Orbit Post Sitemap
Brothers, hold tight to the batch of storage chips bought at the bottom, don’t get itchy hands, hold steady.
Last night’s Nvidia earnings report sends one clear signal: AI is accelerating, storage is insufficient, and prices will have to rise. Q2 revenue hit $96.2 billion, doubling year-over-year, with next quarter guidance at $108 billion. Jensen Huang’s exact words: “AI has reached an inflection point, computing power equals revenue”; but the CFO was more blunt — memory costs are rising more sharply than expected, and shortages will last at least until the end of fiscal 2028.
1. Price hikes are not just slogans
Nvidia itself has hinted to major clients that next year AI server cabinet costs will rise over 15%, mainly due to continued increases in HBM and DRAM contract prices. Their “cost items” are the “profit items” for Samsung, SK Hynix, and Micron.
2. The shortage is structural
The three major manufacturers are dedicating 70% of new wafer production to HBM. By 2026, high-bandwidth memory capacity will be fully booked, with core customer orders lined up through 2028. Next-generation rack storage costs have soared from over $300,000 to around $2 million, accounting for more than a quarter of total costs. This is not a cyclical fluctuation but a physical capacity bottleneck.
3. AI acceleration is the underlying theme
Q2 data center revenue was $89 billion, up 117% year-over-year, with fiscal 2028 revenue growth guidance at 70%, crushing analysts’ 44% estimates. The more models and denser inference, the tighter the binding between storage and computing power.
So SKHY MU $SNDK, these top storage players, are not just “theme speculation,” they are the water sellers for AI capital expenditure. If you bought at the bottom, don’t get shaken out by a few days of volatility; just wait patiently until the grapes are fully ripe $BTC am the mid-term intelligence guy.
$SNDK presents both opportunities and risks on the table.
Let's start with the positives. When Nvidia $NVDA released its $108 billion revenue guidance, AI, memory, and storage all soared. SNDK and $MU Micron are beneficiaries of AI infrastructure. Sandisk itself is the preferred choice for AI data center NAND and enterprise-grade SSDs, supported by contracts and buybacks. Kioxia is investing ¥6.3 billion in Japan to build a NAND factory, with Sandisk as a#BTC surges then falls back, options expiry amplifies the key level battle
$BTC surged from 62,000 all the way up to 81,000, touched 81,000 on the 25th and then reversed, now hovering around 78,000 — a typical "failed breakout + derivatives week" combo.
On Friday, about $640 million nominal BTC options on Deribit expire, with a Call/Put ratio of 0.83, slightly more bullish calls but a large number of calls stacked at the 75,000/80,000 key levels, with the highest open interest in 75,000 calls (about 236 million). Market makers' gamma hedging will pull the price toward the pain points, so before expiry the price will "pin" and needle back and forth between 75,000 and 80,000, which is much more likely than a one-sided breakout.
My judgment:
• 80,000 is not a real resistance, it’s a “false top” created by the options wall plus the psychological round number. The surge then fall is due to profit-taking on calls above and market makers selling hedges, not a complete spot market bearish reversal.
• 75,000 is the real watershed; breaking below it targets 72,000 support; reclaiming 80,000 with continued ETF net inflows means shedding the options burden.
• Expiry ≠ direction, it just removes the pinning. Volatility expands 24 hours before and after expiry; directional choice depends on who takes over after settlement — if spot ETF inflows can cover shorts, that’s a real breakout; otherwise, it’s a bull trap during settlement week followed by continued consolidation.
In terms of trading, don’t stubbornly guess a one-sided move between 75,000 and 80,000; reduce leverage one notch, wait for the close to pick a side and follow. This week’s profit comes from "volatility money," not "trend money."NVIDIA released its latest earnings report overnight, once again playing out the familiar script: all data outperformed market expectations, but after-hours price volatility was limited, indicating strong signs of positive news being realized.
Quarterly total revenue reached $96.2 billion, doubling year-on-year. Data center business contributed as a core force, with revenue of $89 billion, a year-on-year surge of 117%. Jensen Huang stated that AI has reached an industry inflection point, with computing power genuinely converting into enterprise revenue. Data confirms this: hyperscale customer revenue increased +102% year-on-year, and AI cloud and industrial customers grew 138%. Next quarter's revenue guidance is $108 billion, higher than the market expectation of $104.8 billion.
Even with explosive earnings, after-hours trading continued to weaken. This is exactly the same logic as SK Hynix and SanDisk SNDK: high growth has already been fully priced in by the market, and impressive results are expected results, making it difficult to drive stock prices sharply higher. This year, Nvidia's stock price has only risen 12%, with valuation and market imagination already fully tapped #PCEToJacksonHole #AIMonetizationBroadens #BTCOptionsExpiryTest Who exactly is the bull and who is the bear in $ETH? Here's an unconventional theory that is actually true! Right now, the biggest bull in the BTC market is the one holding the short positions! For example, Boss Ten recently closed his long positions. The moment he closed those longs, it caused a huge market shock. Many thought that closing longs meant he was bearish, but actually, it's the opposite. After closing his longs, if he wants to go long again, he must buy back in, making him the largest potential off-exchange bull, always ready with cash to jump back in. Conversely, if he still holds massive short positions that are not closed, to take profit on those shorts, he still needs to buy back. So by only holding shorts, he is actually the biggest off-exchange bull. To make money in this market, you must understand this theory clearly. Who is really the bull? Who is really the bear? Just wait—one day when Boss Ten closes his short positions, the market will experience a big rally just like when he sold his longs last time, shooting straight up. Those holding longs just need to be patient. Currently, both BTC and ETH are consolidating sideways. ETH is fluctuating around 2500 with a 20-point range, BTC is fluctuating around 78,000 with a 2000-point range. Once the consolidation ends, the market will break out and move to the next level. Aggressive traders can go long at the bottom of the range and short at the top, but the closer to the end of the consolidation, the greater the risk. $BTC 5.17%, this number,
is the real ceiling pressing down on $BTC.
Lately, what is everyone focusing on? ETF inflows, the halving narrative, on-chain data... but there is one thing almost no one has seriously discussed — the 30-year US Treasury yield.
A year ago, it was 4.89%. Now, 5.17%.
You might think, it only rose by 28 basis points? But the problem is, these 28 basis points are quietly rewriting the entire pricing logic of the crypto market.
What does that mean? Let me break it down for you.
The long-term US Treasury yield is essentially the "risk-free return anchor" for the whole market. When it stands above 5%, what does that imply? It means you don’t have to do anything; just buy US Treasuries and you can earn 5% passively.
At this point, if you look at BTC, to retain institutional funds, it has to offer a risk premium higher than 5%. But given BTC’s volatility, why would institutions willingly take on that extra risk?
So you’ll notice a very contradictory phenomenon: funds are still flowing into ETFs, the on-chain narrative is still being discussed, but BTC’s price just can’t rise.
It’s not that the bulls are weak; there’s an invisible ceiling pressing down on it.
What’s more troublesome is that this ceiling is hard to break in the short term. The US government’s refinancing pressure is increasing, issuing new debt to pay off old debt, and this cycle itself keeps yields elevated. It’s not something that can be brought down by a single speech or one data release.
Also, if you look at the Treasury’s recent expansion of long-term bond buybacks, yields did briefly fall, but quickly bounced back. Why? Because relative to the $40 trillion total debt, that buyback scale is just a drop in the bucket. What the market really cares about is not how much debt the Treasury buys at once, but how much more it will borrow in the future, whether inflation will rise again, and whether the fiscal deficit can be controlled.
So the current situation is very clear: the ceiling over BTC is not drawn by technical analysis, it’s welded on by US Treasury yields. Only when the 30-year yield trends down from the high of 5.17% will institutional funds massively flow back into risk assets, allowing BTC to truly break free from this invisible constraint and enjoy a loose liquidity environment.
