Orbit Post Sitemap
📊 The overall market is broadly down, but MNT bucks the trend with a +5% gain. Last week, Mantle deployed the "Super Portal" to Solana via the Chainlink CCIP bridge—L2 assets now directly access the Solana ecosystem, heating up the cross-chain liquidity narrative. Coupled with Bybit's deep integration, demand on the exchange side is strengthening.
⚡ Analysis: The cross-chain narrative has short-term momentum, and MNT is indeed one of the few L2s with actual output. However, a $1.5B market cap is not cheap in this sector, so chasing the price higher requires setting proper stop-losses.
💭 Is the ultimate fate of L2s a unified multi-chain or fragmented chains?
$SOL $LINK U.S. stock market opening forecast tonight (August 11, Beijing time 21:30) $BTC $ETH $SNDK
Core background: Tonight is just one day before the CPI data; the heavy inflation data will be released tomorrow night at 20:30. Global funds are generally proactively reducing positions and waiting to observe, making it difficult to have a sustained one-sided big market; additionally, South Korea's SK Hynix weakened with volatility today, putting pressure on the sentiment in the memory sector.
1. Overall market outlook
High probability of a slight gap down / flat open with narrow fluctuations
Full-day characteristics: range tug-of-war, severe sector divergence. Funds are unwilling to heavily bet on direction, trading volume shrinks compared to usual; all fluctuations will remain restrained, with main funds holding back to wait for tomorrow's CPI results.
The Nasdaq and S&P heavyweight tech giants (Microsoft, Amazon, Google) show stronger resilience; semiconductors and memory sectors have greater volatility and are the main battlegrounds on the market.
2. Key sectors | Semiconductors & Memory (your focus)
Yesterday, the Philadelphia Semiconductor Index dropped sharply by 2.94%, with profit-taking pressure at high levels still present; combined with SK Hynix weakening during the Asia-Pacific session, sentiment formed a negative transmission.
1. Micron MU
Short-term resistance: $895; short-term support: $870, strong support at $854
Opening scenario: high probability of a slight gap down testing support.
✅Key signal: Holding $854 maintains high-level oscillation; a valid break below $854 opens short-term correction space.
2. SOXL (3x Semiconductor Bull ETF)
Follows the Philadelphia Semiconductor Index, with volatility amplified by leverage; leveraged ETFs have higher risk of loss in volatile markets and are not suitable for long-term holding or speculation.
3. SanDisk SNDK
Closed up against the trend yesterday, relatively stronger than the memory sector, indicating fund divergence and a higher probability of an independent trend.
3. Three intraday scenario simulations
Scenario 1: Weak oscillation (highest probability)
Slight dip at open, insufficient rebound strength, oscillating downward all day.
Cause: Funds preemptively hedging, memory sector continuously digesting previous large gains;
Effect: U.S. stock market closes weak tonight, further suppressing tomorrow's South Korean market and SK Hynix opening sentiment.
Scenario 2: Range-bound oscillation recovery (medium probability)
After testing support, stabilizes and rebounds, fluctuating back and forth within the range all day with no clear direction.
Feature: Neither rise nor fall can effectively break high or low points; trading is very difficult, chasing highs and selling lows easily leads to repeated losses.
Scenario 3: Strong one-sided rebound (low probability)
Trigger condition: positive news appears, funds preemptively speculate on favorable CPI expectations.
Premise: Memory sector rallies with volume, recovering most of yesterday's losses; if volume is low during the rise, it will likely spike and then fall back.
4. Very important trading discipline
1. Do not hold heavy positions overnight! The decisive event is the CPI at 20:30 tomorrow night; data can easily cause extreme gaps; leveraged instruments multiply risk.
2. Distinguish between "short-term sentiment fluctuations" and trends: tonight's market is more about funds hedging and trading, not changing the medium-term major trend; the real direction awaits CPI confirmation.
3. Cross-market linkage closed loop: tonight's U.S. market close directly determines the opening tone of South Korea's SK Hynix and KR200 tomorrow morning.
5. Key monitoring time points
21:30 market open, the 30 minutes before open form high and low points, serving as the intraday watershed;
Focus on whether Micron can hold the $870 support and whether the Philadelphia Semiconductor Index can stop falling and stabilize.
After the CPI data is released at 20:30 tomorrow, I will interpret it immediately and simultaneously simulate a complete response plan for memory and Korean stocks. #本周三CPI公布，9月加息定价会改写吗？ #现货ETF资金分化，BTC卖压仍在 #财报观察员：AI基建财报接力登场 #FinancialReportObserver: AI Infrastructure Earnings Reports Take the Stage
The first batch of SpaceX restricted shares were unlocked, and the market was originally expecting a sell-off, but it didn't happen. Instead, the stock price rebounded consecutively, returning close to the IPO price. Short sellers covered, the float increased, and sentiment eased suddenly. But don't celebrate too early—there's still about a 7% unlock window on August 20, so the selling pressure afterward needs to be watched.
Tougher tests lie ahead. Tonight, Lumentum and CoreWeave will release earnings after the market closes; on the 12th, Coherent; on the 13th, Applied Materials. Cisco is also on the watchlist. Optical communications, computing power cloud, semiconductor equipment, enterprise networks—this lineup perfectly covers the upstream and downstream of AI investments.
My view is: the story has been told long enough; now it's time to verify. Are the orders really landing? Can the revenue be realized? Can profits support the high valuation? Many domestic optical module companies have already "turned positive" in their quarterly reports, and overseas markets are waiting for the same answers.
Earnings season is never just about the current quarter's numbers; guidance and capital expenditure rhythms are also key. After reading these reports this week, we can roughly sense whether AI infrastructure will continue to sprint or start to catch its breath.
My own stance: don't chase sentiment, just wait for the numbers. If there is real strong growth, it's not too late to get on board. $BTC $ETH #闪迪8月13日投资者日临近，财报分歧待解
Last earnings report exceeded expectations but caused a big drop, trapping many investors
I still hold over 1300 long positions myself, and it's not easy to stay calm.
The core market disagreement isn't about the current earnings, but whether AI storage demand has peaked short-term or is just slowing down. The previous surge pushed expectations too high, so any slight guidance miss triggers a sell-off.
The August 13 investor day is quite critical; it depends on whether management dares to provide clear long-term order guidance. I won't blindly add to my position to lower cost now, nor do I plan to sell off casually. I'll wait for solid signals from this meeting before making moves.
$SNDK Several key signals in today's market highlight the focus points.
BTC is oscillating around 64,000, with a total market cap of 2.27 trillion, down 1.55%, while ETH has risen 2.12% to 1874. Gold has surpassed $4414, up over 7% for the week, as capital is repricing fiat currency credit. Crude oil rebounded to $82, with the Strait of Hormuz blockade causing geopolitical risk premiums to re-enter. The market is waiting for Wednesday's CPI, and the market has already entered a sideways consolidation mode.
The liquidation map shows BTC currently at 64,037, right in the middle position. Below, 63,300 to 63,700 is a dense liquidation zone for longs; breaking below 63,500 could trigger a chain reaction of long liquidations. Above, 64,300 to 65,000 is the short liquidation pressure zone; breaking above 64,400 could trigger a short squeeze. Both bulls and bears are waiting for the CPI data to provide direction; no one wants to make the first move before the data is released.
Geopolitical news has become tense again. Trump stated that the US military has 100% control over the Strait of Hormuz, with the blockade acting like a steel wall, open to non-Iranian vessels but not allowing entry into Iranian ports, while demanding Iran compensate for casualties and damages caused by the conflict. Iran claims that an agreement with Oman on a new shipping route is near completion, but significant differences remain between the two sides. Oil prices have rebounded from $75 to $82, pushing inflation and geopolitical risk premiums higher again. The Wednesday CPI data combined with the oil price rebound means inflation expectations face dual upward pressure.
After weaker-than-expected nonfarm payroll data, the market has significantly lowered the near-term rate hike probability, but this week's CPI and PPI will be key for the next policy pricing. Oil prices have surged sharply due to the Hormuz stalemate, potentially pushing inflation expectations higher again. The weakening employment data and rising oil prices create a tug-of-war, leaving the Fed's short-term path highly uncertain, with the market very sensitive to data.
