules has indeed exploded, revenue doubled, and EPS exceeded expectations, yet it only rose slightly 1.8% after hours. The market's reaction seems a bit lukewarm, probably because they thought good expectations had already been chewed up. Everyone knows the optical module sector is booming, but expectations are set high and slightly less explosive than expected, so the stock price is not very convinced. A typical "good news but not enough surprise" scenario.
Tonight, Coherent will continue up. This company is also driven by optical communications + lasers, with both revenue and guidance likely to be solid. But whether the market will offer a premium depends on whether it can deliver stronger data than Lumentum, such as improved gross margins or increased new orders.
Tomorrow night, AMAT (Applied Materials) will be the main event, the leading semiconductor equipment company, and its guidance directly affects the confidence of the entire chip industry chain in the second half of the year. If AMAT says customers are still expanding production, then the logic for AI infrastructure will remain strong; If it starts to be cautious, then it will have to reconsider.
Also, Thursday's $SNDK investor day has been lukewarm lately. Let's see if they can come up with new stories, such as enterprise-grade SSDs or AI storage solutions, with expected volume expansion.
After mastering these three strategies—Coherent, AMAT, SNDK Investor Day—you can basically see whether AI infrastructure is continuing to tell stories or it's time to carefully calculate the details. Personally, I lean toward the former, but the market has now become numb to good data that is "within expectations"; only by exceeding expectations can there be room to expand.
As for the $SPCX unlock, over 300 million shares will be released on August 20, so we need to understand the pressure on the chips. With such a scale of unlocking, no matter how strong the fundamentals, short-term liquidity shocks are unavoidable, so you need to be cautious in position management.
Overall, the long-term logic of AI infrastructure is sound, but short-term fluctuations will not be small. Earnings season is a process of distinguishing truth from falsehood—real gold fears no fire, and those who swim naked will eventually be exposed. Let's watch as we walk 📡
#AIInfraEarningsWatch #CPIToResetFedBets #AIInfraFundingDivergesAfter four years of shelving, SK Hynix suddenly restarted its Dalian NAND factory—AI is too hot, we can't wait any longer
---
📰 1. Event Overview: A factory that had been shut down for four years suddenly resumed operations
On August 11, according to an exclusive report by South Korea's Seoul Economic Daily, SK Hynix has officially resumed construction of its No. 2 NAND flash memory production base in Dalian, China.
This factory began construction in May 2022, and after completing the main framework, it stagnated for four years due to the sluggish storage industry. Now, against the backdrop of explosive demand for AI data centers and a surge in enterprise solid-state drive (eSSD) demand, this "zombie factory" has been reactivated.
🏗️ 2. Production Capacity and Timeline: A 50% increase in capacity, with production to begin in the first half of next year
· Capacity Planning: The new production line will deliver about 50,000 wafers per month, and with the existing 100,000 wafer capacity at Dalian No. 1 Factory, the total capacity of the Dalian base will increase by about 50%.
· Production Commissioning Timeline: Production equipment will begin in November this year, with a mass production system established and officially launched in the first half of next year
· Operating entity: Solidigm, a NAND subsidiary under SK Hynix, is responsible for promoting the event
· Technical Route: Dalian focuses on mature low-layer products (about 100 layers of NAND), while Korea's Cheongju M17 factory concentrates production of high-end products over 300 layers, forming a differentiated division of labor between China and Korea
🔥 3. Why restart now? —AI is too "hungry."
1. NAND prices surged nearly tenfold in one year
The expansion of AI data centers has driven a surge in demand for enterprise-grade solid-state drives (eSSD), with NAND flash prices rising nearly tenfold within a year. The investment that had been put on hold for four years has finally reached its moment to "settle accounts."
2. AI servers "not satisfied"
Major cloud vendors are accelerating their deployment of AI data centers, with procurement volumes of high-capacity, high-reliability NAND products continuing to rise. Faced with AI's appetite, existing production capacity is completely insufficient.
3. Dual-track layout formed
Dalian is responsible for volume sales of mature layer products, while Qingzhou M17 handles high-end products over 300 layers. Using Dalian's mature capacity to "drive volume" and Qingzhou's high-end capacity to "make money" — walking on two legs is more steady than on one leg.
📈 4. Impact on SK Hynix: Short-term Positives, Long-Term Double-Edged Sword
Short term: Tightening supply, stock prices and sectors have begun to react
After the announcement, Southern Double Long SK Hynix ETF (07709) surged 5.5%; The A-share semiconductor equipment sector followed suit, rising accordingly. The market interpreted this as a signal of increased certainty in capacity expansion, directly driving demand for semiconductor equipment procurement.
Mid-term: Capacity release next year may change the supply-demand landscape
Monthly wafer delivery increased by 50,000 units, but its share of global NAND supply is limited. More importantly, new capacity will only be gradually released next year and cannot ease supply tightness in the second half of 2026. As long as AI demand maintains its current growth rate, this batch of new capacity will most likely only ease pressure moderately and will be unlikely to reverse the overall supply shortage.
Long-term: Customer structure is a double-edged sword
SK Hynix is deeply tied to NVIDIA in the HBM field, and the recent expansion of the Dalian factory further focuses on the enterprise-level NAND market. Some analysts point out that almost all of the company's core profits come from Nvidia's single HBM orders. This is an advantage when AI demand surges, but once it adjusts, the risks also concentrate.
⚠️ 5. Risks and Challenges
1. Equipment import restrictions
The U.S. continues to tighten export controls on semiconductor equipment to China, and whether the Dalian factory can smoothly introduce the required production equipment remains uncertain.
2. Korean brokerages worry that storage cycles are peaking
Some Korean brokerages have begun worrying that the memory chip industry may peak, slashing target prices for Samsung and SK Hynix by about 30%. Between "capacity expansion" and "cycle peak," the market is repricing.
3. Concerns about long-term overcapacity
If AI demand growth slows and competitors like Samsung and Kioxia expand production simultaneously, this batch of new capacity could turn from a "timely rain" into "adding insane damage."
💎 6. Summary
SK Hynix's restart of its Dalian NAND factory is the most direct evidence that AI computing power demand has spilled over into the memory industry chain. The sudden resumption of work at the factory after four years of shelving isn't because of policy changes, but because AI is simply too "hungry"—demand for enterprise-grade SSDs is exploding, NAND prices have risen tenfold in a year, and if production isn't started soon, it's really too late.
Dalian's capacity expansion is SK Hynix's "second battlefield" in the AI era—HBM is responsible for "making big money," while NAND is responsible for "running large volumes." In the short term, this is a confirmation signal of strong AI demand; In the medium term, a 50% capacity increase next year could change the supply-demand landscape; In the long term, whether AI can continue to "starve" is the fundamental variable determining the final return on this investment.
$SKHYNIX $SKHY #CLARITY延期，SEC拟推进监管规则补位
最近在看美国加密监管的进展，我发现一个很有意思的格局：国会立法被延后，但SEC准备抢跑，用部门规则来补监管缺口。
CLARITY法案虽然已经过了参议院银行委员会，但是全院表决要推迟到9月才能进行，正式落地还遥遥无期。
可SEC这边没有干等着，计划在8月14日召开公开会议，要审议加密资产投资合同发行、融资豁免还有安全港相关安排。
也就是说，国会层面的顶层立法还没走完，SEC打算直接利用现有的权限，去界定代币发行、项目融资的相关规则。
在我看来，这是非常关键的转折点。
原本市场都在盼CLARITY法案落地，给整个行业一套清晰的法律框架。现在变成法案延期，监管规则先行补位。
这里就留下两个很大的疑问：
第一，靠SEC的行政规则，能不能真正填补美国加密市场的监管空白？
第二，这种“监管先行”的模式，会不会直接改写往后美国加密行业的合规发展路径。
后续这次8月14日的会议，大家一定要重点盯，里面出来的细则，会直接影响整个市场的走向。 #财报观察员: AI infrastructure earnings report debuts one after another
PCE turns negative month-on-month, GDP growth slows to 1.5%
After discussing the two newly released U.S. data sheets, it feels like the market is about to start tugging back and forth again.