In one sentence: no easing in US Treasuries, no flight for BTC.
#美国核心PCE持平上月，沃什杰克逊霍尔讲话如何定调？ $ETH $NVDA In a few weeks, $BTC rose from 62,000 to 81,000, and $ETH from 1,900 to 2,500. The increase is nearly 30%, and market sentiment has reversed sharply.
But this time is fundamentally different from the wave in April last year. Last time was a macro-driven warming trend lasting several months; this time it feels more like a short-term resonance caused by ETF inflows + short squeeze + oversold rebound, with concentrated strength but questionable sustainability. Since August, BTC spot ETFs have seen a cumulative net inflow of about $3 billion. Institutions are indeed buying, but after BTC surged to 81,200, it quickly pulled back to around 78,000, indicating strong profit-taking willingness as well.
ETH is relatively weaker, still some distance from its previous high. The most critical factor next is the Jackson Hole Fed statement; if hawkish, this rebound may stop here, if dovish, there is hope to open up upside space.
Before the direction is clear, there is no rush to call a bull market or to flee the top. Let the market find its own way; this is more reliable than drawing conclusions prematurely. #美国核心PCE持平上月，沃什杰克逊霍尔讲话如何定调？ The market is rising, but no one dares to say clearly who is truly betting. The flow of on-chain funds reveals risks earlier than prices.
1) Price and Funds
BTC slightly pulled back, ETH rebounded, SOL surged, but fund flows did not synchronize. New wallets on Hyperliquid opened short positions on ETH and BTC with 20x leverage, betting 5 million USDC on a decline. This is not just a short position; it’s new wallets, high leverage, and large assets combined, seeming to test the system’s resilience.
2) This Round of Hot Topics
CrowdStrike’s quarterly performance shows AI-driven cybersecurity demand hitting new highs, shifting market sentiment toward defense. AI is seen as a catalyst for security rather than a threat. This subtly resonates with the “security” narrative of crypto assets. If AI truly becomes infrastructure, security investments will rise but will also increase overall risk premiums.
3) How I Would Interpret It
The bulls’ logic is: AI improves efficiency, drives security demand, benefits the tech sector, and indirectly supports the crypto ecosystem’s infrastructure. The bears focus on: if AI models are abused, they may trigger new types of attacks, protocol-level vulnerabilities could be amplified, and liquidation risks may rise.
What to Watch Next? Watch whether the Hyperliquid platform issues risk warnings, observe changes in ETH/BTC on-chain staking rates, and see if new attack events are recorded on-chain. If no clear risk exposure appears, current fund movements still need verification.
For informational and market scenario analysis only, not investment advice. Crypto assets are highly volatile; please conduct independent research and manage risks.Fundamental Research Report $OKB / OKB (Exchange Token) $3.20
One-sentence conclusion: OKB ($OKB) comprehensive score 51/100, rating narrative outweighs execution. Breaking down in three layers, the company team has cash reserves, the protocol network shows signs of paid usage, and token value capture has been realized.
Fundamental breakdown: OKB (token $OKB), exchange token sector. Main focus is OKX platform token. Comparable to BNB, CRO. Traditional centralized platforms charge 15-40% commission, user data is not controlled by users. On-chain trustless transactions have lower fees, token incentives convert early users into contributors. Average transaction value $50-500/month, settlement requires USDC or fiat. Narrative-driven sector, usage drops 60-80% in bear markets. Positioned as an end-to-end vertical platform. Product implementation: protocol layer is officially running, on-chain dashboard shows protocol fees accumulating, with evidence of paid usage. Latest version not found, 60 valid commits in the last 90 days.
User side, address MAU not disclosed, DAU not disclosed, 24h trading volume $80.00M, TVL not found. Wallet addresses do not equal unique monthly active users; concentration of large addresses may overestimate real user count. Revenue side, user fees not disclosed, supplier revenue about 80-90% of user fees (to LPs and nodes), protocol treasury income $2.00M, token holder buyback and burn annualized no burn mechanism. 24h trading volume is business flow, not revenue. Company profit does not equal protocol profit, protocol profit does not equal token holder profit. Code side, 60 valid commits in 90 days, 25 active contributors, latest version not found. GitHub is grade A evidence for direct verification. Investment background, company equity financing checked via PitchBook/Crunchbase (grade A), token private and public sales checked via whitepaper, release schedule, and on-chain unlock contracts (grade A), market makers and ecosystem funding are grade B, not representing long-term VC holdings, technical integration checked via API/SDK evidence (grade B), strategic partnerships and logo walls are grade D. NVIDIA GPU usage does not equal NVIDIA investment, exchange listing does not equal exchange strategic investment.
Token side, total supply 1,300,000,000, circulating 950,000,000 (73.1%), FDV $4.20B, next unlock 2026-Q4 (accounts for +3.50% of circulation), annualized burn/buyback no clear mechanism. Must buy tokens to use product? Partially yes, medium value capture (staking/discount/governance). Compared with peers (uniform criteria, no cross-sector comparison): Circulating market cap: OKB $3.00B, BNB undisclosed, CRO undisclosed. FDV: OKB $4.20B, BNB undisclosed, CRO undisclosed. Annual revenue: OKB $2.00M, BNB undisclosed, CRO undisclosed. Monthly active addresses or users: OKB undisclosed, BNB undisclosed, CRO undisclosed. Figures based on public data snapshots, some missing data supplemented by official or industry sources. Valuation: circulating market cap $3.00B, FDV $4.20B, P/S 1500.0x, FDV divided by revenue 2100.0x. Pessimistic view discounts $3.00B by 50-70%, neutral range oscillates, optimistic expects revenue doubling, burn implementation, enterprise clients entering, FDV P/S aligns with top players. Summary: fundamentals solid (score 51/100). Token value capture realized (buyback/burn/Gas). Circulating market cap relatively expensive compared to fundamentals, overleveraged expectations, FDV moderate. Risk warnings: short-term large unlocks dumping, protocol income long-term zero, token demand relies solely on incentives (if incentives stop, usage collapses). Tracking metrics: protocol fees weekly, burn amount, active address retention, TVL/loan balance, GitHub version releases. Data from public sources for reference only, not investment advice. If indicator deviation exceeds 30%, reassessment needed.
Logic provided, decision is yours.
#FundamentalResearchReport #Crypto #Research #OKXOrbitWhat BTC is most worth watching now is not how much it has risen, but whether $80,000 can be reclaimed.
Looking at the latest market, $BTC is currently around $78,800, rebounding intraday from the $77,600 area, indicating there is still support below for the time being. But after yesterday's surge to $81,000 followed by a quick pullback, it has already proven that selling pressure above is heavy.
Next, the short-term focus is on two moves:
First, hold $77,600.
As long as this level is not effectively broken, the current situation looks more like a high-level consolidation after a rise, and a pullback is not necessarily a bad thing.
Second, break through $80,000 again.
If volume increases and it stabilizes above $80,000, market sentiment is likely to be reignited, with the next target naturally near the previous high of $81,000.
But if $77,600 is lost, don’t stubbornly hold onto the bullish logic; short-term caution is needed for a further pullback near $76,500.
BTC right now is actually very simple:
Watch $77,600 for support, $80,000 for breakout, and $81,000 for direction.
The real opportunity often isn’t chasing when sentiment is hottest, but waiting for the market to show its direction.#BTC冲高回落，期权到期放大关口博弈
After BTC surged to the 80000 level and then pulled back, a large number of options are concentrated to expire this week. 80000 has become the concentrated strike price for call options, and the gamma hedging effect is causing oscillations and tug-of-war near this key level. As expiration approaches, market makers dynamically adjust their positions; when the price nears the critical strike price, upward moves intensify selling pressure, while downward moves trigger passive buying, significantly increasing market volatility.