US tech stocks are under pressure; Nvidia fell nearly 3%, Apple weakened noticeably, but Microsoft, Amazon, and other cloud and software stocks showed relative resilience and gains. Intel announced a possible $15 billion stock sale to support AI and manufacturing businesses. High-valuation, high-elasticity AI hardware is the first to feel pressure amid rising inflation concerns, with capital favoring cloud and software stocks that have realized revenue and clear commercialization paths. This divergence also exists in the crypto market, as shown by the differing trends of BTC and ETH.
In terms of operations, it is recommended to maintain flexible positions before the CPI data and avoid heavy bets on direction. BTC is currently at 64,037; long positions at 62,288 should be held with stop-loss moved up below 63,500. The first resistance is between 64,300 and 64,500; a volume breakout and stable hold above this can be an opportunity to add positions, with the first target between 65,000 and 65,500, and a breakthrough target between 66,500 and 67,000. The defense bottom line is between 63,300 and 63,700; if volume breaks below this, exit first and wait for clear direction before re-entering. CPI is the real judgment point. Nonfarm data has already shaken the table halfway; CPI will determine the nature of this market cycle. If CPI is weak, rate hike expectations will continue to fall, and BTC has a high probability of breaking through 65,000 directly. If CPI is strong, rate hike expectations will surge again, and BTC will retest 63,000 to 63,500. Do not act before the data; follow the direction after the data is released. Set stop-loss properly and act only when direction is clear.
$BTC $ETH $GRVT #财报观察员：AI基建财报接力登场 #本周三CPI公布，9月加息定价会改写吗？ #英伟达推动5000亿美元AI基建融资 Today $LAZR dropped by 4.08%, and almost all of its holdings including $AXTI, $lite, and $cohr experienced heavy declines.
But, let's look at the changes in this ETF's holdings today?
COHR 4,824 → 5,580 increased by 15.7%
AAOI 18,873 → 21,869 increased by 15.9%
LITE 6,855 → 7,933 increased by 15.7%
AXTI 51,672 → 59,890 increased by 15.9%
SIVE 174,169 → 201,875 increased by 15.9%
IQE 1,258,756 → 1,459,012 increased by 15.9%
AEHR 16,741 → 19,401 increased by 15.9%
SOI 6,008 → 6,960 increased by 15.9%
Almost every company we focus on increased their holdings by 15.9%. In other words, the overall holdings increased by 15.9%, which is roughly $5 million added proportionally to each company.
Why, despite the market decline and LAZR itself dropping, can they still continue to increase their positions?
This indicates new capital subscriptions. In other words, for some institutions and large investors, today's pullback is seen as a bottom-fishing signal.On the eve of the CPI release, the entire market enters a wait-and-see window, with US stocks and the crypto space synchronously awaiting the inflation report
📰 Macro Highlights | July CPI is about to be released, a key turning point for all risk assets
At 20:30 Beijing time on Wednesday evening, the US July CPI inflation data will be officially announced. JPMorgan warns that this data could cause the S&P 500 to fluctuate by as much as 2% in a single day. US tech, aerospace, storage sectors, as well as $BTC and $ETH, will all be influenced by this data.
Current market consensus expectations: overall CPI year-on-year at 3.4%, core CPI year-on-year at 2.5%. Recently, Federal Reserve officials have issued hawkish remarks, stating that inflation cannot be taken lightly and not ruling out the possibility of further rate hikes, breaking the market's previous overly optimistic expectations of rate cuts.
The current market situation is interesting: with weaker non-farm payroll data, traders are preemptively betting on easing. BTC-ETF has maintained net inflows for several consecutive days, with weekly inflows exceeding $850 million. BlackRock's IBIT is the main recipient of funds, but prices remain stuck in a range, unable to rise or fall.
The root cause is that funds are reluctant to make large bets on the CPI outcome in advance.
Three scenarios to clarify in advance:
1. Inflation significantly below expectations: rate cut expectations rise again, Nasdaq, storage, aerospace sectors rebound, BTC tests resistance at 65500-66000, ETH will have greater elasticity.
2. Inflation within expected range: likely a spike followed by a pullback, buying the expectation and selling the fact, the market continues to oscillate within existing levels.
3. Inflation rebounds again: US Treasury yields rise, risk assets collectively under pressure, high-valuation growth stocks in US stocks face valuation cuts, BTC and ETH will open downward correction space.
A detail worth noting: the Shiller PE ratio (CAPE) for US stocks has reached 42 times, second only to the internet bubble period. The overall valuation tolerance for US stocks is very low. Once inflation exceeds expectations, the correction pressure on high-valuation sectors will be significant.
Currently, whether trading US individual stocks or mainstream crypto, it is not suitable to heavily bet on direction; keep positions for after the data release. Tuesday, 2026.08.11
US stocks fluctuated, gold rose, Bitcoin independently declined. Bitcoin is now in a completely independent market. After moving up with gold a couple of days ago, it suddenly reversed yesterday, indicating that price increases are still not driven by liquidity but rather by arbitrage and selling pressure.
On August 10, Bitcoin ETFs had a net outflow of 144 million. Ethereum ETFs had a net inflow of 14.6 million.
Now let's look at the most fundamental reason: the emergence of AI data centers has caused a significant electricity shortage. Compared to current computing power and mining output, if mining machines are not upgraded, under current conditions, transforming into AI data centers can yield profits ten times higher. Many US mining companies have already secured orders and started this transformation, so selling pressure from miners on Bitcoin remains very high. However, AI electricity consumption is not limitless. Although there are calls about power shortages now, a balance will eventually be reached. When that balance point arrives, there may be a need to transform back to mining machines. Therefore, miners and mining companies are currently facing a difficult choice: whether to transform, upgrade, or hold on. Regardless of the choice, miners will inevitably continue to sell Bitcoin for some time in the future.
Furthermore, MicroStrategy was revealed yesterday to have sold over 1,600 BTC last week, which may be one reason for Bitcoin's independent decline yesterday. Once MicroStrategy opened the door to selling coins, it became unstoppable, selling this many coins weekly. At this pace, they would sell out in less than five years. However, they certainly won't only sell without buying; a large holder who only accumulated coins has turned into one who sells high and buys low. ETF holdings have also returned to levels from two years ago. So who exactly will drive Bitcoin's price up in the market?
Market Analysis
Bitcoin's independent decline yesterday was indeed unusual. Although it fell, it likely won't drop much further and remains within a volatile range, consistent with our previous view of initial suppression followed by a rise. It's uncertain whether CPI will rise. Currently, 63,500 is a strong support level.
The US stock market's optical communication sector experienced a significant drop, possibly due to some issues with earnings expectations. After earnings reports are released, it will likely rebound, so there's no need to panic. Buying US stocks on dips remains a good strategy.
Crypto Fear & Greed Index: 37 (Fear) If even the world's largest Bitcoin die-hard bears are selling at a loss, can you still hold onto your BTC?
On August 10, Strategy submitted a filing to the SEC.
The data is painful:
From August 3 to 9, they sold 1,690 BTC at an average price of $64,262, cashing out $108.6 million.
All proceeds were used to repurchase STRC preferred shares.
This is not the first time. In six weeks, they sold a total of 6,916 BTC, worth $429 million.
They realized a cumulative loss of about $93.12 million.
What was their cost basis?
$75,385 per BTC.
Now BTC is hovering around $65,000.
They lose $11,000 on every coin sold.
The man who once said "never sell" is now cutting losses.
But the other side of the story is—
Some are selling off heavily, while others are frantically buying.
Strive increased its BTC holdings by 6,236 coins in Q2.
They bought a total of 12,237 BTC in the first half of the year.
Their total holdings rose to 20,167 BTC.
BitMine continued to expand its ETH holdings in July while also repurchasing shares.
One company is selling, two are buying.
The corporate treasury logic is shifting from a single narrative of "only buying, never selling" to a new phase of "increasing holdings, selling, repurchasing, and cash management coexisting."
So why is Strategy selling?
Because they ran out of money.
They posted a net loss of $8.2 billion in Q2.
STRC preferred shares have been trading below the $100 par value, making new debt financing impossible.
Dollar reserves need to be replenished to $4.65 billion.