In June, PCE surprisingly fell 0.1% month-on-month, marking the first single-month negative since 2020, and year-on-year declined to 3.7%. Core PCE held steady at 3.3%. On the surface, inflation appears to be cooling quite noticeably, and the urgency of short-term rate hikes has eased considerably.
But the annualized GDP in the second quarter was only 1.5%, clearly below the market's 2.1% expectation. Interestingly, after excluding exports, inventories, and government spending, private consumption data actually surged to 3.9%, the highest since 2023.
To put it simply: the overall economy is weakening, but ordinary people are still spending heavily.
This puts the Fed in a dilemma: on one hand, inflation is declining; on the other, domestic demand remains strong. With about a 63% chance of a rate hike in September, it's highly likely to reprice again.
The biggest question now is whether this PCE weakness marks the beginning of inflation decline or a temporary monthly fluctuation influenced by oil prices. The final answer will still depend on the July data.
The recent market is likely to fluctuate repeatedly, so aggressive chasing is not advisable. Be patient and wait for the direction to clarify.The CLARITY Act is stuck in the Senate, and the SEC no longer intends to wait. On Friday, August 14, a public meeting was held, planning to launch its own "Regulation Crypto" issuance system.
Key highlights: Proposed 4-year startup exemption for 5M, 4-year startup exemption for 5M, 1-year fundraising exemption for 75M, and a safe harbor for compliant token issuances. This is equivalent to overseeing the gap in the absence of Congressional legislation.
But don't celebrate too soon—this is just a 'proposal', following a notice-and-comment process, which historically takes 12–18 months. CLARITY's 9/15 vote, requiring 60 votes, remains another main thread. Dual regulatory tracks run in parallel, but certainty is still far off.
If the SEC legislates itself, do you think it's more reliable than Congress? Let's talk in the comments.
#CLARITY延期, the SEC plans to advance regulatory rule supplementation Wow, NVIDIA has been reported to be developing the next-generation open-source large model Nemotron 4, with at least 1 trillion parameters, aiming directly at the world's leading open-source model.
Chip giants are rushing into open source, but their calculations stick to the old ways: open ecosystems to expand AI applications, and then continue to drive demand for their own GPU computing power. On August 11, Nvidia closed down 0.02%, but the stock price barely reacted when the news broke.
At the same time, Jensen Huang cooled down the $500 billion AI financing plan on X: Nvidia's support scale does not exceed 25% of individual project opportunities, based on residual value, only filling the gap and not replacing independent underwriters. Previously, the market worried that Nvidia would bear too much risk exposure in this major plan, but after clarification, related credit risk indicators fell back on August 11.
The shovel seller started by burning models themselves, but the ledger was still the hash power book 😂
#AI基建融资升温, Nvidia and Intel are diverging in their paths #标普收盘再创新高, the 8,000-point level is expected to heat up
At the close of U.S. stocks on August 11, the S&P 500 fell 0.32%, the Nasdaq dropped 0.6%, and the Dow fell 0.34%. On the surface, it looks like a typical pullback, but when you break it down, the contrast is striking: uncertainty in US-Iran negotiations supported oil prices, Brent crude rose 1.4% to $88.91, and the S&P energy sector rose 1.1%; Meanwhile, Alphabet fell 3.8%, and Amazon dropped 2.1%.
Rising oil prices are not a one-way button for the market. Energy companies may benefit, but higher pressures on transportation, production, and inflation will expose highly valued tech stocks to tougher interest rate expectations.
This also explains why $BTC is affected. Crypto assets may not directly depend on oil prices, but they can receive secondary shocks through risk appetite, the dollar, and interest rate expectations.
The real theme last night was not "the stock market has fallen," but rather that funds are reassembling between energy and growth assets.实在是抱歉，这波我真看不懂谁在接，$VELVET短期确实很难看好。
昨天大家跟着我的逻辑，做空$DOS 的应该都吃到肉了，今天再来分析下一个短期仍不看好的代币（ velvet:native）。
这个项目是个 DeFAI 叙事的链上交易 + 组合管理终端，YZi Labs 等投过，Binance Alpha IDO 出来过。6月靠叙事（AI + 合成 Pre-IPO）从低位猛拉到接近2刀，现在回落到0.45-0.6区间，从ATH已经腰斩再腰斩。
说一下做空逻辑：
➫解锁压力实打实：8月10日前后有一批解锁（大约2%左右供应，涉及insider、private、community等），金额不小。
➫市值和真实使用严重脱节：MC两亿多，TVL长期在百万级别甚至更低。交易量很大一部分是刷Gems冲榜、做空投、刷推荐。
➫出货痕迹：6月大涨期间，有团队关联和DWF相关地址往交易所转过不小量。
➫营销味道重：空投+排行榜活动吸引一堆KOL撸毛宣发。短期热度能撑，但热度一过就是接盘。
➫历史黑料：2024年发生前端钓鱼事故，合约审计早期也出过严重问题（后来修了），信任已受损。
做空可以考虑Unlocking—these two words almost come with a horror movie soundtrack in the crypto market. Whenever any project announces a "token unlock" schedule, the community immediately becomes alarmed; many people's first reaction is "the price is going to crash again." But is the truth really that simple? Recently, @Tokenomist_ai conducted a study on 236 unlocking events, offering a different perspective. The real severe negative impact mainly concentrates on early-stage projects—those with very small circulating supplies where the unlocked amount is disproportionately large relative to the circulating supply. In other words, unlocking itself is not scary; what’s scary is a sudden surge of disproportionate supply during moments of low liquidity. The market gets spooked not because the sell orders are too large, but because the market depth can’t support that selling pressure. The essence of fear is a loss of proportional balance, not the numbers themselves. Looking at the Top 300 projects with less than 30% of their tokens unlocked is even more interesting. The following ten projects are currently still in a state where most tokens remain locked: RaveDAO’s $RAVE is 23.03% unlocked, Backpack’s $BP is 25.00%, Raydium’s $RAY is 26.00%, Lighter’s $LIT is 26.10%, Zama’s $ZAMA is 26.46%, Plasma’s $XPL is 26.89%, SentientAGI’s $SENT is 27.15% #今晚CPI公布，9月加息定价会改写吗？
Tonight's US July CPI is arguably the most important recent milestone, directly rewriting the pricing for the Fed's September rate hike.
Let's first review the previous nonfarm payrolls situation: July nonfarm employment unexpectedly decreased by 23,000, and May and June data were revised down by a total of 103,000, showing weakening employment. Initially, the market adjusted down the September rate hike expectations accordingly. But market sentiment quickly reversed. Now, CME rate tools show a 52% probability of holding rates steady in September, and a 48% probability of a 25bp hike, with the split almost even and disagreement at its peak.
The market has left all suspense to tonight's CPI.
Currently, the market consensus expects: overall CPI monthly rate at 0.1%, core CPI monthly rate at 0.2%; year-over-year overall CPI expected at 3.4%, core CPI at 2.5%.
Two scenarios can be simply outlined:
1. Inflation continues to cool as expected, combined with weak employment data, the weak economic logic dominates, and September rate hike expectations will fall again, which is relatively favorable for risk assets.
2. Conversely, if core inflation exceeds expectations, the pressure to hike rates returns immediately. The dollar and US Treasury yields rise, and whether it's the US stock market or crypto market, they will have to re-digest the September policy risk, with volatility significantly increasing.