Many traders interpret options expiration as a signal for a sharp drop, but I personally disagree. Options do not forcibly crash the market unilaterally; they mainly amplify existing market trends and do not create entirely new directions. Coupled with the current market being in a greed zone and continued inflows into spot ETFs, the large-scale bullish structure remains intact. However, short-term profit-taking is accumulating, so the risk of a pullback is real.
In practice, do not use the maximum pain point as a direct forecast level; it should only serve as a supplementary reference. Maintain a base position in spot holdings and avoid adding chips at high levels; for contracts, be sure to reduce leverage and decrease high-frequency trading as expiration nears. The 80000 level is being contested repeatedly; only a volume-backed hold above it can open upward space. If it effectively breaks below 76200, the short-term pullback space will further expand.
Besides the options event, the Jackson Hole speech is also a key variable. The combined disturbance from derivatives and macro news will amplify two-way volatility, so heavy one-sided bets are not recommended.
Key follow-ups: the gain or loss of the 80000 level, volatility changes around options expiration, and the linkage with U.S. Treasury yields.#财报观察员：英伟达超预期，软件收入开始兑现
NVIDIA's latest earnings report once again exceeded expectations, with its pre-market stock price rebounding strongly by nearly 5%! But this time, what excites Wall Street the most is no longer just how many GPU hardware units were sold, but that its software and ecosystem service revenues are beginning to truly scale.
This marks a qualitative leap in the moat of the AI computing power leader:
Breaking free from the pure hardware cycle curse: The market has always worried that the hardware procurement boom would peak. Now, the explosion of the CUDA ecosystem, enterprise AI software, and cloud service subscriptions proves the success of its transformation into a high-margin software platform.
The flywheel effect of a computing power closed loop: Hardware establishes a monopoly, software locks in customers. NVIDIA not only controls the computing infrastructure but also fully controls the deployment lifeline of upper-layer AI applications.
Industry chain synergy surges: NVIDIA's strong and unexpected performance has directly boosted partners like Marvell and optical communication customized chipmakers, completely shattering short-term doubts about AI investment returns.
From selling shovels to becoming an ecosystem ruler, how high do you think NVIDIA's software narrative can push its market value?
$NVDA $MRVL #财报观察员 #英伟达 #AI算力 #美股 #半导体
IllustrationThe full set of US July PCE data has been released, showing overall strength. The market's expected combination of "inflation continuing to cool + economic weakening + Fed accelerating rate cuts" did not materialize.
Several key points to note from this data:
① Core PCE year-on-year at 3.3%, unchanged from last month
Core inflation did not further decline. Although it did not significantly exceed expectations, the logic of "inflation continuously falling" has been interrupted again. PCE remains significantly above the Fed's 2% target, indicating inflation stickiness still exists (Reuters).
② Overall PCE year-on-year rose to 3.7%
Up from the previous 3.6%, also higher than the prior market expectation of 3.6%. On a month-on-month basis, July PCE rose 0.2%, and core PCE also rose 0.2%, showing no clear signs of inflation cooling (Reuters).
③ Consumer side remains resilient
Personal consumption expenditures in July grew 0.2% month-on-month, and personal income grew 0.4%, indicating US household demand has not significantly slowed. In other words, the current US economy is not a simple "high inflation + recession" scenario but more like an economy with resilience and stubborn inflation (Economic Analysis Bureau).
④ Durable goods orders also clearly stronger than expected
US July durable goods orders rose 1.1% month-on-month, above the market expectation of 0.5%, and excluding transportation equipment still grew 0.4%. Manufacturing and durable goods demand are not weak, further weakening the market's imagination of "rapid economic cooling forcing the Fed to quickly pivot to easing" (The Wall Street Journal).
In short:
The US economy shows no clear recession, consumption remains resilient; inflation has not continued to fall rapidly.
For the Fed, this is not a data combination particularly conducive to "rapid rate cuts."
The market previously expected:
Inflation cooling → rate cut expectations rise → US Treasury yields fall → USD weakens → gold, BTC and other risk assets benefit
But what we see now is closer to:
Inflation stickiness remains + economic resilience persists → Fed's rate cut space is limited
Therefore, short-term market rate cut expectations will naturally be suppressed. Data shows that after the PCE release, the market's pricing probability for a Fed rate hike in September has clearly increased (Reuters).
For gold:
This data is overall bearish for gold.
The reason is simple: inflation has not clearly cooled, the economy has not clearly weakened, and the Fed has no sufficient reason to quickly pivot to easing in the short term.
Therefore, gold is more likely to face three pressures in the short term:
USD strengthening + US Treasury yields rising + rate cut expectations cooling.
So currently, gold is more suited to guarding against high-level volatility and pullback risks.
Do not simply interpret short-term price action as "continuous one-way rise after data release" just because the medium- to long-term gold logic remains strong.
Short term is more likely to be volatile digestion, waiting for the Fed to further release policy signals.
For BTC and ETH:
For $BTC and $ETH, this data is also not a major positive.
One of the macro catalysts the crypto market most needs now is:
The Fed pivoting back to easing and improved liquidity expectations.
But this PCE data does not provide such a signal.
On the contrary, sustained high inflation and resilient consumption will make the market continue to wait for more economic data to confirm whether the Fed has room to cut rates.
So in the short term:
BTC may not immediately plunge, but relying solely on macro logic to sustain a continuous one-way rally is clearly more difficult.
ETH is the same.
More likely is:
Rally → profit-taking → pullback → rebound → continued volatile consolidation.
Especially with Fed Chair Wash about to deliver the Jackson Hole speech, funds are unlikely to bet heavily on direction prematurely.
The real main event is still ahead.
PCE is just the first test.
Next, the market is really waiting for the Jackson Hole symposium and Fed Chair Kevin Wash's speech.
This year's Jackson Hole symposium will be held from August 27 to 29, with Wash expected to speak at 10 PM Beijing time on August 28. This is his first Jackson Hole keynote speech as Fed Chair (EBC Financial Group).
So what the market really cares about next is not whether Wash simply says "rate cut" or "no rate cut," but:
How will he define the current inflation environment?
If Wash signals a hawkish tone, emphasizing inflation remains the main risk and even hints at maintaining high rates in the short term, then:
The USD and US Treasury yields may continue to be supported, and gold, BTC, ETH will face short-term pressure.
Conversely, if Wash views current inflation as more transitory, expresses concerns about economic growth and employment, and leaves room for future rate cuts, then the market may reprice easing expectations.
At that time, gold, BTC, and ETH will have a chance to receive truly meaningful macro catalysts.
So don't rush to guess the direction now.
PCE has already told us:
Inflation has not clearly cooled, the economy has not clearly slowed, and the logic for rapid rate cuts has been temporarily doused.
Next, it depends on how Wash sets the tone at Jackson Hole.
Gold should guard against short-term pressure and volatility, and BTC, ETH should be cautious of back-and-forth consolidation.
The real direction may only become clearer after Wash's speech.
#美国核心PCE持平上月，沃什杰克逊霍尔讲话如何定调？ Bitcoin's move back toward $80,000 has two distinct supports, but they should not be treated as equally durable. K33's record one-day short squeeze explains much of the initial acceleration, while the subsequent decline in futures open interest suggests that fuel is already fading. The stronger test is whether last week's $1.92B of US spot ETF inflows can offset profit-taking as roughly $6.44B in BTC options expire Aug 28, with positioning concentrated around $75,000-$80,000. My read: holding this range after expiry would be more constructive than the squeeze itself because it would point to demand, not forced covering, carrying the rebound. Not advice, just analysis.