Preferred stock dividends must be paid, and debt interest must be serviced.
They used to buy BTC by issuing stock; now they sell BTC to pay bills.
At the end of June, the board authorized the sale of up to $1.25 billion in Bitcoin.
So far, only $430 million has been sold.
There is still over $820 million left to sell slowly.
What’s really worth pondering is not how much Strategy lost.
It’s that the narrative of "corporate Bitcoin treasury" itself is being redefined.
What did the market believe before?
"Company buys BTC = bullish on BTC = BTC will rise."
The logic was simple and crude, but effective.
Now?
Companies can buy, sell, repurchase shares, and replenish cash simultaneously.
Buying is a tool, selling is also a tool.
BTC on corporate balance sheets has shifted from a "faith asset" to a "liquidity tool."
This isn’t necessarily bad, but it’s definitely different from before.
To be honest—
Corporate treasuries are evolving from "one-way pumps" to "two-way regulators."
What does this mean for the market?
The good news: 840,447 BTC are still held by Strategy, accounting for 4% of total BTC supply. The bulk remains untouched.
The bad news: If the financing environment continues to deteriorate, part of that 4% will slowly turn into sell pressure in the market.
More importantly: Will other BTC-holding companies follow suit?
Strive is buying, BitMine is buying, but Hyperscale Data is selling, and Trump media is reducing holdings.
The divergence among companies is turning into a long-short game in the market.
Finally, three questions for you:
First, if Strategy keeps selling, will you panic and follow?
Second, if Strive and BitMine keep buying, will you follow to bottom-fish?
Third, when "never sell" turns into "sell when needed," how much faith do you still have in Bitcoin?
$BTC $ETH $SOL #Strategy再卖1690枚BTC，企业财库出现分化 Officially declared completely dead.
Only 0.15% of the total hash power supported it throughout, basically meaning it died right after launch.
The miner team Roughnecks, who led the fork, gave up voluntarily shortly after.
The entire forked chain mined only 2 blocks from start to finish, then completely stalled.
Now this chain is more than a hundred blocks behind the Bitcoin main chain, falling further and further behind, completely unable to catch up.
Bitcoin soft forks require at least 55% of the total network hash power to be valid.
This time it never even exceeded 2.53%, doomed to fail from the start.
Michael Saylor specifically posted a calculation.
This forked chain needs to mine another 2015 blocks to reach its first difficulty adjustment. At the current hash rate and block speed, this process will take 25 years.
Just waiting for the first adjustment will take over twenty years, so there’s no need to discuss what comes after.
The forked chain doesn’t even meet the basic conditions for stable operation, let alone ecosystem and users.
Saylor said something very practical.
Anyone has the right to fork Bitcoin; the network is open, and the main chain can choose not to follow.
But for a forked chain to survive, it must have security, practical value, capital scale, and a user base—missing any one of these means the forked chain is just air.
Bitcoin’s consensus cannot be forcibly declared by a mere proposal.
It requires genuine recognition from miners and users across the network.
Market data also reflects the current attitude: BTC at 65085, ETH at 1919, gold at 4348, mainstream asset trends diverge, and funds have not shown obvious disturbance due to the fork event; the market is already immune to such proposals.
This rapid collapse of BIP-110 taught everyone a lesson.
Bitcoin’s core moat is the consensus built jointly by hash power and users.
Trying to leverage a niche proposal to shake the entire Bitcoin network is simply unrealistic.
Miners vote with hash power, dousing cold water on all schemes that try to fork easily.
The market’s demand for Bitcoin’s underlying stability far outweighs various niche radical reform proposals.
This consensus is truly not something that can be shaken casually.
I’m still holding my positions in BTC, ETH, and $GRVT #财报观察员：AI基建财报接力登场 #本周三CPI公布，9月加息定价会改写吗？ #英伟达推动5000亿美元AI基建融资 According to foreign media reports, Apple has begun verification testing of Changxin Memory's DRAM memory chips, covering iPhone and MacBook product lines. The potential supply target is primarily models sold in mainland China. Currently, it is only in the evaluation and testing phase, and no formal procurement orders have been finalized. Currently, Apple's DRAM supply chain is highly dependent on Micron, Samsung, and SK Hynix. Driven by AI computing power demand, leading memory manufacturers are shifting their capacity toward high-margin HBM, leading DRAM supply is shrinking and prices continue to rise. Apple faces mounting hardware cost pressure, and bringing in new suppliers aims to diversify supply chain risks while increasing bargaining power with the existing three major storage manufacturers. However, there are multiple practical obstacles to the implementation of cooperation. On the policy front, U.S. lawmakers have already written to Apple, demanding it abandon purchasing domestically produced memory chips. Apple needs to seek regulatory approval from the U.S., which can only purchase standardized finished chips and cannot carry out deep custom development. On the supply side, Changxin Storage's current capacity is nearly at full capacity, making it difficult to release large-scale incremental orders in the short term; On the technical side, Apple has strict standards for memory power consumption, bandwidth, and stability, with lengthy certification cycles. Institutions generally believe that even if implemented, entry-level models will be prioritized, with the high-end Pro series basically excluded in the short term. The news triggered a shockwave in global storage sector sentiment. US stocks like Micron and SanDisk, and South Korean stocks like SK Hynix have all experienced temporary fluctuations. Market divergence emerges: in the short term, testing is more of a bargaining tool; In the medium to long term, if Apple completes the introduction, it will reshape the global DRAM supply landscape.#英伟达推动5000亿美元AI基建融资 $NVDA
NVIDIA's move is essentially the old end-of-cycle "equipment trust + revolving financing" routine: Apollo, Goldman Sachs, and other PE firms raise funds and lend to AI buyers; after buyers purchase GPUs, NVIDIA recognizes revenue, effectively injecting off-balance-sheet debt and hidden leverage into the system.
The $500 billion is currently just a memorandum of understanding target figure, not actual cash, carrying the risk of inflated demand bubbles. More critically, there is a maturity mismatch: GPUs depreciate and become obsolete in about 3-5 years but are used as collateral for long-term debt.
History has examples like Lucent, Cisco, Motorola: at cycle peaks, massive credit was secured with rapidly depreciating assets, ultimately ending in crashes. Currently, strong AI demand masks this fragility, but once demand wanes, this "coin" composed of private credit, GPU collateral, and revolving financing will flip, and the hidden leverage will backfire on the entire chain.
However, the A100 chip launched in 2020 is still widely used globally, extending equipment life cycles up to ten years; the current market's computing power supply continues to fall short of demand, with rental prices for H100 and Blackwell series chips rising steadily, and spot computing power is scarce. This directly proves that demand comes from real business applications across industries—pharmaceuticals, manufacturing, finance, retail are all ramping up AI transformation, not just bubble demand driven by pure capital speculation.
Jensen Huang added that traditional semiconductor cycles are driven by consumer electronics replacement waves like phones and computers, showing clear boom-bust cycles; but this AI infrastructure round is a global, cross-industry digitalization necessity. Countries are all building computing power foundations, with chip, HBM, power, and data center land shortages across the board. Capacity catching up with demand is a long-term theme, and there is no short-term basis for demand collapse.
Responding to accusations of "deliberately amplifying leverage," Jensen Huang explained that the financing platform was created to address real industry pain points: global cloud providers and AI labs invest hundreds of billions in expanding computing power, and relying solely on corporate cash would severely squeeze R&D and business expansion space. The S&P closed at a new high again, with expectations for 8000 points heating up, and this is the kind of time when people are most likely to get carried away.
JPMorgan Chase raised its target, and the reasoning is straightforward: strong Q2 earnings, AI capital expenditures are starting to show returns, and cloud business, orders, and cash flow are all improving. It sounds like a beautiful bull market script. But what I think is more worth being cautious about is that the market is treating "AI spending finally paying off" as the new default assumption.
Once this assumption holds, valuations can continue to rise; once a certain earnings season proves that returns on investment are not that quick, the pullback will also be swift.
8000 points is not a magic number; it’s more like a deposit the market is willing to pay for future earnings. The current question is not whether the US stock market can rise, but whether this deposit has been overpaid.
If you only look at the new highs of the index, you would think everything is strong; if you break it down, you will find that the rise increasingly depends on AI giants turning promises into real money.