Now, employment and inflation signals are pulling in opposite directions. Which side the Fed will lean toward will be an important clue from tonight's data. In terms of trading, I think it is safer not to bet on the direction in advance. The restart of memory chip capital expenditure and shareholding restructuring strengthened supply-side pricing power, but Micron's 27% pullback from its peak indicates divergence in the market regarding the pace of downstream demand digestion, with position rebalancing risk becoming the short-term dominant factor.
The secondary market showed a divergence between bulls and bears, with $SKHYNIX current price rising to 1,048 and the underlying stock rising 4.7% in a single day. At the same time, Micron pulled back 27% from its June high, reflecting a shift in risk appetite from chasing highs to structural discernment. Dalian's capacity plan increased to 150,000 wafers and Kioxia completed a 14.19% equity restructuring, prompting bull funds to reallocate positions within the sector.
The order of driver transmission is: First, supply-demand squeeze caused by inventory dropping to about four weeks and DRAM price increases of 65% quarter-on-quarter and NAND price increases of 75% in Q1; Second, JPMorgan Chase raised its 2026 global storage market size forecast to $969 billion, triggering a valuation restructuring; Third, increased industry concentration leading to macro inflation expectations.
The trigger for the upward breakout scenario is that $SKHYNIX holds above the key support at 1,000, with price increases remaining effective throughout the year. If the price surges to the 1,100-1,150 range, it will confirm the premium support for valuation from AI demand. This scenario failure signal is that the price breaks below the 1,000 support level, at which point long positions will need to close positions to lock in existing profits to avoid deep pullbacks.
The trigger for a downward pullback scenario is that Micron's 27% drawdown represents downstream digestion delays spreading across the entire industry, with profit-taking concentrated and profit-taking. If the 1,000 stop loss protection level is breached, the price may drop to the valuation retest zone mapped by the previous low of 134 prominence. This scenario signal is that the Philadelphia Semiconductor Index will regain ground and push tokenized stocks above 1,150.
If the new 50,000 wafer capacity at Dalian NAND Fab 2 is released more rapidly, although it will expand market share in the long run, the market may interpret this in the short term as capital expenditure expansion, changing the sustained pricing of the supply-demand gap. Whether capital sentiment can shift from defensive rebound to trend-driven upward depends on whether the price increases can offset liquidity discounts during the US market off-market.
In the next 7 days, it is important to watch $SKHYNIX's position accumulation at the key 1,000 level, and whether the Philadelphia Semiconductor Index can break through its 0.87% single-day gain to sustain the following momentum.
#存储股抛压缓和, is the AI memory bull market still stable? #CLARITY延期, the SEC plans to push regulatory rules to fill the gap#财报观察员：AI基建财报接力登场 目前人工智慧基礎設施領域的企業財報正相繼公布，整體進程宛如一場無縫接軌的接力競賽。市場與投資人所高度關注的共通核心，無非是在過去段落中被大量資本砸入的AI基礎建設投資，究竟能否真正轉化為實質且可觀的帳面營收與利利潤？ 以算力租賃服務商 CoreWeave 為例，其第二季度的營收表現極為亮眼，達到了 $25.8 億，增幅高達一倍；與此同時，尚未交付的積壓訂單總額更是攀升至令人咋舌的 $1040 億，這直接帶動其股價在盤後交易中大漲 9%。儘管公司目前仍維持虧損狀態，但虧損幅度已明顯小於市場原先預估。究其根源，當前的虧損完全是源於企業為了應對市場需求而進行的激進產能擴張，而前端買方的需求力道依然非常強勁，並未出現衰退。 另一方面，光通訊模組大廠 Lumentum 同樣展現了市場對光模組的爆發性需求，其單季營收實現了倍數成長，每股盈餘（EPS）也順利超越法人預期。然而，其盤後股價僅小幅上揚 1.8%，市場反應顯得相對冷靜。這種偏向平淡的市場回饋，反映出投資人可能早已將相關利多預期提前反映在股價之中。 展望接下來的市場動向，今晚登場的 Coherent 財報📌 SanDisk SNDK investor preview analysis on August 13
I. Basic Information of the Conference
Investor Day will be held at 9 a.m. Eastern Time on Thursday, with key disclosures:
1. HBF storage technology roadmap and commercialization timeline
2. Technical details of BiCS10 10th generation 3D NAND
3. SSD expansion planning and long-term supply cooperation agreements
Both technologies are designed to address AI storage bottlenecks and are key highlights of this rally.
2. Breakdown of Two Core Technologies
1. HBF (jointly developed by SanDisk ×SK Hynix)
Positioned between high-end HBM memory and regular SSDs, it fills the gaps of both:
- HBM is costly and has small capacity; Traditional SSDs have low bandwidth and slow read/write speeds, while HBF strikes a good balance
- Uses 8/16-layer NAND stacking, with a maximum of 512GB per module, bandwidth 0.4~3TB/s, and UCle interfaces supporting direct connection to GPU and CPU
- Google and Tenstorrent have entered the ecosystem, focusing on solving the storage wall problem of AI large model inference: no matter how strong the GPU computing power is, if data reading lags behind, it will be dragged down
2. BiCS10 10th Gen 3D NAND (Jointly developed by SanDisk + Kioxia)
- 332-layer stacked QLC flash, bit density increased by 59% over the previous generation, sample deliveries to begin in the second half of the year
- CBA wafer bonding process optimizes logic circuits and memory arrays separately, offering high density and low power consumption, targeting AI data lakes and RAG knowledge base storage scenarios
3. Current Stock Fundamentals
1. The stock price fell from the June high of 2354 to 1214, a drop of nearly 48%, with a price-to-earnings ratio of only 6 times, indicating significant valuation compression
2. Last period's revenue was 8.97 billion yuan, exceeding expectations, but the earnings guidance was lowered by 250 million yuan, weakening short-term sentiment in the capital market
3. Institutional target price: Citibank offers 2500, analyst average price 2220, indicating ample premium room for expectations; Executives slightly reduce holdings at high levels to suppress sentiment, but this is not considered core negative factors
4. Key Market Judgments
The market trend depends entirely on the clarity of HBF commercialization:
1. ✅ Roadmap and mass production timeline are clear: The technical story is expected to generate revenue, funds will re-raise valuations, and stock prices will enter a recovery rally
2. ❌ Ambiguous planning and unclear implementation timeline: Still stuck in the thematic concept, the stock price is likely to fluctuate prolongedly in the 1200 range and bottoming out
Essentially: The market is now watching whether HBF and BiCS10 can transform from purely technical concepts into tangible performance revenue.
Information is for reference only and does not constitute investment advice. #SK Hynix advances NAND capacity expansion, storage supply expectations rise. #AI基建融资升温, NVIDIA and Intel see divergent paths. #今晚CPI公布 Will the September rate hike pricing be rewritten? The CPI is coming tonight—some are happy, some are worried
Bitcoin and $BTC fell for two consecutive days, while Ethereum also slipped from around 1,930 to 1,880.
My view:
No rate hike in September, but short-term pressure hasn't fully been released. Nonfarm payrolls turning negative, credit tightening, and inflation trend downward—all three factors combined, so Washe has no reason to force a rate hike. He's been hawkish since taking office, but the data makes it clear—employment is shrinking, core inflation is cooling, and raising rates now doesn't make sense.
CME prices a rate hike probability of around 45%, which sounds intimidating, but the U.S. Treasury market has already signaled — the 2-year yield is stagnant, and the market simply doesn't believe it.
What truly hurts the crypto world are two forces: ETF funds are flowing, with a net outflow of 145 million on August 10; oil prices are at the top, the Hormuz agreement has not materialized, and Brent is approaching $90. Inflation won't come down, and even if Walsh wants to relax, he can't find a way out.
$XAU gold soared, breaking through the $4,400 mark.