#BTCOptionsExpiryTestBTC rose 23% this week. Previously, the market was suppressed by cautious sentiment, with short positions piling up more and more. As a result, two large-scale liquidations directly shattered expectations: $1.37 billion was liquidated on the 19th, and another $739 million on the 21st. Notably, after squeezing out high leverage, the market did not immediately enter a crazier leverage-adding phase. Perpetual open interest fell back to 284,000 BTC, and the funding rate returned to neutral, indicatiThe more widespread AI becomes, the more important cybersecurity may become. CrowdStrike CEO George Kurtz recently mentioned that in the future, we will not only need to protect computers, servers, and accounts, but also "protect every AI Agent." Behind this statement lies a very real issue in the next phase of AI: when AI truly starts doing things for people, security risks will also be amplified. 1. AI Agents Are Different from Previous Software Older software was mostly "operated by humans," while AI Agents might read emails, call databases, execute code, connect payments, or even complete entire tasks themselves. The more capable they are, the greater their authority. This means that if an Agent is attacked or its permissions go uncontrolled, it may affect not just an account, but also internal data, systems, and business processes. 2. The deeper AI penetrates the enterprise, the more security becomes like infrastructure. In the past, cybersecurity mainly protected terminals, servers, clouds, and identities. But in the future, there will be an additional layer of protection: AI Agents capable of performing tasks themselves. Therefore, AI development may not necessarily weaken the security industry; on the contrary, it may make security demands more complex. Companies like CrowdStrike, Palo Alto, and Cloudflare have future opportunities not just to "use AI for security," but to provide security capabilities for the AI era itself. 3. What insights do investors have for many people nowadays, when they look at AI, the first thing that comes to mind is chips and modelsOil prices have fallen, which looks like a risk cooling down, but it’s more like traders rushing ahead of the “resumption of navigation” expectations.
The US has expanded sanctions on Iran, which should logically tighten the energy market. But after the Strait resumption talks progressed, oil prices first went down, indicating the market is more concerned in the short term about whether the channel can reopen rather than how tough the sanction rhetoric is.
However, I don’t think the risk has been lifted. If any link in the chain—channel, insurance, shipowners, warships, port inspections—is not agreed upon, oil prices could be lifted again. The most annoying thing about the energy market is that it often lets people relax first, then uses a sudden detail to bring inflation expectations back.
This drop looks more like a traffic jam easing a bit, not a full-speed recovery to smooth flow. Don’t be too quick to take the oil price decline as a macroeconomic positive.
#伊阿敲定临时航道，美对伊制裁加码 In the past 24 hours, the crypto market continues to show a very clear characteristic: $BTC is still a variable that drives the general sentiment, but the level of reaction of each altcoin group is completely different. Market data on August 27 showed that $BTC is rising slightly, while $ETH, $BNB and especially $SOL have more pronounced gains; $XRP is going in the opposite direction. This shows an important thing: the rise of BTC does not mean that all altcoins are rising, but the flow of money is diverging according to each ecosystem, narrative, and level of risk. Below#美国核心PCE持平上月，沃什杰克逊霍尔讲话如何定调？
US July Core PCE rose 3.3% year-over-year and 0.2% month-over-month, with Q2 real GDP revision holding at 1.5%. Although inflation has not worsened, it remains significantly above the Fed's 2% target, and the urgency for rate cuts is once again suppressed by stubborn inflation stickiness.
Tonight, all eyes in the market are on Kevin Warsh's heavyweight debut at Jackson Hole:
Policy weight reshuffle: Between slowing employment and sticky inflation, how highly does the new chair prioritize price stability? This is the key gauge to assess whether September will unexpectedly lean hawkish.
Clear signaling or Tai Chi on rate path: The market urgently needs a clear framework—what conditions will trigger further tightening, and what conditions will open the door for rate cuts.
Global asset pricing reset: If the speech fails to ease divergences, long-term US Treasury yields and the dollar index will continue to tug at high levels, while valuation recovery for gold, US tech stocks, and BTC will still face volatile tests.
Do you think Warsh will send a hawkish tightening signal at this annual meeting, or maintain neutrality to stabilize the market?
$BTC $SPX #FederalReserve #PCE #JacksonHole #Warsh #Macroeconomics
Generate an illustration for this tweet Marvell Earnings Preview: Testing the Quality of the AI "Water Seller"
NVIDIA's better-than-expected earnings have ignited market enthusiasm, and now the pressure shifts to Marvell. As a key player in AI network connectivity chips, its performance will verify whether the AI market can expand from the "compute core" to the "peripheral infrastructure."
Ordinary investors only need to focus on three core indicators:
Data Center Revenue Growth: This is Marvell's lifeline. If custom AI chips and optical interconnect chips maintain over 50% sequential growth, it indicates that major companies are aggressively buying "network cables" and "switches," and AI infrastructure is still in an acceleration phase; otherwise, it may signal inventory reduction or cautious sentiment.
Enterprise Networking and Carrier Business: This traditional segment has been a drag in recent quarters. This time, signs of halting the decline and rebounding must be seen; otherwise, even if data center revenue surges, poor overall revenue will cause the stock price to "die in the light." As long as there is no significant negative growth, combined with a data center boom, it is a solid positive.
Gross Margin and Guidance: AI chips have higher gross margins, so an overall margin increase indicates a higher proportion of high-priced AI products. Also, pay attention to management's guidance for the next quarter and whether they dare to give optimistic forecasts.
Strategy: If all three indicators are strong, consider opportunities for catch-up in the optical module and switch supply chains; if only data center is good, a wait-and-see approach is advised; if guidance is conservative, be cautious of a pullback. NVIDIA has proven there is gold in the "mine," Marvell needs to prove that "shovels" and "transport vehicles" can also sell at good prices. Tonight's/early morning's data is the moment of truth!Today's crypto market is in a pretty clear state: many coin groups are adjusting at the same time, from large-cap, Layer-1, DeFi to meme coins. It's worth noting that this is not just a story of a few tokens. When many assets fall together in a short period of time, the market is often reflecting a general change in risk appetite. And the first variable to look at is still $BTC. 🔴 1 - $BTC declined, but the deeper cause lies in the macro Bitcoin is currently under pressure after failing to maintainLast night, Nvidia's earnings report had only one idea:
AI acceleration + storage shortage + price hikes = storage stocks continue to rise.
Nvidia's "cost" is the "profit" of storage manufacturers.
1. Storage prices are about to rise
NVIDIA has notified major clients that AI server prices will increase by more than 15% next year, mainly because storage prices have gone up again, and NVIDIA can only follow suit.
2. Severe storage shortage
Supply bottlenecks will persist at least until fiscal year 2028.
Next-generation rack storage costs soared from 370,000 to 2 million, up 435%, with market share rising to 25%-30%.
Samsung, SK Hynix, and Micron will have already sold out their HBM capacity by 2026.
3. AI is accelerating
Q2 revenue was 96.2 billion, doubling year-on-year, with next quarter's guidance at 108 billion.
Revenue growth guidance for fiscal year 2028 is 70% (analysts expected only 44%) #PCEToJacksonHole #AIMonetizationBroadens #BTCOptionsExpiryTest #黄金ETF大额吸金，避险资金如何重配
Gold and Bitcoin are simultaneously attracting capital. Over the past five trading days, ETFs for both have collectively absorbed more than $7 billion, setting a historical record for the same period. BlackRock's IBIT saw a net inflow of $1.5 billion, and GLD had a net inflow of $3.4 billion. Both sides are buying—not an either-or choice, but both are needed.
These two assets are telling the same story—the US dollar credit is loosening. Ray Dalio directly advised reducing bond holdings, allocating 10% to 15% to gold, and holding a small amount of Bitcoin. The 90-day correlation between Bitcoin and gold has risen to its highest since the pandemic, with capital treating both as aligned tools to hedge fiat credit risk.
Here’s my view. The simultaneous strengthening of gold and Bitcoin is more important than how much each individual asset has risen. Capital flowing simultaneously into non-sovereign assets from both sides indicates the market is pricing in the same thing—the loosening of US dollar credit, and US Treasuries are no longer considered "risk-free assets."