#标普收盘再创新高，8000点预期升温 📊 $SOL Contract Liquidation Express (August 16)
According to liquidation data, longs and shorts are repeatedly slaughtering each other, with the "dog whales" harvesting back and forth...
Time Total Liquidation Long Liquidation Short Liquidation
1 hour $10,400 $10,400 $0
4 hours $21,000 $12,600 $8,444.21
12 hours $328,300 $179,800 $148,500
24 hours $2,704,700 $2,487,100 $217,600
From the $SOL liquidation data, in 1 hour long liquidations crushed shorts, shorts were zero, a blitz attack killing longs at the start; in 4 hours the long advantage sharply narrowed, ratio dropped to 1.49 times, short squeeze pressure significantly increased; in 12 hours the long advantage continued to narrow, ratio dropped to 1.21 times, longs and shorts tending to balance; in 24 hours longs surged again, long liquidations soared to $2.48 million, 11.4 times that of shorts, killing longs made a comeback. The "dog whales" completed repeated slaughter between longs and shorts on SOL—kill longs → short squeeze → kill longs again, cumulative liquidations exceeded $2.7 million. Everyone control your positions well, don’t get harvested back and forth.
🔥 Market Weather Vane | August 16
Today's three hot topics point to the same theme: the capital feast of AI infrastructure is facing a brutal test from "burning money" to "making money".
🏗️ AI Infrastructure Earnings Relay: The market only recognizes "real money"
In Q2 earnings season, Wall Street’s logic has completely shifted. Amazon AWS revenue surged 37% year-over-year, Microsoft Azure soared 43%, and the three major cloud businesses combined grew 48%. What truly pushed Amazon into the $3 trillion market cap club was AWS hitting its highest growth rate in 18 quarters.
However, the market does not "buy just because it’s AI." Meta delivered better-than-expected earnings but fell after hours because AI investments have not yet generated independent revenue; Nvidia rose only 2% for the whole week, investors are holding their breath for the August 26 earnings report. The market now rewards not "who invests more," but "who earns faster."
📊 CPI Released Tonight: The Scale for September Rate Hike Hangs in the Balance
At 20:30 Beijing time on August 12, the US July CPI will be released. The Cleveland Fed predicts July overall CPI will rise slightly by 0.09% month-over-month, but core CPI is expected to rise 0.21% month-over-month, rebounding compared to June’s flat month-over-month.
Currently, the market prices in about a 44%-55% chance of a September rate hike. If tonight’s data beats expectations, the hawkish camp will quickly expand; if moderate, rate hike expectations may further fade. This data will be the first domino deciding the direction of the September FOMC meeting.
💰 Nvidia Drives $500 Billion AI Infrastructure Financing: GPUs Become "Investable Assets"
On August 10, Nvidia announced cooperation with Apollo, BlackRock, Blackstone, Goldman Sachs, and KKR to establish an independent computing power financing platform, aiming to leverage over $500 billion in third-party capital. Jensen Huang personally went to Wall Street, with all six institutions present.
However, on the day the news was announced, Nvidia’s stock price fell about 2.8%, wiping out over $70 billion in market value. Michael Burry, the prototype of "The Big Short," publicly warned that the "circular financing" model may repeat the lending chaos before the 2000 internet bubble burst. Jensen Huang emphasized that AI computing power is now "a new era essential infrastructure equivalent to electricity and the internet."
💎 Summary
The AI infrastructure earnings season proves one thing: the market no longer pays for "stories," only for "returns." Nvidia’s $500 billion financing plan is the climax of this capital game—and the biggest bet. Every basis point of tonight’s CPI may determine the macro tone of this gamble. #财报观察员：AI基建财报接力登场
#本周三CPI公布，9月加息定价会改写吗？
#英伟达推动5000亿美元AI基建融资 NVIDIA teams up with Wall Street for a massive $500 billion move, stock price dips first as a sign of respect—Old Mo explains how this relates to your positions
NVIDIA made a big move last night.
On August 10, Jensen Huang officially announced signing a memorandum of understanding with six Wall Street giants: Apollo Global Management, Blackstone, BlackRock, BofA, Goldman Sachs, and KKR. They are creating an independent computing power financing platform aimed at leveraging over $500 billion in third-party capital over the long term, specifically to build data centers and purchase NVIDIA chips for cloud providers, AI labs, and enterprises.
Jensen Huang said: "This is truly the first time technology chips have become an investable asset class." He wants to liken GPUs and data center infrastructure to commercial real estate or toll roads—assets that can be mortgaged, financed, and generate cash flow. BlackRock CEO Fink went further, saying the significance of this is comparable to the birth of mortgage-backed securities in the 1970s.
After the news broke, NVIDIA’s stock price dipped as a sign of respect.
It fell about 1.1% in early Monday trading, closing down 2.9%, wiping out nearly $70 billion in market value. Meanwhile, NVIDIA’s 5-year credit default swaps jumped nearly 6 basis points in one day, doubling insurance costs since late May. Despite the positive news, the stock fell and CDS rose, indicating the market is worried about three things.
First, concerns about "circular financing."
The market worries that NVIDIA helps customers finance → customers use the money to buy NVIDIA chips → NVIDIA’s revenue grows → then NVIDIA uses that money to support the next customer. How much real demand is there in this cycle? Michael Burry, the inspiration for "The Big Short," publicly warned that NVIDIA has pushed circular spending to unprecedented levels. "Doomsday prophet" Jim Chanos mocked this, comparing it to the financial engineering before the 2008 financial crisis.
Jensen Huang personally responded. He said the $500 billion "is neither NVIDIA’s revenue nor a commitment from a single fund or client." Each project is independently evaluated by financial institutions for client demand, utilization, cash flow, and residual value; NVIDIA only provides the platform. The company may provide financing support up to 25% of the project size for some projects.
Second, credit risk is rising.
The CDS surge indicates the market believes NVIDIA is bearing implicit off-balance-sheet risks. Major clients are highly concentrated—the top cloud providers and AI labs contribute the vast majority of chip orders. This tightly bound lending and procurement model amplifies industry volatility risks.
Third, the banks’ attitude is intriguing.
Some analysts point out that NVIDIA turning to private equity for $500 billion precisely shows that traditional commercial banks are cautious about huge investments in computing power centers. If commercial banks were willing to lend, NVIDIA wouldn’t need to bypass them to seek private equity.
Does this matter to the crypto market? Yes, and significantly.
First, AI infrastructure and crypto share the same underlying resources. Data centers, GPU computing power, electricity—AI is competing for the same resources Bitcoin miners use. Morgan Stanley predicts that from 2026 to 2028, major cloud providers’ combined capital expenditures will reach $3.5 trillion. If AI computing power expansion squeezes miners’ hardware access and electricity costs, it will have a tangible impact on the Bitcoin network.
Second, institutional funds are flowing simultaneously into AI infrastructure and crypto assets. Last week, Bitcoin and Ethereum ETFs saw a combined net inflow of $1.1 billion. AI infrastructure financing and crypto ETF inflows share the same institutional capital—BlackRock, Blackstone, and Goldman Sachs are involved both in NVIDIA’s financing platform and in the custody chains of crypto ETFs. If $500 billion-level AI infrastructure financing materializes, it will further consolidate the "tech + crypto" institutional capital alliance.
Third, AI narratives and crypto narratives are merging. NVIDIA’s "AI factory" logic—computing power as revenue, hardware as assets—overlaps at the fundamental level with the crypto world’s "Proof of Work" and "decentralized computing power."
Back to the market.
BTC latest price is about 64000-64200, fluctuating between 63800-65000 in 24 hours. ETH is around 1870-1880. Bitcoin didn’t follow NVIDIA’s drop or rise today—indicating the market is still digesting the news and doesn’t see it as a direct catalyst for crypto assets.
Old Mo has a few words on trading.
NVIDIA’s move has no direct short-term impact on the crypto market. What really matters is how institutional capital allocates between AI infrastructure and crypto assets. If BlackRock, Blackstone, and others treat AI computing power as an "investable asset class," then the institutional allocation logic for crypto assets as "digital gold" will be reexamined.