Ethereum $ETH is more sensitive to interest rate expectations than BTC, and staking yields are directly related to funding costs. Currently, it is grinding between 1,850 and 1,920; if CPI is moderate, it is very likely to break through 1,930, with the next stop at 2,000. Bitcoin should definitely break through 65,000
$SPCX On this side, the first batch of 911.5 million shares unlocked on August 6 was not sold, yet the stock price actually rose 23% over two days, approaching the $135 IPO price. However, on August 20, 319 million shares remain unlocked (about 7%), about 700 million shares in September, and nearly 700 million shares in October, with chips still flowing outward.
$SNDK Thursday Investor Day was the highlight. Q4 revenue surged 372%, but then fell 7% in after-hours trading—what the market wants is the HBF mass production milestone and BiCS10 schedule, not past data. If you can pay clearly, the stock price will be repriced; If you can't, the 1,200 will still have to be scrapped.
Tonight at 8:30, we will focus on core month-on-month growth and rent. CPI will only reinforce one judgment: there will be no rate hikes in September, and these assets will react before the market reacts.
#今晚CPI公布, will the pricing for a rate hike in September be rewritten?
#霍尔木兹海峡通航协议未落地, oil price risks are heating up
#现货ETF资金分化, BTC selling pressure remains $GOOGL
$SPCX
$GOOG 有点猛了说实话！最近其披露对外投资持仓展示其就持仓 $SPCX 一家公司就占了大约 95%！
投资 9 亿，15 年时间，现在价值 900 亿美元，回报超 100倍……
和大家聊一下细节部分内容……
根据 Alphabet 截至 2026 年 6 月 30 日的 13F，它一共披露 29 项公开证券持仓，总市值约 990.8 亿美元；其中 SpaceX 就有 551,189,500 股，季度末价值约 941.8 亿美元，占整个 13F 组合约 95.05%。
不过这里有一个非常关键的细节：这并不是Google最近突然花900多亿美元买了SpaceX。
恰恰相反，这其实是一笔持有了超过10年的超级长期投资。
2015年1月，Google就参与了SpaceX融资。Alphabet后来在自己的10-K里明确披露，当时Google向SpaceX投资了 9亿美元；
当时媒体报道Google获得大约 7.5%的股份，对应SpaceX估值约120亿美元。Google和Fidelity那一轮合计投资10亿美元，获得接近10%的SpaceX。
换句话说：
2015年：Google投入约9亿美元。
2026年Q2：13F里SpaceX持仓价值约941.8亿美元。
当然，这其实也不能简单粗暴地直接理解成“Google赚了104倍”，因为中间还有稀释、股权变化、拆股等因素；
但它依然可以算是Google历史上最漂亮的一笔战略投资之一。
而这次13F还有一个很容易误解的地方，我看也没有人聊。
那就是SpaceX今年6月上市以后，Google长期持有的这部分股权才第一次作为公开证券进入13F。
所以这次13F真正告诉我们的不是Google刚刚买入了SpaceX，而是上市以后，我们终于第一次清楚看到了Google手里到底还有多少SpaceX。
而且941.8亿美元是6月30日的市值，对应13F披露的551,189,500股，折算大概就是 170.86美元/股的季度末市场价格，并不是Google的实际成本价。
Google真正最早的一笔成本，还是十多年前那9亿美元。
更有意思的是，看看Alphabet剩下那5%的公开投资组合：
$PL Planet Labs约 11.7亿美元；
$ASTS AST SpaceMobile约 7.95亿美元；
$ARM约 6.95亿美元；
另外还有一批AI、生物科技和软件公司。
所以Google过去投资其实一直有一个挺明显的特点：
我觉得也很有意思，来多讲两句……
它不是只买和自己主营业务完全一样的公司，而是愿意非常早地押注未来十年可能变得重要的基础设施。
AI、自动驾驶、生命科学、卫星、通信网络都是如此。
而SpaceX这笔投资尤其典型。
2015年的SpaceX还远远不是今天的SpaceX，当时Starlink甚至都没有正式商业化，Google就已经拿出9亿美元下注。
十多年以后，SpaceX已经把火箭、Starlink、AI、卫星通信甚至未来轨道数据中心逐渐串成了一整套基础设施。
所以我自己看到这份13F以后，反而更加坚定了之前对于 $SPCX 的长期判断。
短期股价当然会受到估值、财报、CAPEX和解禁影响，我之前该看空的时候一样看空。
😂😂😂
但是长期来说：我依然非常看好SpaceX。
有时候真正的大机会，并不是靠持有100家公司找到的。
而是在真正理解一家公司以后，愿意陪它十年。
其实再多聊一两句，一个好的企业投资另外一个好的企业，左脚踩右脚不是什么稀奇的事情，上次我在聊腾讯的时候就聊到过，腾讯其实在国内投资了非常多比较不错的企业，这其实也是一种盈利的模式。
后面我再给大家深度挖掘一下可以聊的东西，我们下期再见…… Before $BTC emerges, who might launch first, $ETH or $SOL? The key is not the rally, but the "relative strength."
Recently, the market has been discussing the 'knockoff rotation,' but it's still too early to call it the 'knockoff season.'
BTC continues to fluctuate around key areas, while in recent days, SOL showed clear relative strength, once hitting a two-week high.
When judging whether ETH or SOL will be stronger in the next phase, I won't just look at USDT trading pairs.
The real highlights are:
ETH/BTC and SOL/BTC.
Assuming BTC rises 3% and ETH rises 2%, ETH appears to rise on the surface, but ETH/BTC is actually weakening.
Conversely, if BTC consolidates sideways, SOL rises by 3%, and SOL/BTC continues to break out, it indicates that funds are actively increasing their SOL allocation.
Capital rotation is usually closer to:
BTC stabilizes → ETH/SOL strengthens compared to BTC → mainstream altcoin transactions have increased → further spreading funds into popular sectors.
There is another macro variable that cannot be ignored: tonight's CPI.
Research data shows that changes in CPI expectations have some predictive power for the realized volatility of altcoins like ETH and SOL. In other words, if tonight's inflation data significantly deviates from expectations, ETH and SOL could be more volatile than BTC.
So now, I won't chase a coin just because it's "resistant to a drop," but wait for three conditions:
BTC no longer hits new lows; ETH/BTC or SOL/BTC begin to strengthen steadily; The rise is accompanied by increased spot trading.
Only when all three conditions appear simultaneously does the credibility of capital rotation significantly increase.
Risk boundary: If BTC breaks below core support, the counterfeit rotation logic needs to be recalculated. True strength is not about BTC rising and then rising accordingly, but about when BTC moves sideways or even slightly pulls back, there are still funds willing to actively buy.
#今晚CPI公布, will the pricing for a September rate hike rewrite? #现货ETF资金分化, BTC selling pressure remains $CRWV $KORU
CRWV：
Current price is 104.29, 24-hour +17.71%, range 84.61-105.46, stuck between 103.10-104.63 in the past 2 hours. Over 15 minutes, the main increase surged near 88 to 105, then volume contracted and sideways thereafter, funding rate 0, OI about 1.51 million USD, more like a revaluation and short covering after earnings stimulus. It is CoreWeave, a US AI cloud computing company. On August 11, official Q2 revenue was $2.575 billion, backlog about $104 billion, a clear catalyst. Risk is tough: net loss $626 million, expansion is too costly, if 103 cannot hold, don't hold on.
KORU：
Current price is 18.55, 24-hour +8.67%, range 16.61-18.63, up from 17.88 in the past 2 hours. 15-minute volume concentrated in the 17.65-18.23 breakout phase, with about 1.92 million USD traded in the past hour, fee 0, open interest about 3.26 million USD, more like a recovery in Korean equity risk appetite. KORU is the Direxion 3x Korean stock market long ETF, benchmarked against MSCI Korea 25/50, with weights biased toward SK Hynix and Samsung Electronics. No confirmed recent catalysts yet; future depends on whether Korean tech stocks, exchange rates, and the US AI chain can take over. Risk is 3x ETF volatility wear; if you can't get past 18.63, don't chase hard.