Gold is driven by real interest rates, safe-haven demand, and central bank allocations, while Bitcoin is driven by liquidity, ETF buying, and leverage changes. The driving logics are not exactly the same, but both ultimately point in the same direction—the systemic increase in allocation weight to non-sovereign assets.
Continued synchronized inflows on both sides indicate capital is systemically increasing allocations to non-sovereign assets. If divergence occurs one day, that will be the time to reassess. For now, the direction is very clear.
$BTC $BTC #TGABuybacksVsFiscalRisk
This is not a liquidity injection, but an invisible game between the Treasury and the Federal Reserve.
As the U.S. Treasury accelerates its long-term bond buyback program, as the TGA (Treasury General Account) balance drops from 800 billion to 650 billion, and as the market debates whether this is QE or Treasury-led — what we are witnessing is not a simple liquidity operation, but a struggle for fiscal dominance by the Treasury "secretly easing" during the Fed's rate hike cycle.
The Treasury's buyback program is essentially a "reverse bond issuance" — using cash to repurchase issued long-term government bonds. This lowers long-term yields, effectively amounting to implicit easing. The market is underpricing this because traditional analysis frameworks still view monetary policy as the sole variable.
Implications for the crypto market: If the buyback program maintains a scale of $20-30 billion per month, it equates to injecting about $300 billion liquidity into the market annually — enough to offset part of the Fed's QT impact. This is a structural positive for risk assets, but the market has not fully priced it in yet.
TGA buybacks vs fiscal risk, your judgment —
A. Optimistic about liquidity improvement, increase BTC holdings
B. Wait and see, until buyback scale is clear
C. Go long on long-term government bonds, betting on yield decline
👇 Type the letter in the comments!#GoldVsBTCETFFlows
This is not a safe-haven showdown, but a war of asset allocation between "old money" and "new money."
When the gold ETF (GLD) sees a weekly outflow of $1.2 billion, and the Bitcoin spot ETF (IBIT) sees a weekly inflow of $1.8 billion, and both experience a rare simultaneous price increase but diverging capital flows — what we are witnessing is not a zero-sum asset choice, but a structural shift where "old money" continues buying gold futures, while "new money" begins allocating to Bitcoin through ETFs.
The outflow from GLD does not indicate a bearish view on gold, but rather some investors taking profits at the historical high of $4600. Meanwhile, the inflow into IBIT shows institutions are reallocating part of their "alternative value storage" from gold to Bitcoin — Bitcoin’s liquidity and transparency advantages are gaining recognition from traditional asset allocators.
Implications for the crypto market: If this "gold outflow → BTC inflow" trend continues, it will be the strongest empirical evidence supporting Bitcoin’s "digital gold" narrative. But the key prerequisite is: BTC needs to break above 80K and hold, otherwise the narrative will be temporarily shelved.
Gold vs BTC ETF flows, your call —
A. Follow the capital flow, buy BTC
B. Allocate to both sides to hedge risk
C. Bet on gold’s pullback, short GLD
👇 Type your letter in the comments!#HormuzFlowsVsSanctions
It's not oil price volatility, but the tug-of-war between the “Hormuz Premium” and “Iran Sanctions.”
When Iran conducts live-fire military drills in the Strait of Hormuz, when the U.S. announces a new round of sanctions on Iranian individuals and entities, and when the market repeatedly prices between “sanctions driving up oil prices” and “strait blockade threat premium” — what we are witnessing is not just geopolitical news, but the implied volatility of crude oil options approaching 100%.
The Strait of Hormuz carries about 20% of the world's oil shipments. Iran's military drills are seen as an “escalation of bargaining chips” in negotiation strategies, but the market is more concerned about an “accidental misfire” during the drills. Brent crude oil fluctuates between $105-$110, and the “tail risk” priced by the options market is rising.
Transmission path to the crypto market: rising oil prices suppress risk assets through inflation expectations, but crypto’s “safe-haven narrative” may temporarily outperform traditional risk assets during geopolitical risk escalations. Watch whether the short-term negative correlation between BTC and oil prices will break.
Hormuz vs Iran sanctions, your judgment —
A. Buy crude oil futures to hedge inflation risk
B. Hold BTC as a geopolitical safe-haven asset
C. Wait and see, do not bet on geopolitics
👇 Type the letter in the comments!🤖 SoftBank Isn’t Just Buying a Robot Company — It’s Betting on Physical AI. SoftBank is reportedly in talks to acquire a majority stake in 1X Technologies at a valuation of around $6B. What makes this interesting? 1X reportedly received 10,000+ NEO orders, yet hasn’t delivered a single robot to customers yet. That tells you the market is already pricing in the future of humanoid robotics, not current revenue. The bigger signal is SoftBank itself. After its $5.4B ABB robotics deal, this move wouTonight's US core PCE data was fully released, showing overall resilience beyond market expectations and prompting a reassessment of the upcoming rate cut pace. Core inflation did not continue to fall as some hoped, but remained flat compared to last month, while personal consumption expenditures and durable goods orders both slightly beat expectations. Putting these numbers together sends a very clear signal: the intrinsic momentum of the US economy remains, inflation cooling has temporarily paused, and market optimism about rapid rate cuts has naturally been dashed with cold water. For risk assets, this is not exactly good news. The delay in rate cut expectations means the timing of liquidity easing has been extended, making it difficult to expect a one-sided market rally in the short term. In such a macro environment, Bitcoin and Ethereum tend to maintain range-bound fluctuations, repeatedly testing the upper and lower levels rather than finding a clear direction. For ordinary investors, this stage tests patience the most, and it is also the biggest taboo to chase gains and sell losses. The situation facing gold is even more direct. As a product highly sensitive to real interest rates, a decline in rate cut expectations usually means higher holding costs, making it difficult for gold to gain strong upward momentum in the short term. A more likely scenario is pressure consolidation, accompanied by increased volatility. If you blindly enter the market at this time, you may face considerable psychological pressure; it's better to wait and see for clearer macro signals. Of course, data from a single month may not change the long-term trend, but it does remind the market: the last mile of inflation is often more winding than imagined. Next, market attention will turn to the Jackson Hole meeting and Federal Reserve officials$SUI direct conclusion: $SUI continues to decline steadily; it's not a problem with the chain itself. The core issue is the token unlocking selling pressure combined with tightening macro liquidity. No matter how good the technology is, it can't stop the continuous supply of tokens.
The primary long-term weakness of $SUI is the monthly linear unlocking, with early investors and team tokens continuously released every month. The selling pressure is constant, and the buying side struggles to keep up, so the selling pressure will persist for a long time. There is another unlocking window in September.
Secondly, PCE inflation exceeded expectations, delaying rate cut expectations, and the overall altcoin environment is relatively cold. A large amount of capital in the public chain sector is flowing to the SOL ecosystem. Although $SUI has a solid Move technology foundation, its ecosystem lacks blockbuster applications, and on-chain funds continue to shrink, with narrative heat fading.
From a technical perspective, there is a buildup of trapped positions above. Every small rebound triggers selling pressure from those unlocking tokens, resulting in poor rebound sustainability. Until Bitcoin breaks out of its range, mid-sized public chains will find it difficult to have independent rallies.
In terms of trading, don't blindly bottom-fish just because the price has dropped a lot. Focus on whether there is a volume breakout above key resistance. Without a breakout, the risk of steady decline remains.#财报观察员：英伟达超预期，软件收入开始兑现 Nvidia's earnings blow up the scene! But don't just get excited, this is the "real signal" ordinary people should watch
Folks, Nvidia's Q2 revenue doubled directly, and it rarely forecasted a 70% increase again in fiscal 2028! The data center business is soaring wildly, and AI computing power demand simply can't stop. But note, it even said "delivery is limited by supply capacity" — it's not that no one is buying, it's that they can't produce enough!