Bitcoin is expected to continue fluctuating between 64000-65000, with resistance at 64800-65000 and support at 63800-64000. ETH is expected in the 1870-1920 range. Don’t heavily bet on direction before Wednesday’s CPI release.
$500 billion AI infrastructure financing—do you think it’s bullish or bearish for the crypto market? Let’s discuss in the comments. $BTC $ETH $GRVT #英伟达推动5000亿美元AI基建融资 Just checked SOL, hovering around $76, down about 2% in 24 hours. This price is quite awkward—the on-chain data is extremely strong, but the coin price just doesn't follow, a typical case of "applause without sales."
The chain is actually quite robust. Last week alone, it processed 1.01 billion transactions, a historical high. RWA is the biggest highlight; Solana accounts for 96% of tokenized stock trading volume, with 3.86 billion transactions in June alone. DeFi locked value is 4.79 billion, stablecoin market cap is 15.96 billion. Four ecosystem applications even have daily revenues exceeding Hyperliquid.
Why isn't the coin price rising? The core issue is just one—the money earned on-chain isn't flowing to SOL holders. Most of the profits are taken by market makers and project teams, causing a serious disconnect between the coin price and ecosystem activity. The funding rate is also problematic—perpetual contract funding rates have surged to 0.01%, the last time it was this high SOL was above $200. Positions total $1.8 billion, with speculative leverage highly crowded. If resistance near 78 repeatedly holds, high-cost long positions closing could trigger a stampede.
The two most critical short-term factors:
CPI data is released tonight, which may break the current balance. Governance proposal SGP-0003 aims to increase SOL daily burn from 650 to 7,500-9,000, accelerating annual inflation reduction to 30%. Currently, 73 validators support it; if 65.16 million SOL are staked before August 18, it can enter voting. Once passed, the SOL supply structure will be fundamentally changed.
Technically, 75.48-76 is the first line of defense. Above, 78-80 is the core ceiling; breaking through there opens the way to 82-90. On the downside, if 75.5 doesn't hold, 74-75 may be tested.
Honestly, fundamentals are indeed improving, but with short-term high leverage and macro uncertainty, I wouldn't heavily position here. I'll wait for CPI results and proposal outcomes before acting. It's more reliable to move once the direction is clear than to gamble now.
Personal opinion, not investment advice.
$SOL $BTC $ETH #英伟达推动5000亿美元AI基建融资
Just came across a news piece: Nvidia is teaming up with six Wall Street giants to push a $500 billion AI infrastructure financing. Meanwhile, Intel is also raising $15 billion through stock issuance. Both companies announced financing on the same day, but their approaches are completely different, which is quite interesting.
Nvidia's move is indeed clever.
They aren't putting up their own money; instead, they brought in Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR to create an independent financing platform. The goal is to leverage over $500 billion in third-party capital in the coming years to help customers build data centers and buy GPUs.
In simple terms, Nvidia acts as the middleman—customers want to buy chips but lack funds? They borrow from Wall Street. The borrowed money is used to build data centers and purchase Nvidia GPUs. Nvidia sells the chips, Wall Street earns interest, and customers get computing power.
Jensen Huang specifically clarified that "this is not circular financing." But honestly, Nvidia's customers borrow money to buy Nvidia products, and ultimately, repayment depends on the computing power generating revenue—whether this counts as circular financing is for the market to judge.
Intel's situation is much worse.
They directly issued $15 billion in common stock, which dilutes existing shareholders. After the announcement, their pre-market stock price dropped over 3%.
What is the money for? Capital expenditures for 2026 have already been raised above $20 billion, and 2027 will be even higher. The problem is Intel is currently unprofitable and has to keep pouring money in, so the market naturally isn't buying it.
The fundamental difference between the two financing models:
Nvidia is playing "leverage"—not spending its own money but using Wall Street's funds to help sell its chips. Intel is playing "blood transfusion"—its own cash flow is insufficient, so it must dilute shareholders to survive.
One uses other people's money to run its business; the other sells its own flesh to get others' money. The difference is clear.
For the market, this has two sides:
The positive side is that the funding source for AI infrastructure is fully opened; a $500 billion scale means GPU demand is guaranteed for the coming years. The negative side is the market starting to worry about Nvidia's "off-balance-sheet debt"—if these loans can't be repaid, will it eventually impact Nvidia itself? Nvidia's CDS is already soaring.
To be honest, the long-term logic is sound; AI definitely requires heavy investment. But how long this "borrow money to buy chips" model can last, no one knows. The market is still speculating on expectations; when concerns about bad debts arise, it could trigger another bloodbath.
This is my personal opinion and does not constitute any investment advice. #本周三CPI公布，9月加息定价会改写吗？
Let's talk a bit about tomorrow night's CPI.
The nonfarm payrolls already surprised on the downside, lowering the probability of a September rate hike significantly, but the folks at Bank of America still insist on a "September hike," arguing that the Fed cares more about inflation. So tomorrow's data is even more critical than the nonfarm.
What is the market expecting now?
Overall CPI year-over-year is expected to drop from 3.5% to 3.4%, core from 2.6% to 2.5%. Month-over-month, overall up 0.1%, core up 0.2%. The Cleveland Fed's own Nowcasting model calculates overall YoY at 3.42%, roughly in line with market expectations.
But interestingly, the prediction market and CME are at odds.
On Polymarket, the probability of holding rates steady in September is 63%, Kalshi even more extreme at 65%. CME FedWatch shows a 51.2% chance of a hike and 48.8% chance of no change. That's a difference of over ten percentage points. Traders on one side or the other are bound to be proven wrong.
A few key points:
A month-over-month increase is highly likely. June CPI fell 0.4% MoM, July is expected to rise 0.1% — the lagged effect of oil price rebound is starting to show. Core service inflation may remain sticky, and housing costs won't cool down quickly.
For BTC, there are three scenarios:
CPI below expectations (YoY 3.3% or even lower), September hike probability continues to drop, BTC could surge back to 65,500-66,000.
CPI meets expectations (3.4%), market will likely consolidate between 63,500-65,000, waiting for the next wave of data.
CPI above expectations (above 3.5%), hike probability spikes again, BTC might retest 63,000 or even 62,000.
Personally, I think the probability of meeting expectations is the highest. But no matter what, the market will likely remain sideways until the data is out tomorrow night. My position is light; I'll wait for the data to land before making moves.
This is my personal view and does not constitute any investment advice.$BTC
$ETH
$SOL
I think tonight's CPI data shouldn't look bad
So I added some positions again
Mainly because I believe the CPI data should be okay, at least it won't exceed expectations.
Regarding oil prices, although they rose in July, the average price wasn't much higher than in June.
As for wages, last week's wage data showed both annual and monthly rates lower than previous values, especially the monthly rate which was significantly lower. Wages are an important component of goods and service costs.
On consumption, the July consumer confidence index (preliminary) was lower than the previous value, indicating no demand growth pushing prices up.
Regarding the "pubic hair theory," we previously discussed that the non-farm payroll data was a step for the Fed, and the Labor Department's CPI data should be similar. The goal is to prevent the Fed from raising interest rates in the near term. #财报观察员：AI基建财报接力登场
The AI infrastructure earnings season officially kicks off tonight.
Let's first review the timeline:
Tonight (August 11) after market close, Lumentum and CoreWeave will report simultaneously. Tomorrow night, Coherent and Cisco follow, and the day after tomorrow, Applied Materials will close the show. Four big tests in four days, covering optical communications, compute leasing, network equipment, and semiconductor equipment.
Lumentum: The bellwether of optical modules
Market expects Q4 revenue of $988 million, a year-over-year surge of 105%. Last quarter already delivered $808 million, up 90% year-over-year. AAOI's last quarter earnings confirmed demand — revenue of $191.9 million, up 86% year-over-year, with 800G products about to scale. Lumentum is a core optical component supplier in the supply chains of Google TPU and Nvidia GPU; barring any major issues, the optical module line is stable in the short term.
CoreWeave: The cash-burning machine of compute leasing
Market expects Q2 revenue of $2.555 billion, up 110% year-over-year. But the real shock is capital expenditure — expected $7.9 billion in Q2, adjusted loss per share $1.18, loss widened by 339%. Backlog orders at $668 billion, contracted power capacity 3.1 gigawatts. Orders are plentiful but cash burn is intense; the market is watching whether losses can narrow.