#CRWV #KORU #AI算力 #韩国科技股📊 $RE contract liquidation express (August 12)
According to liquidation data, RE shows a pattern of rapid short-term direction switching and medium- to long-term bullish overshooting:
· Short-term (1H/4H): 1-hour long liquidation at $87.89, short liquidation at **$0, shorts completely neutralized; 4-hour long at $87.89 (unchanged from 1-hour), short at $8,588.24, bears crushing long positions by 97.7 times**, shorts crushing long positions at 97.7 times**, short-term period extreme reversal from short squeeze to long selling, short chasers targeted and blown.
· Medium- to long-term (12H/24H): 12-hour long liquidations at $23,700, short at $10,300, bulls at 2.3 times bears; 24-hour bulls at $56,300, short at $39,500, bulls at 1.43 times bears. The scale of long liquidations in medium- to long-term cycles continues to expand, but the 24-hour multiple narrows significantly, with bulls and bears becoming more balanced.
· Total liquidations have exceeded $95,800, with long positions at $56,300, accounting for nearly 59%. Bulls dominate slightly, but the gap between bulls and bears is not large.
⚠️ Risk warning: RE short-term trend extreme switching (1H short squeeze →4H long sell), very high risk of double kill between long and short; 24-hour long-short multiple narrowed to 1.43x, indicating unclear direction. Leverage is recommended to be compressed to within 3x; do not chase rallies or cut losses; strictly control positions while waiting for direction to clarify.
🔥 Market Barometer | August 12
Today's three hot topics point to the same theme: the market is fully transitioning from "storytelling" to a stage of "handing in the answer sheet"—the capital feast of AI infrastructure is entering its first round of return validation.
🏗️ Cloud vendors submit financial reports: AI investment enters return validation phase
In Q2 earnings season, the four major cloud providers delivered their first "report card" of AI investment. Amazon AWS's revenue was $42.2 billion, +37% year-on-year, marking the fastest growth in 18 quarters; Microsoft Azure was +43% year-on-year, with annual Azure revenue surpassing $100 billion for the first time; Google Cloud's revenue was $24.8 billion, an 82% year-on-year increase. The combined cloud business revenue of the three companies was about $116.2 billion, up about 43% year-on-year.
More importantly, order backlog. AWS orders backlog reached $496 billion, triple-digit year-over-year growth; Google Cloud order backlog reached $514 billion; Microsoft's business RPO rose 84% year-over-year to $678 billion—visibility into future revenue is improving.
But the cost is just as real. Amazon's free cash flow turned from positive $18.2 billion to negative $7.6 billion over the past 12 months; Google's free cash flow is under short-term pressure. The combined quarterly capital expenditure of the four companies has soared to $151.4 billion.
The market is voting with its feet: rewarding companies that can turn hash power into real cloud revenue, and punishing narratives that only invest without return.
📊 CPI released tonight: The scale for a rate hike in September hangs in the balance
At 20:30 Beijing time on August 12, the US July CPI will be released. The market expects the overall year-on-year CPI to fall from 3.5% to 3.4%. Before the data release, CME data showed the probability of a rate hike in September remained at 51.2%.
Deutsche Bank expects CPI to rise 0.15% month-on-month and core CPI to 0.26% month-on-month. The Cleveland Fed forecasts overall CPI for July to rise slightly 0.09% month-on-month, with core CPI 0.21% month-on-month. If tonight's data exceeds expectations, the hawkish camp will quickly expand; If moderate, rate hike expectations may fade further.
💰 Nvidia's $500 billion vs. Intel's $20 billion: divergence between two paths
On August 10, two chip giants simultaneously announced financing plans.
NVIDIA has partnered with six institutions—including Apollo, BlackRock, BlackRock, Goldman Sachs, and KKR—to establish an independent computing power financing platform, aiming to leverage over $500 billion in third-party capital. Jensen Huang stated: "Computing power has now become infrastructure like electricity and the internet. "Essentially, it's about turning GPUs from consumables into foundational assets that can be financed." After the announcement, Nvidia's stock price closed down 2.86%.
Intel announced a $20 billion common stock issuance, marking the largest single equity financing since its IPO in 1971. On the day of the announcement, the stock price closed down 4.06%. Both paths point to the same conclusion: competition in AI chips has escalated from a technology race to a capital race.
💎 Summary
Cloud vendors proved AI demand is real with 43% revenue growth, but the $151.4 billion quarterly capital expenditure also reminds the market that the pace of burning cash has never slowed; Every basis point in tonight's CPI could determine where the scales of September rate hikes will tip; Meanwhile, the 500 billion and 20 billion yuan financing plans announced by Nvidia and Intel on the same day mark the AI race officially entering a new phase of "capital-intensive." When industry logic, macro narratives, and capital strategies converge on the same day, the market is fully moving from "storytelling" to "handing in the answer sheet." #今晚CPI公布, will the pricing for September rate hikes be rewritten?
#今晚CPI公布, will the pricing for a rate hike in September be rewritten?
#黄金站上4400美元, demand for risk avoidance is heating up 🦅Interpretation of AI infrastructure financial report cycles: Results confirm whether AI demand is real, and multiple leading companies have released their reports one after another
1. The core logic of this round of AI earnings season
Optical module, computing cloud providers, and semiconductor equipment companies have collectively disclosed their financial reports, with only one core criterion: whether the AI capital expenditure invested by major companies has truly been converted into real revenue, to judge whether the AI industry chain demand is purely thematic hype or genuine demand.
2. Breakdown of disclosed target performance
1. CoreWeave ($CRWV Computing Power Rental)
After-hours stock price rose 9%, Q2 results far exceeded expectations:
1. Quarterly revenue was $2.58 billion, a year-on-year increase of 112%;
2. Net loss of $626 million, better than the market expectation of $757 million, with losses narrowing significantly;
3. Backlog of orders reaches $104 billion, with core clients covering leading AI companies like OpenAI, Meta, and Microsoft.
Conclusion: Demand for computing power leasing continues to surge, losses are only caused by capacity expansion, and downstream demand is completely fine.
2. Lumentum (optical module manufacturer)
Q4 revenue was $1.01 billion, doubling year-on-year, with EPS earnings per share exceeding expectations by $0.28;
By simultaneously binding Google TPU and NVIDIA GPU to two major AI computing power routes, the demand for AI optical modules has been realized.
However, it only rose 1.8% after hours, reflecting that the market had already priced in the rally expectations, and the positive results were actually realized.
3. Key Targets to Be Disclosed in the Financial Report
1. Coherent (after market close on August 12)
Core supplier of AI optical interconnection, with market expectations for revenue of 1.98 billion yuan, up 30% year-on-year. Financial reports are used to verify whether the growth rate of the AI optical communication industry continues to accelerate.
2. Applied Materials AMAT (after market close on August 13, leading semiconductor equipment company)
Expected revenue is 9 billion yuan, a year-on-year increase of 23%. As a leading semiconductor equipment company, performance reflects chip manufacturers' willingness to expand production and determines upstream capacity cycles.
4. Other Key Events
1. SpaceX ($SPCX) unlocking cycle
The first batch of shares unlocked on August 6 did not trigger a sell-off, and the stock price held above the IPO issue price; However, on August 20, the second round of unlocking occurred, with 319 million shares unlocked, accounting for about 7% of total share capital. In subsequent September and October, large shares were still unlocked in batches, with chip supply continuously increasing, which will suppress price elasticity for a long time.
2. SanDisk SNDK (August 13 Investor Day)
The HBF technical roadmap and the commercialization timetable for BICS10 will be announced; Includes grid trading arbitrage cases, showing stable grid strategy returns during range-bound market fluctuations.