The more critical signal is hidden on the software side: Salesforce's AI product annual revenue is approaching $4 billion, CrowdStrike and Synopsys have both raised their expectations. What does this mean? AI is no longer "burning money telling stories," but turning into real cash! The market evaluation logic has changed: no longer looking at who shouts the loudest, but who can turn AI into orders, renewals, and cash flow.
For us ordinary investors, don't just watch whether Nvidia's stock price surges or not, watch whether the "AI monetization chain" is complete. Chips are the starting point, but software, security, and design tools are the key links where profits settle. Marvell is about to release earnings, can the network connection link take over? This will be the touchstone to test whether AI growth can run through the entire industry chain.
Risk warning:
High growth comes with high volatility, do not chase highs or panic sell. Current valuations already reflect optimistic expectations, it is recommended to build positions in batches, control position size, and add more only after continuous performance verification.$SOL spot market is currently caught between the valuation premium from long-term consensus upgrades and the yield decline caused by short-term capital shifting to stablecoin payments.
U.S. stock spot ETF funds have surpassed the $1 billion mark, establishing a risk-hedging baseline for institutional positions, while on-chain RWA market capitalization has reached $2.01 billion, demonstrating a reduced reliance of real settled funds on high-risk appetite.
The core variables driving valuation are prioritized as follows: Alpenglow mainnet deployment progress, high-frequency DeFi liquidity depth, and the transmission of short-term capital risk appetite.
The bullish scenario triggers during the deployment of the Alpenglow consensus upgrade in Q3 to Q4 2026. If transaction confirmation time shortens from 12.8 seconds to 100-150 milliseconds, and block delay remains within the 200 milliseconds required by the Firedancer client, the price will re-anchor upward to the institutionally estimated range of 210 to 250.
The invalidation signal for this upward scenario is that before the consensus upgrade lands, if the mainnet experiences technical failures or high-frequency ecosystem funds redeem early, the long position cluster will be rapidly squeezed out.
The bearish scenario triggers under tightened macro risk appetite and a delayed announcement of the Alpenglow upgrade. At this time, the low-yield mode of stablecoin transfers cannot support the current risk premium, and long positions may be liquidated down to the 60 range.
The invalidation signal for the bearish scenario is that even if macro inflation expectations fluctuate, if on-chain high-frequency DeFi trading volume breaks through against the trend, it will directly erase the valuation discount caused by technical delays.
In the next 7 days, close attention should be paid to the net inflow scale of U.S. stock spot ETFs and the hedging changes in the daily settlement volume of on-chain stablecoins.
#Anthropic估算30万亿美元市场，IPO叙事能否兑现？ #Meta巨额和解后股价走高，风险定价重估 #JaneStreet持有闪迪5%，AI存储估值再受审视Hello everyone, I am Ruyi
Anthropic has thrown out a $30 trillion story. What does $30 trillion mean? It's close to the entire annual GDP of the United States, equivalent to a quarter of the global GDP. It's even larger than the $28.5 trillion TAM when SpaceX went public, and the combined revenue of 191 tech companies in the S&P 1500 index last year was only $2.4 trillion.
But whether this $30 trillion pie can actually be eaten depends on three calculations by the market:
First, can computing power costs be reduced? The capital expenditure for large model training, chip procurement, and computing infrastructure is huge. Wall Street's bet is that as model training and inference efficiency improve, the proportion of computing power and labor costs in total revenue will gradually decrease. But this assumption requires continuous technological breakthroughs to be validated.
Second, can customers continue to pay? The $30 trillion TAM is based on the assumption that AI will replace human work. However, as more companies adopt industry-specific models built on cheap open-source systems, there is great uncertainty about how much market share Anthropic can capture.
The impact on BTC is indirect but cannot be ignored. The market is already discussing the risk of funds flowing from Bitcoin to AI IPOs. Bitwise advisor Jeff Park bluntly stated that Bitcoin is being used as a source of funds to support popular deals like SpaceX and Anthropic that "everyone suddenly must have." #美国核心PCE持平上月，沃什杰克逊霍尔讲话如何定调？ @币圈超短王马大帅 Volatility and Sharpe Ratio
$BTC volatility is 21.5%, Sharpe ratio 1.92, comfortably profitable on a risk-adjusted basis. $ETH volatility is 44.5%, Sharpe -1.10, meaning it's a losing effort with negative risk premium. July core PCE at 3.3% met expectations, Fed rate hike expectations cooled, liquidity environment favors high-volatility assets, but $ETH has not converted fundamentals into returns. $BTC's low volatility plus positive returns indicate institutional investors holding steady; $BTC ETF net inflows exceeded 2 billion U in August as evidence. $ETH's high volatility shows retail investors are speculating but no trend consensus formed yet.
Capital Attraction Comparison
$ETH cumulative net inflows 729 million U, $BTC only 485 million U, $ETH attracted 50% more capital. $ETH ETF recorded the largest single-day inflow since October. In terms of open interest (OI), $ETH rose from 5.4 billion U to 6.1 billion U, up 13%, while $BTC rose from 7.9 billion U to 8.3 billion U, only up 5%. Smart money is betting on $ETH's recovery rally. However, $BTC's OI has been net outflow in the past two days, $ETH outflow yesterday but inflow of 284 million U today, frequent capital turnover indicates large long-short divergence. Fee-wise, $BTC dropped to 0.0046% with bulls retreating, $ETH returned to 0.01% standard level, more neutral. 今日市场主线：Nvidia 财报明显超预期，AI 与半导体重新成为风险偏好主线；但美国 PCE 仍偏热，Fed 政策风险没有解除。BTC/ETH 维持高位整理，黄金重新反弹，原油继续受中东谈判预期压制。 亚洲股市今天连续第三日上涨，Nvidia 财报后的盘后涨幅一度约 4.7%；与此同时美元指数约 99.12，市场正在等待 Jackson Hole 的政策信号。 ① BTC／ETH：高位整理，资金回流仍提供支撑 已确认事实： BTC 今天大致围绕 79,000 美元震荡。最新市场数据约 78,950 美元附近；近期 BTC 现货 ETF 已重新出现连续资金流入，过去一轮连续流入累计接近 20 亿美元，对这轮 BTC 从低位反弹构成重要资金支撑。 ETH 同样维持近期急涨后的高位结构。从相对强弱看，ETH 在 BTC 横盘阶段仍具有一定补涨特征。 分析判断：BTC 🟠 震荡偏多；ETH 🟢 相对偏强。 BTC 第一支撑继续看 77,600–78,000，强支撑 76,500–77,000；上方先看 79,500–80,000，真正突破确认区仍在 80,000–81,200。 ETH BTC surged and then retreated, and option expirations further disrupted the market
In the past couple of days, the Bitcoin market has surged and then fallen. Besides profit-taking at high levels, another factor has amplified volatility—the concentrated expiration of options. Right now, bulls and bears are fiercely squeezing at key levels, and the market is being pulled back and forth.
Simply put, options themselves can't directly push or dump the market, but they act like an amplifier. A large number of contracts are settled immediately, with huge strike chips piled up at the 75,000 and 80,000 levels. Market makers have to keep adjusting positions to hedge, and as soon as the price approaches these levels, there are many spikes and back-and-forth sweeps. Right now, the coin price is stuck in the middle, pushing up 80,000 is tightly pressed down, and 77,500 is the key #PCEToJacksonHole #AIMonetizationBroadens #BTCOptionsExpiryTest The recent gains have been pretty good, up 58.09% in the past week, with the current price at $0.14+
However, the fiercer the rise, the more you need to watch the moves of the big players.