Coherent & Cisco: Continuing tomorrow night
Coherent management guidance is revenue between $1.91 billion and $2.05 billion; FCC is still pushing policies to ban imports of Chinese optical modules. Cisco has raised its AI infrastructure order target from $5 billion to $9 billion, with full-year AI revenue expected to be recognized at about $4 billion.
Applied Materials: Closing on Friday
Market expects earnings per share of $3.36, up 35% year-over-year, revenue of $9 billion, up 23%. Semiconductor equipment is upstream of AI chips; guidance is more important than current data.
To be honest
The core focus of this earnings round is one thing — whether AI infrastructure orders can really convert into hard cash. Optical communications are relatively stable, compute leasing is watched for loss narrowing, network equipment is watched for AI revenue proportion, semiconductor equipment is watched for guidance. The US stock market is already at historic highs; any earnings slip now could cause a big sell-off. Personally, I will control my position size and wait until all these earnings reports are out before deciding the next step.
Personal opinion, not investment advice. $XRP's candlestick looks exactly like the script of "bottom fishing halfway up the mountain"
Around 1.50, there was a long lower shadow, which looks like a reversal, right? But the problem is—after the long lower shadow appeared, there was no strong buying follow-up
A truly strong rebound should be a series of consecutive bullish candles pushing upward. XRP is following the path of "sideways → bearish candle piercing the lower shadow → continuing downward"
Now XRP is hovering around 1.03-1.04, having dropped 44% year-to-date. On Polymarket, 67% of people are betting it will fall below 1 dollar in August
Analyst Chart Nerd believes XRP may first dip to 0.7 dollars—before a strong rebound appears, it still needs to drop further
It's not just XRP's problem; the entire altcoin market is being drained.
The Altseason Index is only 42/100, BTC dominance has surpassed 58%. Funds are concentrating into Bitcoin, altcoins are bleeding
XRP ETF net inflows plummeted 93% last week, while BlackRock and Fidelity's Bitcoin ETFs handled 850 million in one week; XRP's money isn't even a fraction of that
In short, the altcoin winter is not over yet
No buying follow-up after the long lower shadow, ETF funds flowing out, Bitcoin absorbing liquidity—three signals all point in the same direction: another drop is coming
My judgment: XRP is unlikely to hold 1 dollar; the next stop is around 0.95-0.98, with an extreme case possibly down to 0.7.
Not just XRP, most altcoins are following the same script. Not all altcoins will fall together, but when the overall trend is down, the odds of going long against the trend are very low #财报观察员: AI infrastructure financial reports debut in succession Recently, global AI infrastructure companies have been intensively disclosed in their financial reports, with chip, server, storage, and cloud vendors delivering their report cards one after another. The AI industry has officially entered its "accounting moment," and the market is no longer just chasing capital expenditure stories, but is placing greater emphasis on order conversion, cash flow, and profit quality. Upstream hardware performance was the most prominent. Storage, AI servers, and high-speed interconnect hardware companies benefited from expanded computing power, with revenue and profits rising sharply. NVIDIA is about to enter a critical earnings window, with the implementation of a $500 billion computing power financing framework. The market is closely watching GPU shipment guidance and actual downstream procurement demand. Although storage companies like SK Hynix hit record profits, their stock prices fluctuated sharply after the financial report was released, reflecting market concerns about whether the high prosperity could be sustained. At the cloud vendor level, Microsoft, Amazon, and Google have continuously raised AI capital expenditures and invested heavily in computing infrastructure, but massive capital investments keep suppressing free cash flow, creating a pattern where "hardware profits, cloud providers expand revenue." On one hand, orders and revenue are growing rapidly; on the other, data center construction and GPU procurement bring massive capital expenditures. When ROI will improve has become a core question for institutions. Current market divergence is intensifying: an optimistic view believes AI computing power demand will remain rigid for the long term; The risk points are concentrated in the rapid depreciation of hardware iterations and the false demand driven by leverage. If downstream AI commercialization falls short of expectations, the high capital expenditure model will face pressure. Earnings season is currently filtering for stocks with truly fundamental strengths, and thematic speculation is gradually fadingThis afternoon, I was writing my weekly report in a café, and the two people at the next table were talking loudly. One was talking about an AI coin he bought last month that doubled but he didn’t sell, and now it has dropped back to his cost basis. The other said, "At least you didn’t lose money," then fell silent. I glanced down at my screen; the price of $BTC has moved less than three hundred dollars since this morning, digesting last week’s high-volume bearish candle all day.
ETH gas prices have been hovering in single digits for a week straight. Based on past experience, such ultra-low gas prices don’t last long because on-chain activity always picks up again. But this time it’s different: Solana’s cross-chain bridge traffic is steadily increasing, with the amount bridged daily from ETH doubling compared to last month. ETH is losing part of its originally owned liquidity.
I noticed a very complex operation from an address on $AAVE today. First, it borrowed USDT using WBTC, then swapped USDT for USDC, used USDC as collateral to borrow ETH, and finally swapped ETH for stETH to deposit into a liquidity pool. This entire process involved three lending protocols and two DEXs, with an annualized yield of about 0.8%. People running this strategy clearly aren’t chasing high returns but are seeking positively skewed return opportunities, profiting from overlooked arbitrage.
$LINK continues to see accumulation today. One address has been consistently withdrawing from exchanges over the past week, with small amounts but very high frequency—about one withdrawal every two hours—accumulating over 100,000 LINK withdrawn. This high-frequency small withdrawal pattern looks more like an oracle node replenishing collateral rather than retail behavior. On the $MKR side, someone is doing something similar: an address withdrew a sizable amount of MKR from Coinbase today and directly deposited it into Spark’s sDAI pool, earning savings interest while maintaining MKR governance exposure.
$UNI’s long-tail trading volume on the front end shrank a bit again today, but there’s a detail—a market maker address simultaneously adjusted positions on $CRV and $CVX, swapping CRV for CVX at a one-to-one ratio. This usually means someone expects a change in value distribution within the Curve ecosystem, with CVX undervalued relative to CRV. $PENDLE’s yield curve steepened today, with ongoing purchases of long-term PT. Combined with last week’s large $ENA position, it looks like someone is setting up a multi-month interest rate trading strategy.
$SOL saw several cross-chain stablecoin inflows today, all from the same ETH address, splitting amounts more than yesterday, with the smallest around a few hundred U and the largest just over ten thousand. This method looks like spreading deposits across multiple lending protocols to earn interest while keeping liquidity readily withdrawable. If you see more addresses doing the same, it indicates a genuine lack of better yield opportunities in the market. $SUI and $APT didn’t hit new lows today but had weak rebounds, basically moving slightly with the broader market. $SEI’s trading volume was even lower than yesterday, with the thickest buy order on the order book just over thirty thousand dollars—this depth can be easily broken by a slightly larger retail trader. $TIA and $INJ had no independent moves today, completely following the market rhythm, with volatility narrowing to recent lows.
The AI sector didn’t continue yesterday’s relative strength today. $FET and $RNDR briefly surged in the morning but were quickly pushed back down. The sell pressure came from several fixed addresses, not the same entity, but with very regular intervals—selling batches every fifteen minutes, resembling a TWAP sell strategy. $TAO had a large transfer from an exchange to a private wallet today, close to six figures; such a withdrawal size is usually beyond retail capability. $AR continued low-volume sideways trading, while on $NEAR, an address has been consistently withdrawing from Binance at fixed times daily for five consecutive days, accumulating a significant amount.
The RWA sector was quieter today, but $ONDO and $OMNI address activity didn’t completely stop. The address that tested slippage ranges yesterday had no new moves today, but another newly created address bridged from Polygon to ETH and bought US Treasury tokens, with a small amount and the exact same path, indicating someone is running a structured automated strategy. A large USDC-to-DAI swap appeared today in $CRV’s stablecoin swap pool, with well-controlled slippage and execution price almost equal to market price. This trade looks like an institution adjusting stablecoin allocations, not arbitrage.