5. Overall Summary
Tonight's Coherent and tomorrow's AMAT earnings report, combined with Thursday's SanDisk Investor Day, three key pieces of information will directly determine the future direction of the AI infrastructure market:
If earnings exceed expectations, the narrative of high AI prosperity will continue; If earnings fall short of expectations, funds will return to valuation logic and start squeezing thematic bubbles. #财报观察员: AI infrastructure earnings report debuts one after another #今晚CPI公布, will the pricing for a rate hike in September be rewritten? $BTC
Whale increased holdings by 46,420 BTC over 60 days, but network activity dropped by 45%
Addresses holding over 10,000 BTC net increased their holdings by 46,420 BTC (about $2.97 billion) over 60 days, with tokens continuing to concentrate from scattered holders to super holders. Long-term holder (LTH) supply declined simultaneously, further confirming the trend of token concentration.
On the other hand, network transaction volume fell 45% year-on-year, creating a stark structural contradiction between on-chain activity and whale accumulation. Whales are accumulating, but the market lacks trading depth, and short-term rebound momentum is insufficient. This contrasts with the previous report on miner whale liquidation (6,494 BTC sold in 20 days) that pointed to different batches of whales—market participants' behavior is clearly different. CoreWeave surged sharply after hours, and I also hope it pulls back a bit—not to surge too aggressively all at once
Last night, CoreWeave's financial report was indeed impressive: revenue of 2.58 billion yuan, more than doubling year-on-year, backlog of orders reached 104 billion yuan, and after-hours sales jumped about 15%. To be honest, this data is quite important for restoring confidence in the AI infrastructure sector.
But then again, while the earnings report is good, the rapid rise is a bit uncertain. The market has just surged, so I actually hope it can pull back and not fully hit expectations all at once. After all, the company's debt is indeed high, and interest expenses have more than doubled. After every big rally, profit-taking always comes out. Moreover, SpaceX has made new moves, starting to sell surplus computing power to foreign markets. Meta may also develop its own cloud business. Although the threat is not great for now, it will somewhat lower the valuation ceiling.
With 104 billion yuan in orders, AI demand remains, but prices have risen to this level. Moving slowly is more stable than taking everything in one step; a pullback can actually give you more room to come.
#财报观察员: AI infrastructure financial reports make their debut — $CRWV #今晚CPI公布, will the pricing for a rate hike in September be rewritten?
The CPI data is set to be released at 8:30 tonight, and rate hike expectations have now reached a 50-50 split—CME FedWatch shows a 52% chance of rates unchanged in September and a 48% chance of a 25 basis point hike. The market expects the overall CPI annual rate to fall from 3.5% to 3.4%, and core CPI to drop from 2.6% to 2.5%.
Core CPI is the real winner
Expected month-on-month growth of 0.2%, annual rate of 2.5%. The Cleveland Fed's Nowcasting model shows core CPI rose 0.21% month-on-month in July. If the actual data is close to 0.2%, basically in line with expectations, the market may not experience much volatility. But if it exceeds 0.25%, the probability of a rate hike in September is very likely to return above 60%.
Chicago Fed President Goolsbee took a hawkish stance ahead of the CPI release, saying, "Inflation is currently the biggest problem." Both hawks and doves at the Fed are waiting for this data; who wins and who loses depends entirely on the numbers.
Three possibilities
Core CPI in line with expectations (0.2% month-on-month, 2.5% year-on-year): The probability of a rate hike in September remains around 50%, with BTC fluctuating in the short-term range of 63,000-65,000. SanDisk Grid continues to run its volatility.
Core CPI below expectations (0.1% or lower): The probability of a rate hike in September may drop to 30%-40%, $BTC there is a chance to break through 65,000 or even challenge 66,000. $XAU Gold may break through 4420 and continue to rise.
Core CPI exceeds expectations (0.3% or above): The probability of a rate hike in September could jump directly above 70%, and BTC could pull back to 62,000 or even 61,000. Gold is under short-term pressure.
My position
$SNDK SanDisk's grid has been running for 7 and a half days, with 3,439 arbitrage trades, grid profit of 39U, unmatched losses reduced to -29U, total profit +9.79U. Price around 1290, lower boundary 1219, upper bound 1490, sufficient space.
Regardless of CPI data, as long as prices fluctuate within the range, the grid will continue to move. The only thing to watch out for is that the data far exceeds expectations and causes prices to break below the lower boundary, but the strong parity price of 919 still has more than 300 points to spare, so there's no need to panic for now.🦅 In-depth analysis of capital games: institutional ETFs bottom-fishing, whale miners selling, BTC and ETH trends diverging completely
1. Core Situation: Funds are tugging at both bulls and bears, BTC is stuck in a sideways consolidation
ETH saw nearly 1.1 billion ETF inflows in one week, while BTC prices fell below the 64,000 mark. The market polarization essentially reflects a tug-of-war between buying institutional funds and selling whale chips. Tonight's US July CPI data is the biggest short-term variable, with the market widely forecasting a year-on-year drop to 3.4%; Persistently declining inflation raises interest rate cut expectations. After market risk appetite recovers, ETH's upward resilience has always been stronger than BTC's.
Many people wonder: with BTC spot ETFs continuously seeing large net inflows, why does the market still fluctuate around 64,000? The core answer is that institutions are buying on one side, while whales and miners are selling aggressively. The two funds are fiercely contested, naturally leading to a short-term market stalemate.
1. Buyers: Wall Street ETFs are steadily accumulating long-term funds
Last week, US BTC spot ETFs saw net inflows for five consecutive trading days, totaling $853 million. BlackRock's IBIT alone accounted for $694 million, accounting for over 80% of total inflows.
BlackRock has also stated that this batch of funds is mainly for long-term allocation. Even if many institutions are investing at high levels of $100,000 and $110,000, they won't panic and sell their losses at the moment. Simply put, Wall Street institutions are slowly buying at the $60,000 price level to support the bottom, with strong buying resilience.
2. Sellers: On-chain whales and miners are fleeing at a loss, with overwhelming selling pressure on existing holdings
On-chain monitoring data clearly reflects selling pressure:
1. An anonymous whale sold a total of 7,513 BTC over three weeks, equivalent to approximately $487 million;
2. A suspected miner address deposited 6,494 BTC to Binance within 20 days, selling at an average price of $64,798, with a market value of $421 million;
One of the shares was bought at a high price of $116,111 a year ago, but now has a 44% loss and is painfully cutting losses.
Not only are retail investors cutting losses, but large holders holding large amounts of chips are also cashing out, with existing selling pressure continuously suppressing BTC's gains.
Summarizing the current landscape: ETF long-term funds keep taking over, on-chain whales + miners keep selling; Buying is new institutional funds off-exchange, selling is the exit of existing tokens on the exchange. Both sides are fiercely clashing at the 64,000 level, BTC volatility has dropped to its lowest point of the year, and bulls and bears are currently evenly matched.
2. Key Divergence: ETH accumulation efficiency surpasses BTC, and the two diverge in trend
Recently, the capital flows of the two major coins have started to show a clear gap:
Single-day capital flow: BTC ETF saw a net inflow of about $112 million, ETH ETF saw a net inflow of about $56.78 million;
Seven-day dimension: ETH ETF accumulated a net inflow of 118,500 Ethereum, valued at $225 million.
Although ETH's total funds are smaller than BTC's, considering market capitalization, ETH attracts more efficiently and concentrates tokens faster.
The advantage on the downside is also prominent: BTC dropped nearly half from its peak, while ETH fell from 4800 to 1900, a drop of over 60%. Greater room for decline means better valuation cost-effectiveness and a higher rebound ceiling.
3. CPI data determines the future market trend
1. CPI below expectations, cooling inflation
Rising expectations of rate cuts and loose liquidity have strengthened risk assets. ETH, due to its high concentration of chips, has fallen even harder, so its rebound will be much more explosive than BTC;
2. CPI exceeds expectations, inflation rebounds
With rate hike expectations rising again, the entire crypto market will also face pressure and correction, with no standalone strong stocks.