The venture capital firm Hack VC, this shameless associated address, is suspected of selling $3 million $ENA through Wintermute
4 hours ago Address 0x2a5... 590CF transferred 21.85 million ENA to the Wintermute deposit address through multiple transfers
#PCEToJacksonHole #AIMonetizationBroadens #BTCOptionsExpiryTest Nvidia truly remains the strict father of AI stocks, leading a group of smaller players in a collective rebound after the market closed
But if you look closely, the initial drop after the earnings came out was the first to turn around, until CFO Kress said during the call that FY2028 revenue is expected to grow by 70%.
Comparing the financial data with previous expectations, it did exceed, but there was no extreme positive news:
Revenue was $96.2 billion, compared to market expectations of about $92.2 billion;
Adjusted EPS of $2.22, expected to be around $2.10;
Data center revenue was $89 billion, compared to an expected $85.1 billion;
Next quarter's revenue guidance is $108 billion, exceeding the publicly available consensus estimate of about $104.2 billion, and does not include any Chinese data center computational revenue #PCEToJacksonHole #AIMonetizationBroadens #BTCOptionsExpiryTest NVIDIA's Q2 FY2027 earnings once again exceeded expectations: revenue reached $96.2 billion, more than doubling year-over-year; adjusted EPS was $2.22, also beating Wall Street.
Data center revenue hit $89 billion, accounting for over 90% of total revenue. The key point is not just that chips are still selling, but that the software layer is starting to convert installed base into revenue.
Jensen Huang's exact words nailed the logic: AI has reached an inflection point, tokens are already profitable, and computing power equals revenue. The market's past concern was the "shovel-selling cycle peaking"; this earnings report provides counter-evidence that after the shovels are sold, the platform continues to take commissions.
ACIE generated $40.3 billion in a single quarter, soaring 138% year-over-year and up 25% quarter-over-quarter, with growth clearly outpacing hyperscale customers. This is the entry point for software monetization—enterprise deployment, sovereign cloud, NeoCloud; customers are buying not just GPUs, but the entire stack including continuous CUDA optimization, Nemotron models, and the Agent toolchain.
A stronger signal from the earnings call is the opening of distribution channels: AWS is deploying an additional 2 million GPUs and integrating Nemotron into Bedrock and SageMaker; Amazon's warehouse robots will run the full suite of physical AI software including Omniverse, Cosmos, Isaac, and Jetson. Software is no longer just a chip bundling bonus; it is starting to follow cloud and production lines through subscriptions and API calls.
#财报观察员：英伟达超预期，软件收入开始兑现 Just glancing at the market, BTC is hovering around $79,000. The August 27 batch of coins performed quite impressively, with several lines worth paying close attention to. $HYPE the real center today, breaking through $83.6 to set a new all-time high, up about 3% in 24 hours. On-chain trading is quite interesting—the suspected a16z institution dumped another 36 million USDC in the past 24 hours to buy HYPE, with a total holding of 4.679 million tokens (about $381 million), an average cost of $65.6, and an unrealized profit of about $74.4 million. On the other side, a whale sold 301,900 HYPE today, cashing out $24.4 million and making over $5.3 million in profit before exiting. Bulls and bears are facing off at the $80 level, with resistance at $82.46 on the upper band and support at $77.37 on the 4-hour chart. Additionally, Hyperliquid just launched AQAv2 on August 26, with about 90% of its revenue going to buy back and burn HYPE, expected to generate $135–$200 million in buying funds annually. The first buyback will only be confirmed on October 3. $TRUMP dropped about 4.5% today, currently around $2.28. On-chain data is not looking good. The largest investor Zhou was recently reported to have six dishonest cases in China involving tens of millions of RMB, and there is also a money laundering case under trial in the UK. Although he spent $100 million on WLFI tokens, this negative news definitely suppresses short-term sentiment. $SNDK closed at $1,499.37, up 1.26%. YearCHIPUSDT: Review of USD.AI narrative, actual business, and meme coin potential.
USD.AI is not about "issuing an AI coin," but about building a computing power asset credit market:
1. Users mint USDai with stablecoins, then enter sUSDai to earn yields;
2. AI cloud providers and GPU operators apply for loans using installed, revenue-generating GPU equipment;
3. The CALIBER framework maps hardware ownership and disposal rights into on-chain certificates, configuring liens, insurance, custody, and default handling;
4. Loan interest, issuance fees, and reserve asset income flow into the protocol and sUSDai yield system;
5. CHIP governs loan admission, interest rates, risk parameters, fee routing, and insurance collateral modules.
Officially, loans typically use a maximum LTV of about 75%-80%, amortized over roughly three years, prioritizing GPUs that are already installed and capable of generating income; CALIBER attempts to make on-chain certificates correspond to enforceable physical property rights. This design addresses the mismatch between traditional bank approval cycles and rapid GPU depreciation, rather than simply betting on GPU price appreciation. #财报观察员：英伟达超预期，软件收入开始兑现
I find today's market movement even more interesting.
After Nvidia's earnings report was released, the market's feedback was very direct: AI demand is still strong, and storage has clearly become a bottleneck. Nvidia's supply chain commitment jumped from $119 billion last quarter to $279 billion, more than doubling, mainly to secure key capacities including memory.
So it's not surprising that Micron, SanDisk, and Hynix all saw varying degrees of gains today. Especially since Nvidia itself emphasized tight memory supply, this serves as a strong fundamental confirmation for the storage sector.
I now feel that the previous round of adjustments was more about valuation cuts rather than logic cuts. The stock price will definitely fluctuate in the short term, but as AI computing power continues to expand, demand for HBM and DRAM is unlikely to suddenly fade.
So regarding the storage sector, I still say: don't be scared off by short-term volatility; the real story may just be beginning. Brothers, I'm bearish on this wave and shorting as well! $BTC can't hold the C80 level, a downtrend has already formed.
Yesterday, Bitcoin hit a high of $81,237 intraday, breaking the 80k mark for the first time in three months. But it was immediately slammed back down, dropping to around $78,000 at one point.
80k is like an iron wall; even if it breaks through, it might not hold! Today, BTC spot/perpetual contract prices are fluctuating between $78,500 and $79,000, with an early dip to $78,296.
The market data says it all.
In the past ten days, Bitcoin rose from $62,800 to $78,900, a 25% increase, marking the best ten-day gain this year. But within 24 hours, about $324 million in leveraged positions were liquidated, with over $300 million in long positions wiped out.
More importantly—the market sentiment has turned cold. Santiment data shows the weighted sentiment score dropped to -0.023 on August 26, the first negative value since this rally began. Prices are rising, but sentiment is falling; this divergence often signals that upward momentum is fading.
Why do I insist on being bearish?
First, 80k is a strong ceiling. The $80,000-$82,000 range is a key daily supply zone, with the 50-week moving average around $81,000 forming strong resistance. Bitcoin was rejected at $81,000 and pulled back, testing whether the 24% gain is supported by real buying power.
Liquidity is relatively thin between $80,000 and $82,000, so large orders can trigger very volatile moves.
Second, this rally is driven by short squeezes and lacks sustainability.
Glassnode reports that Bitcoin's roughly 26% rebound since mid-August was mainly driven by record short liquidations, with August 19 marking the largest single-day short liquidation since 2019.
During the same period, BTC futures open interest denominated in coin terms dropped 11%, and perpetual contract funding rates remained neutral overall, indicating no significant new leveraged long positions.
Forced buying can quickly push prices up, but every liquidation-driven buy order closes an existing position rather than creating sustained demand. Once shorts are cleared, the market needs new buyers to keep prices rising.
Third, overbought signals are obvious. The Fear & Greed Index has risen to 83 (extreme greed). The daily RSI is in the 78-81 overbought zone. Prices are far above all short-term moving averages, indicating strong demand for a short-term pullback.
If 80k is rejected again, there is room for another correction. Analysts even warn this could be a trap or lead to a crash down to $45,000.
I'm still holding my short position, opened near 78,340, waiting for it to slowly retrace. If 80k can't hold, the end of the party will be a quiet exit.