The Meme sector was collectively quiet today. $PEPE and $DOGE volatility dropped to recent lows, with wider bid-ask spreads than usual, and market makers unwilling to provide thick liquidity at this level. $BONK and $WIF trading volumes also shrank significantly. Occasionally, someone in the group shouts, "Is this the bottom?" and replies are either memes or "I’ll buy if it drops another half." $FLOKI saw some address activity today—a not-large transfer from an exchange to a contract address. Such moves in Meme usually mean someone is preparing to deploy new liquidity, but the amount
Net buys clearly were: $LINK, $AAVE, $MKR, $ENS, $LDO, $RNDR, $FET, $TAO, $MNT, $METIS. Net sells continued for: $ARB, $OP, $STRK, $BLAST, $SUI, $APT, $SEI, $TIA, $INJ, $PEPE, $BONK, $WIF, $FLOKI. These data only show today’s capital flow; tomorrow morning might be a completely different picture.
In the evening, I went downstairs to buy a coffee and took a couple of sips standing at the shop’s door. The milk tea shop across the street had a line extending onto the sidewalk, with several young people discussing weekend plans. Most things in life have nothing to do with the crypto market. When the market is quiet, you can clearly see who is really working and who is just riding the hype.
The cup is empty, and the sky has darkened. Time to go back and keep watching the data. #财报观察员：AI基建财报接力登场
Damn! The two-year AI infrastructure frenzy is finally coming to an end, and the market is about to conduct a post-mortem: after burning through hundreds of billions, have the orders actually turned into profits?
No matter how good the orders look on the PPT, they must be backed by real gross margins and cash flow, or else the market will just toss them out as trash.
In the optical module segment, Lumentum is the choke point. Both Google’s TPU and Nvidia’s GPU supply chains rely on its laser and optical components.
In Q3, revenue has already hit over $800 million, nearly doubling year-over-year, with operating margins soaring above 30%. Wall Street’s next quarter forecast is directly aiming for $1 billion, doubling again.
The earnings report comes after market close tonight. If revenue and margins beat expectations again, the bears still clinging to storage logic will be even more discredited; if it falters, the optics rally will be immediately busted.
A well-known trader on X bluntly said: AI clusters have expanded from thousands to hundreds of thousands of cards, and compute power is no longer the only bottleneck—data transfer is the real killer. Bandwidth scarcity is replacing compute scarcity, and Lumentum is stuck right in the middle. Some say its production capacity is already booked through 2028; this isn’t a cycle, it’s a structural shortage.
SpaceX is another extreme case: the first batch of over 900 million shares unlocked hit the market, yet the stock price rose from 108 on the unlock day to 136, a 26% increase. The market is still willing to pay for the long-term story of AI aerospace infrastructure.
But on August 20, a second wave of 319 million shares will be released, with even larger amounts coming in September and October. Early holders have enjoyed paper wealth for years and can finally cash out.
Some KOLs on X think: being able to sell doesn’t mean you have to dump, but supply is real. Whether the stock price can hold depends on how the market prices the Terafab project. If it can’t hold, bears will count their money; if it holds, the narrative can last a few more quarters.
The remaining earnings reports this week are for the post-mortem. CoreWeave’s revenue continues to double, backlog orders have piled up close to $100 billion, but GAAP is still bleeding big losses, and capital expenditures are burning like there’s no tomorrow.
Cisco has directly raised AI infrastructure order expectations from $5 billion to $9 billion. Kohler Ant’s orders are already booked through 2028. When Applied Materials reports, we’ll see clearly if chipmakers are truly expanding capacity or just blowing smoke.
Everyone understands that revenue growth has become the entry ticket; what really determines stock price is backlog conversion rate, actual megawatts deployed, whether interest costs are out of control, and if the burn rate can be reduced.
Some call CoreWeave the cleanest litmus test for whether AI infrastructure is a bubble. Demand looks strong, but execution is the real killer.
The sector is rapidly diverging. Some companies are already truly profitable with rising gross margins; others are still frantically grabbing land, raising funds, and burning cash, hoping utilization can climb before the music stops.
Storage just delivered explosive earnings but still got crushed. Now making money isn’t enough; you have to prove how long you can keep it. The burn-for-growth crowd is the same; the market is starting to calculate the return on every dollar.
The story has been told; the market is digging up old accounts. Whoever can truly turn paper orders into higher gross margins and realized cash will continue to be favored. The rest are just cannon fodder for the next wave of buyers. #Strategy sells another 1,690 BTC, corporate treasury shows divergence
Sigh, Strategy sold again.
From August 3 to 9, sold 1,690 BTC at an average price of $64,262, cashing out 108.6 million. All proceeds were used to repurchase STRC preferred shares. Including the 1,638 BTC sold the previous week, a total of 6,916 BTC have been sold over six weeks, raising 429 million. Holdings dropped to 840,447 BTC.
But this is just moving assets from one hand to the other. In the same week, Strategy sold 6.5857 million MSTR common shares through the ATM program, net raising 653 million, of which 650 million was directly transferred into USD reserves. USD reserves increased from 4 billion to 4.65 billion.
These details are worth pondering.
First, they have to sell even at a loss. The average holding cost is $75,385, this sale was at $64,262, a single loss of 18.8 million. But STRC preferred shares have already discounted to $96, so repurchasing is necessary.
Second, cash reserves are the real bottom line. USD reserves at 4.65 billion, repurchase quota still has 785 million left. Saylor’s actions acknowledge one thing: having no cash on hand is scarier than Bitcoin dropping.
Third, BitMine is taking a completely opposite path from Strategy. Strategy is selling BTC to hoard USD, while BitMine bought 7,391 ETH at an average price of $1,888. One is contracting, the other expanding.
Selling 429 million over six weeks is not large relative to the 840,447 BTC holdings. But this is no longer a question of "whether to sell," but "when will the selling stop." When the largest bull starts proving their ability to pay by selling coins, the market doesn’t care how much you sold, but how long you can hold on. $BTC #本周三CPI公布，9月加息定价会改写吗？
#霍尔木兹海峡通航协议未落地，油价风险升温
Today, the main factor affecting $BTC sentiment
might not be a particular blockchain or a new token,
but a barrel of oil.
On August 10, tensions around the Strait of Hormuz escalated again, pushing oil prices up about 5%; meanwhile, the S&P 500 index slightly fell by 0.06%, and the Nasdaq dropped 0.32%. The U.S. Bureau of Labor Statistics will also release the July CPI at 20:30 Beijing time on August 12.
Rising oil prices will reignite market concerns about inflation, and inflation expectations in turn affect interest rates and risk asset valuations. BTC has recently hovered around $65,000 without a clear move, which is not only due to internal bullish and bearish battles in the crypto space but also a wait ahead of macro events.
Sometimes, to understand a single BTC candlestick,
you really have to first see what happened with crude oil. $NBIS options show divergence
Spot price weakens in the short term, Delta capital flow weakens, but options show no panic, instead giving Call premium.
Price closed at $184.11, down 2.05%, with the night session around $186.
190 is an important level, with multiple different market structures converging here. 190–195 was originally one of the most active option zones.
After the spot price broke through, it could not sustain the high price and fully retraced, indicating that on the day before the earnings report, the chasing funds did not control the market.
184 happens to be the largest trading area in the dark pool, essentially returning to the chip exchange zone.
Earnings implied volatility is 11.01%, and at the price of 184, the approximate range is 164–204.
Open Interest (OI) forms a Barbell shape, with 150 Put, 190, 250, and 320 Call expanding, indicating the market is betting on volatility.
If the price stays above 190 before the earnings, the overall bias is bullish. $BEAT rebound lacks strength? Is it like $LAB $RAVE where the whales have fled? Has it been continuously falling?
1. Excessive gains in the early stage
* BEAT once experienced multiple-fold gains, with heavy profit-taking in the market.
* Such projects often fall faster than they rise once funds withdraw.
2. Recently underperforming the market
* In the past week, BEAT has been one of the poorer performing tokens, while some altcoins have already started to rebound.
From a "whale control" perspective
If the following situations occur, I would increase vigilance for whale exit:
* Top 10 addresses hold an extremely high proportion;
* Project wallets continuously transfer tokens to exchanges;
* Community operations stop updating;
* Market makers withdraw buy orders;
* Trading volume keeps decreasing but price continues to decline slowly.