4. Final Summary
At this stage, the market is in a quiet period before the CPI release, with uncertainties hiding:
BTC's advantage lies in strong institutional confidence and stable ETF funding, but long-term on-chain selling pressure has dragged down the upward momentum;
ETH has fallen deeper and is less valued, with better recent capital inflows. Once the market recovers, its elastic advantage will be fully unleashed. $BTC $ETH #今晚CPI公布, will the pricing for September rate hikes be rewritten? #财报观察员: AI infrastructure earnings reports debut in succession. #今晚CPI公布, will the pricing of September rate hikes be rewritten? 🐸 Meme coin traders—this is a hunting ground tailored just for you. The platform is offering a $20,000 BTC prize pool to reward sentiment traders in the market. The rules are simple and direct: buy or deposit $SHIB, $DOGE, $PEPE, with a total of 2,000 winning spots, ending on August 25. But don't treat it as just a marketing campaign—there's a deeper signal behind it. 🔥 On the surface, it's a reward, but at its core, it's a battle for liquidity. The narrative of Meme coins has never been about value discovery, but about close engagement. The platform chose this time window to target the most active period of retail funds in the meme sector. Through lottery-style incentives, it reactivates the wait-and-see capital that has settled in the market, while injecting buying pressure into these three asset categories—smart money sees chip turnover, not just a simple benefit handout. 📊 On the data side, the 2,000 winners are set up with very subtle nuances. It's not a universal model of 'everyone wins,' but rather a mindset of expecting a 'high probability of winning.' This precisely hits the most vulnerable nerve for traders: FOMO and luck coexist. More importantly, the event rules especially emphasize the dual channels of 'buy or recharge,' meaning the platform wants to capture both incremental trading and activate existing positions. Killing two birds with one stone, skillful tactics. 💡 For market participants, the real information is not about the size of the prize pool, but about the platform's bet attitude toward the meme coin sector. During a period of overall cautious sentiment, daring to use real money to incentivize these three tokens speaks for itselfThis wave was stepped on correctly, SK Hynix's stock rose another 4.7% today. What exactly is the storage chip sector speculating on this time?
Still holding the long position on SKHYNIX 997, current price 1,048, floating profit 126%. The stock rose another 4.7% today, rebounding from a low of 134 to now 141.65, roughly a 5% bounce. Let's talk about what's been happening recently in the storage chip sector.
Dalian expansion, restarting the second factory
SK Hynix has restarted construction of the Dalian NAND flash second factory, aiming to increase local production capacity by 50%. The new production line has a monthly capacity of about 50,000 wafers, combined with the existing 100,000 wafers from the first factory, bringing Dalian's total capacity to 150,000 wafers. This factory started construction four years ago but was halted due to industry downturns. Its restart indicates a change in industry expectations. NAND prices have increased nearly tenfold over the past year.
Becoming Kioxia's largest shareholder
Toshiba has been continuously reducing its holdings, and SK Hynix, through SPC2, holds 14.19% of Kioxia's shares, surpassing Toshiba's 14.06%, becoming the largest shareholder. The NAND flash industry landscape is being reshuffled.
The fundamentals of storage chips are indeed shifting
JPMorgan has raised its global storage market size forecast to about $969 billion in 2026, expected to reach $1.44 trillion in 2027, and $1.82 trillion in 2028. AI demand is driving simultaneous growth in storage chip prices and shipments, and this logic remains intact. SK Hynix's DRAM unit price rose 65% quarter-over-quarter in Q1, NAND rose 75%. The company clearly stated that this year, demand from all customers cannot be fully met, and price increases will continue throughout the year. Inventory is down to about 4 weeks.
The secondary market is also moving
At the Korean stock market open, Samsung rose over 4%, SK Hynix rose over 2%. The US storage sector also rose yesterday, with SK Hynix up 4.7%, SanDisk up 2.68%, Seagate up 2.44%, Micron up 0.87%. The Philadelphia Semiconductor Index rose 0.87%. However, Micron has fallen 27% from its June all-time high, indicating significant divergence.
My judgment
The logic of the stock stabilizing + token oversold is still in play. The three news items—Dalian expansion, Kioxia equity, JPMorgan's raised forecast—overlap, keeping short-term sentiment alive. Continuing to hold the 997 long position; the first target of 1,050-1,070 has been reached, next looking at 1,100-1,150. Stop loss moved up to 1,000 to lock in profits.
The storage chip story is not over yet, but it has risen too much in the short term, so watch the rhythm.
Brothers, what do you think about this storage market wave? Let's chat in the comments. Personal views for reference only 🤫 👇$SKHYNIX #存储股抛压缓和，AI内存牛市还稳吗？ 账户仓位分歧雷达
方向共识是真是假，拿账户比例和头部持仓一对就知道。
$DOGE 账户数口径一致偏多，但头部持仓比仍在1下方，人数优势没有变成头部仓位优势。 下跌同时扩仓，卖压有新仓配合，但仍不能只靠OI确认空仓方向。 价格若上行但头部持仓继续偏空，回撤时仍容易出现仓位口径冲突。
$CAP 账户数口径一致偏空，但头部持仓比在1上方，偏空人数没有变成头部空仓优势。 价格与持仓同步回落，这段先按减仓下跌处理。 头部持仓比没跌回1下方前，空头账户优势仍是不完整的共识。
$XRP 账户数已经偏向多侧，头部仓位规模没有跟随，当前分歧来自数量与权重。 下跌没有带来持仓扩张，先看风险敞口收缩何时放慢。 账户端已经偏多，后面就看头部仓位是否愿意把权重压到同一边。$BICO surged sharply before, attracting quite a lot of market attention. After reaching its peak, $BICO entered a prolonged period of consolidation, resulting in no significant rebound. I infer that $BICO is unlikely to have a major rebound in the short term, meaning the $BICO story is temporarily over. Unless something unexpected happens, $BICO will continue to decline. It might even fall below the initial rise level. —————————————————— Let's look at its contract data. We can see that its contract open interest had a small increase yesterday, while the long-short ratio was declining. This indicates that many shorts were active during yesterday's rebound. Looking at a longer timeframe, although the current price is roughly the same as a couple of days ago, the contract open interest is much lower than before, and the long-short ratio has increased significantly. This suggests that the price stability over the past two days is likely due to short sellers taking profits. Short covering has both pros and cons. The benefit is that the short-selling pressure decreases. The downside is that market liquidity also reduces. Liquidity is the foundation of all price increases; without liquidity, any rise is rootless. —————————————————— Based on my judgment, I believe $BICO is unlikely to see significant improvement in the short term. Perhaps in the next cycle, it might experience another surge? That, I am not sure about I know many people still don't think BTC has bottomed out and are still waiting for 40k
I just scanned the altcoins that have dropped sharply over the past two years. After two years of accumulation, some have already started to break out, and some have even surged tenfold. These are signs of the market bottoming out. Although most markets won't rise together until October, the same thing has happened repeatedly over the years, and some coins always start first, leading the way.
Just because you can't get meat doesn't mean you don't have any; if you lack technology, don't blame the market.
#今晚CPI公布, will the pricing for a rate hike in September be rewritten? $BTC $ETH On the eve of the CPI release, the crypto market's "risk appetite + safe-haven" "risk appetite" both failed—I saw this situation clearly on the market. Inside the box, there was a naked gambling direction; this was the most certain judgment at the moment.