Brothers, this short is solid!
$ETH
$SOL
#美国核心PCE持平上月，沃什杰克逊霍尔讲话如何定调？ Institutions have already shown their cards, are you following?
First, let's see to what extent the cards have been revealed. The US spot Bitcoin ETF has had a net inflow of $2.26 billion over six consecutive days, $2.72 billion in August alone, the strongest in 2026. Total assets surged from 76.6 billion in mid-August to 98.56 billion, just $1.44 billion shy of 100 billion. On August 25 alone, 4,284 BTC were bought, with a seven-day total of 26,499 BTC.
And the positions of allocation funds will eventually settle into long-term holders, so today let's see where this batch of long-term capital stands.
Addresses holding coins for over 155 days are considered long-term holders (LTH), the toughest batch of chips in the entire market who have endured the full downturn and haven't sold. Their average cost is now about 50,000, current price is 78,700, with an overall unrealized profit of over 50%.
MVRV rising to 1.25 to 1.5 standard deviations marks historical tops; it reached 2.2 in December 2017 and 1.5 in April 2021. Falling below -1 marks historical bottoms, seen near the green line in 2015, March 2020, and November 2022. This June, this line broke below -1 touching the green line, then reversed upward in the past two weeks, now back to -0.85.
The toughest chips still have a cost basis at 50,000, while the price is already 78,700. Do you understand now? monetization is starting to look less like a single-layer chip story and more like a widening enterprise stack. NVIDIA doubled Q2 revenue year over year and projected about 70% growth for FY2028, even as supply constrained deliveries. Meanwhile, Salesforce AI ARR neared $4B, CrowdStrike delivered record net new ARR, and Synopsys raised its outlook, while slower Okta orders show the trend is not universal.
My read: the strongest signal is not headline growth alone, but AI夜盘盯完最后一眼K线，我默默把TRUMP的仓位截图删了，这波教训值得记进小本本。 你们有没有那种时刻，明明看懂了市场在交易什么，却还是忍不住手痒？ 今天不聊宏大叙事，就当作一本风险管理日记来写。我花了整个周末把手里关注的几个币和美股存储板块重新捋了一遍，发现一个很微妙的规律：市场的钱其实一直在板块之间做选择题，而不是无脑撒钱。 先说结论性的观察，这轮强弱排序非常清晰： - 风险从低到高大概是 OKB 最稳，SOL 次之，ZEC 因为ETF传闻被推得很高但属于情绪透支，BICO 中规中矩，HYPE 完全跟着AI Agent的呼吸走，TRUMP 就是纯情绪筹码，跟基本面没有半点关系。 - 如果Nvidia财报是今晚的暴风眼，受影响最大的其实是 HYPE 和 SOL，因为它们身上的AI叙事和风险偏好标签最重。ZEC 和 BICO 会被情绪扫到，但故事内核不会变。TRUMP 和 OKB 几乎是绝缘体，各自活在独立宇宙里。 我真正想说的是板块强弱这件事，它比价格涨跌更早暴露聪明钱的态度。 你看美股存储三兄弟，SanDisk、Micron、SK Hynix，它们的日成交额加起来能吓死人，但走势却极Tether Gold (XAUT/USDT) trades on OKX at $XAUT $4,607.4 (-0.06%), slipping below key moving average levels as it tests support just above local low levels: MA5 ($4,611.9), MA20 ($4,611.9), and MA10 ($XAUT $4,615.1).
Reclaiming $4,611.9 is required to halt immediate downside momentum and launch a recovery toward upper resistance at $4,615.1 and local peak resistance at $4,628.4.
#PCEToJacksonHole #AIMonetizationBroadens #OKX.ai $ETH current market analysis. Overall, $ETH is currently in a strong rebound driven jointly by a "short squeeze" and institutional capital inflows. The short-term technical outlook is bullish but approaching a key resistance zone.
📊 Market status: A battle between short-term strength and resistance
After fluctuating around 2,498-$2,507, the short-term technical signals are as follows:
· Short-term trend is bullish: Daily moving averages show a bullish alignment, ADX trend indicator is about 46.7, indicating a strong upward trend. Technical indicators show prices at 2,536 and $2,553**.
· ⚠️ Overheat and resistance signals: RSI has reached 77.4 (>70 is overbought), indicating a risk of short-term pullback or high-level consolidation. Additionally, the 2,550 range forms a strong resistance band, with a "sell wall" from whales around the 2,650 area.
· Key support: The first support zone below is at 2,400, with stronger support near $2,370 (a break below may trigger massive long liquidations).
🔍 Driving forces behind: More than just a technical rebound
There is deeper logic behind this rally:
· Short squeeze: The market was extremely bearish before, with crowded short positions. As the price rose, shorts were forced to cover (buy back), with $1.13 billion worth of $ETH-related short liquidations in just one day, intensifying the rally.
· Institutional capital inflows: Spot ETFs are the main engine of the rise. On August 26 alone, the US Ethereum spot ETF had a net inflow of 75,150 ETH (about $184 million), providing sustained buying support to the market.
· Fundamental narrative shift: Network activity has not died down; about $165 billion stablecoins circulate on Ethereum, and L2 network fees have dropped below 2 cents. The ecosystem development logic is shifting from "low fees mean failure" to "low fees promote growth."
📈 Mid-to-long-term perspective: hurdles still to overcome
Although short-term sentiment is hot, analysts' warnings are worth noting: macroeconomic headwinds (such as US stock pullbacks and PCE inflation data) remain a looming threat that could trigger market corrections.
In short, $ETH is at a critical crossroads now: if it can break out with volume beyond 2,823, there is room to advance; if it is resisted and falls back, it may retest support near $2,400. US July core PCE year-on-year 3.3%, month-on-month 0.2%, unchanged for two consecutive months, fully in line with expectations; overall PCE year-on-year 3.7%, unchanged, slightly above expectations.
Inflation has neither continued to decline nor suddenly worsened — this is a "neutral to sticky" report card.
Core PCE has been above the Fed's 2% target for 65 consecutive months. The services side still contributes stickiness, and the decline in goods prices has not fully offset this. Real consumer spending is almost stagnant, yet income continues to grow; neither supply nor demand sides provide evidence that "inflation is dead."
This data comes right on the eve of Jackson Hole. New Chair Kevin Walsh will deliver his first keynote speech since taking office at 10 a.m. Eastern Time on Friday. The nominal theme is "The Impact of Financial Innovation on Payments and Policy," but the market is really focused on the interest rate guidance.
Walsh has deliberately given little forward guidance so far, emphasizing "framing big issues, feeding fewer answers." After the July rate decision, three dissenting votes called for a rate hike, and long-term US Treasury yields rose; the market is growing somewhat impatient with his communication.
The most likely path for this speech is: reaffirming the 2% target has no flexibility, policy remains restrictive, and the option to raise rates is not closed — rebuilding credibility with principled statements rather than directly previewing a rate hike or cut in September.
#美国核心PCE持平上月，沃什杰克逊霍尔讲话如何定调？ Do you know how much money $UNI has made since its launch?
UNI has been the top-ranked DEX in the industry for many years, accumulating fees totaling $5.79 billion. Among them:
1. $ETH contributed $4.48 billion
2. BASE chain contributed $591 million
3. $ARB contributed $316 million
4. Robin Hood contributed $107 million
So why has the UNI price been continuously declining, from a high of $44 to around $4 now?
The total supply of UNI tokens is 1 billion, with 109 million tokens burned so far, leaving a total supply of 890 million; the DAO treasury holds 267 million tokens, and the actual circulating market supply is 623 million.
Of the above $5.79 billion revenue, over 90% has been distributed to LPs. Currently, only about 2% of the revenue is used for buyback and burn.
Compared to HYPE and pump projects, UNI is considered a good company, but the current buyback for shareholders is still too low.