♾️ BlockInfinity 8/11 California Tuesday: On the eve of CPI, both risk appetite and risk aversion in crypto failed, with the bet box showing bare gambling
🌐 Macroscopically
The biggest variable this week = US July CPI (to be released Wednesday). Consensus: Overall year-on-year growth 3.4% (previous 3.5%), core 2.5% (top 2.6%). Two scenarios—(1) Hot-biased data → concerns over stagflation and falling US stocks, Wash's inclination toward rate hikes; (2) <3% → Risk appetite returns, giving back on year-end rate hike pricing. Employment "bad news = good news," but CPI "bad news is bad news." The Fed prices a rate hike in December at ~28bp and September at about 50/50. Timiraos: More and more FOMC members are watching CPI to decide "whether rate hikes are needed"; New Chairman Wash downplays the "data dependence" framework, but the old framework still dominates.
🌍 International situation
Strait of Hormuz remains closed: Iran's Supreme National Security Council reiterates that "the U.S. will not reopen unless conditions are accepted" (demanding an end to the war + unfreezing funds), while the U.S. continues to exert "extreme pressure" on Iran. Hedging signal: Pakistan hints that U.S.-Iran is close to reaching an agreement → Oil prices give back early gains; An explosion occurred at Libya's Zawiya refinery. U.S. military intelligence has determined that Iran's strategic focus has shifted from its nuclear program to the strait, with an assessment of force control that is "lengthy, deadly, extremely costly, and cannot guarantee victory"→ Geopolitical premiums are hard to reduce.
🛢️ WTI $83.20 (+1.3%, having retreated from an intraday high of +5%); 🥇 Gold $4,370 (retreated after a two-month high), silver $64.64 (−1.9%), copper $6.685 (+0.2%).
📈 Technical Aspects (Multiple Cycles)
🔴 $BTC $63,289 (−0.95%): Daily MACD base death cross (DIF95<DEA101 bar−12) turns bearish, breaking below MA20 (64,169), RSI14 46; 4H RSI 35, 1H RSI 32 deeply oversold, 1H Bollinger band only 1.09%, ⚠️ extreme compression inverter. Ranges 62,227–66,397 hold the lower boundary.
🔴 #ETH $1,864 (−0.5%): Weakest leg in the structure, daily MACD bearish, near MA50/EMA50, 4H/1H RSI 33–36 oversold, range 1,820–1,982 margin turning bearish.
🟡 #SOL $75.1 (−0.9%): Relatively the most resilient to declines, daily chart still above MA20 (74.45), bullish alignment at RSI 49; 1H RSI 32, short-term oversold.
🔬 Derivatives / Capacity
🔴 24h liquidation = bullish flood: BTC $55.4M (long blow $50.0M / short blow $5.4M ≈ 9:1), ETH $32.4M (long blow $26.5M), SOL $4.5M (long blow $4.2M) — this round of decline = long squeeze.
🟡 Rates are mildly positive (BTC +0.010% / ETH +0.007%, non-extremes); BTC OI $46.4B, 24h only +0.16% (unleveraged, short positions buying while buying).
🟡 Spot discount −0.11% / −$71 (no buying in US spot session); Fear and greed index 29 (fear); BTC DVOL 35.98 (hidden wave pressure at low level).
📈 U.S. Stocks (intraday, about 1 hour before close)
🔴 Nasdaq slightly weak QQQ 717.9 (−0.7%); AI capex led declines ORCL −4.3% / GOOGL −3.0%
🟢 CRCL +7.4% stands out; TSLA +0.7%、META +1.3%
🟡 Storage Differentiation: SNDK +1.8% / SKHYNIX +2.2% / MU −2.2%; INTC −1.6%
₿ BTC Core Judgment
The daily chart shows a death cross for the first time, with multiple periods in a bearish alignment, but the price remains within the consolidation range (62.2–66.4K) and has not broken through; 4H/1H RSI is deeply oversold + bulls just got wiped out at $50M. Meaning = "Weak but oversold." Before trend confirmation is lacking, chasing short sellers = fueling short squeezes, chasing long = running barefoot before the CPI binary event.
🎯 Key positions: break above 65.8–66.9K to turn long; break below 62–63K to open the next gear; Between the two = noise.
⚖️ Comprehensive judgment
Crypto's dual failure in "risk appetite" and "risk aversion" continues: gold hits a two-month high, oil inflation premiums reignite, but BTC follows the decline but not the rise—the market positions Middle East/oil prices as "rate hikes" rather than "safe havens." Before the CPI release on Wednesday, all directions are binary bets: overheated to suppress risk assets, cooler to ignite a rebound. Current optimal = low exposure and other data; avoid heavy positions at the 65K magnetic level + before the event.
🎯 Today's Trading Perspective (For reference only, not personal positions)
• BTC: No chasing at the mid-range (63–64K); Breaking above 65.8K and holding above = light position to try long, stop loss at 64.8K; Breaking below 62K to close = turning bearish, stop loss at 63.2K. ATR daily chart ~$1,254, stop loss should not be narrower than this.
• ETH: Weakest leg, rebound 1,900–1,940 resistance, light short and stop loss at 1,965; break below 1,820 to accelerate.
• General discipline: Control position before CPI = preferred; Deeply oversold, avoid naked short chase (short squeeze fuel); if confirmation is lacking, don't chase long naked; first wait for a false breakout to be confirmed.
⚠️ Risk events
📅 Wednesday: US July CPI (the highest) + EIA crude oil inventories + OPEC/IEA monthly reports + Tencent Q2 earnings
📅 Thursday: US July PPI + initial jobless claims + JD.com earnings report
⚠️ Non-investment advice, DYORGold has become a hot topic above $4,400, while BTC is around $63,821 and ETH is still consolidating at low levels. The fact that these two scarce assets have not strengthened simultaneously indicates that safe-haven funds are currently more oriented toward traditional assets, and crypto is still trading as risk assets for now. I will verify three points: whether BTC can reclaim 64,000; whether ETH can hold above 1,875 and stop falling; As gold continues to rise, will BTC spot trading volume increase in tandem? If the divergence between strong gold and weak crypto markets continues, the risk-averse narrative has not yet been transmitted; Only if three improvements are made can capital be repriced. Do you think BTC will take over from gold, or continue to fluctuate independently? $BTC $ETH #美股全线走高, crypto stocks led the gains
A rather controversial question: By putting stocks with open trading times into perpetual contracts that run all day, does that improve efficiency, or does it package the risk of time lag as well?
When underlying stocks are closed, news does not pause, but spot prices are temporarily not updated. At this time, derivatives may first fluctuate as expected, then realign after the spot market opens. It may look like the same asset, but in reality, it bears a different set of liquidity and pricing mechanisms.
This doesn't mean the product has problems, but the "familiar company name" easily leads people to underestimate unfamiliar contract structures.On the surface, money seems to have become cautious. But more accurately: the funds have not disappeared; they are waiting for macro uncertainty to materialize. Particularly noteworthy is the recent renewed clear capital demand for U.S. spot crypto ETFs. BTC and ETH ETFs attracted about $1.1 billion in funding in the past week, indicating that some institutional funds have not exited due to short-term volatility. That's also why I'm reluctant to be overly bearish on the market right now. If tonight's CPI comes in below expectations, inflationary pressures continue to ease, and the Fed signals more accommodative policy, risk assets could quickly undergo a round of repricing. At that time, I will focus on these directions: ₿ $BTC remain the core barometer of the entire crypto market. Once the macro environment improves, BTC tends to be the first to gain liquidity support. ♦️ $ETH ETF funding needs and the ongoing development of the Ethereum ecosystem give ETH a strong institutional allocation logic. ⚡ $SOL If the market re-enters a risk-On mode, high-beta assets like SOL are usually more likely to access new risk capital. 🟡 $BNB Backed by a mature ecosystem and ongoing on-chain use cases, fundamentals remain worth watching. 🔵 $OKB As the OKX ecosystem expands, on-chain activity increases, and token utility and deflation expectations continue to rise, OKB remains one of the exchange platform tokens I am monitoring. But here is a very generous one