Currently, public information looks more like:
Unlocking + profit-taking escape + market maker support weakening
Not necessarily a run, but possibly entering the late stage of whale distribution.
Key points:
① Top 10 holdings 84.34%
This is the most alarming data.
Normal projects:
* Top 10 generally 20%-50%
Moderate whale control:
* Top 10 50%-70%
High whale control:
* Top 10 >80%
For BEAT here:
Top 10 holdings = 84.34%
Indicates extremely concentrated chips.
But there is a question:
If most of this 84% is:
* CEX exchange wallets
* Liquidity pools
* Project treasury
Then the risk is not that high.
If 10 private addresses control 84%, that's another matter.
② Developer holdings <0.01%
Data does not look like a rug pull.
Because many rug projects show:
* Developer holdings 10%
* 20%
* 30%
Then directly dump the tokens.
Developers have already cleared their holdings.
This means:
Possibly:
* Tokens have been allocated to market makers
* Transferred to foundation wallets
* Transferred to institutional wallets—not necessarily malicious exit.
③ Smart money increasing positions
Screenshots show smart money adding positions.
Although this label is not 100% accurate.
But if on-chain recently wallets with higher profitability keep buying,
It indicates:
There is still capital willing to take the tokens in the market.
This does not align with whales completely fleeing.
④ Price dropped 51%
This point fits more with:
Whale distribution followed by slow decline
Rather than:
Whale exit causing a crash
Difference:
Exit:
* Drops 80%-99% in one day
* Liquidity disappears
* No one to take the tokens
Distribution:
* Continuous decline over several weeks
* Each rebound is suppressed
* Trading volume gradually decreases
BEAT looks more like the second scenario.
My judgment
Estimated probabilities:
* Whales have already fled: 20%
* Whales still present but in distribution phase: 50%
* Whales shaking out for a second wave: 30%
Currently, the most dangerous signal is not developer clearing holdings.
But:
Top 10 holdings at 84.34%
This indicates BEAT's future price movement largely depends on a few large holders. Japanese government bonds are crashing
The 2-year yield hit 1.615%, the highest since March 1995. The 5-year yield is 2.09%, a 31-year high. The 10-year yield is approaching 2.805%, just a breath away from the 3% warning line.
An economy with a debt/GDP ratio over 240%, and interest rates soaring—this scene is too intense to watch.
What is the market afraid of?
The expectation of rate hikes is getting more urgent—the probability of a rate hike in September has surged to 66%, up from just 30% at the end of last month. The 2-year yield is skyrocketing, betting that the Bank of Japan will take action.
What’s even more contradictory is that the government is pursuing expansionary fiscal policy while the central bank wants to raise rates; the two are working against each other. The demand for government bond auctions has collapsed, directly pushing yields higher.
An even bigger problem is—
Japan is the world’s largest holder of U.S. Treasuries. If it is forced to sell U.S. Treasuries to rescue its own bond market, U.S. Treasury yields will be pushed up, shaking the global asset pricing anchor.
For the crypto space, once liquidity tightening expectations form, all risk assets will suffer.
I’m watching Japanese bonds closely. If the 3% level breaks, it might not just be Japan’s problem, but a global one.The most valuable move on the chessboard is not the checkmate of the king, but the opponent's queen that was quietly maneuvered early on but not yet placed. Sandisk's last quarter's earnings and EPS exceeded expectations, yet the midpoint guidance for the next quarter planted a hint below consensus—this contrast is not a novice's miscalculation, but a grandmaster's classic double threat: seemingly offering you a piece, while in reality, your queen's flank is wide open. The market's violent swings after the earnings report resemble a heavy piece losing its pawn chain protection—there is a clear path forward, yet you don't realize the opponent's bishop has already diagonally sealed all retreat routes.
The truly valuable moves are never the ones you have already seen, but the ones you haven't yet noticed. The investor day on August 13 is the agreed moment for a midgame review. Is the weak guidance a temporary patch or an early concession to real demand weakness? The supply and demand of flash memory chips determine whether the midgame pawn structure is stable or fragmented; the AI storage roadmap decides if you still have the capacity to launch a queen-side offensive. The $14 billion buyback is exchanging heavy pieces for queen-side pawns—you lose immediate attacking power but gain theoretical winning chances in the endgame. However, having an extra pawn in the endgame is not victory but the beginning of torment. Timing the exchange is always more important than quantity; exchanging at the wrong time is equivalent to handing victory back to the opponent.
I've seen too many players panic when the opponent sacrifices a piece. A true grandmaster, at the moment of the opponent's sacrifice, asks themselves: which line does he want? The real value of this midpoint guidance may not be to tell you demand is weak, but to tell you management has already foreseen the next mistake in market consensus. While everyone bets on next quarter's performance, the smart player is already calculating the subtle differences in the opponent king's position in the endgame. The linkage between XIBM is like shadow chess on two boards—the same opening, different variables. You think you see the same game, but the timers on both sides are already off by more than one beat.
In the midgame, the most terrifying trap is not the opponent making a strong move, but making a seemingly smooth move that quietly shifts the entire board's balance. This midpoint guidance is such a smooth strike. It didn't crash the market, but it forced everyone to reassess their relationship with the game. Investor day is more like a "midfield lock"—both sides have time to test their assumptions' errors. But after the lock ends, the real situation will emerge.
The market wants answers on investor day. But a grandmaster's answer is never at the review meeting. The review meeting only convinces the opponent of your own half-believed doubts. So, don't rush to judge that weak guidance as a defensive flaw. Sacrificing a piece may be the necessary path to a decisive kill. When that midpoint guidance written on paper falls like a toppled pawn into the gray consensus zone—the real question is never why it fell, but what the board reveals after it falls.
You know it's not the king's front that has fallen, but the abyss on the queen's side. #sandiskinvestorday A predominantly long whale suddenly reversed, placing $8.05M worth of short orders against SKHX
An old face on the PnL leaderboard on the 7th, 0x66f4...8836, a swing address with $20.36M in equity and $1.75M in historical profits, usually prefers to be on the long side.
But in the 5 minutes starting at 5:33 AM, he consecutively placed 2,000 short orders, executing 7,950 SKHX contracts valued at $8.05M, with an average price around 1012.2.
This is completely opposite to the predominantly long bias in his 58 historical trades, resembling a sudden position shift.
The account is currently fully cross-margined, with no other positions in the same direction; this trade is almost a all-in bet.
Public data shows his win rate is 33.3%, not very high, but the PnL curve holds up, indicating possible skill in stop-loss or risk-reward management.
If SKHX cannot quickly recover the cost zone, whether he continues to add shorts or reverses will be key to observing his true intentions.
If you like my sharing, please consider following.$JD stock price is consolidating narrowly at $33.47, with a $45.2 billion market cap mismatched against $1.32 trillion in annualized revenue, forming the core contention point of whether gross margin can support valuation recovery.
From the price structure perspective, $JD's daily trading volume remains at $7.65 million, with the price in a sideways consolidation. Compared to peers in the same sector, BABA has a circulating market cap of $317.15 billion, PDD has $132.38 billion, and $JD's $45.2 billion market cap shows a characteristic of valuation discount.
In terms of driving factors, the top priority is to observe whether gross margin can match capital expenditure intensity, followed by order growth and market share changes. Macro interest rates suppressing the overall valuation midpoint are considered external variables.
In the valuation recovery scenario, if order growth rebounds and quarterly gross margin stops falling and stabilizes, supply chain monetization efficiency is validated, and the price-to-sales ratio has 20% to 50% expansion potential, the price will break out of the consolidation range near $33.47 and seek an upward breakthrough.
In the downside test scenario, if price wars intensify again squeezing profit margins and AI capital expenditure falls short of expectations, valuation pressure will trigger sell-offs, with market cap facing downside risk toward halving to $45.2 billion.
After confirming the downtrend, if the company significantly increases buybacks, the downward momentum will be interrupted and turn into accumulation and digestion within the range.
This entire deduction fails if macro interest rates rise; once external rate environments increase suppressing overall Chinese e-commerce valuations, the upward logic brought by internal gross margin stabilization will be directly invalidated.
In the next 7 days, key focus will be on the realization of quarterly gross margin and revenue growth data, as well as institutional holdings changes and order backlog data disclosed in the 13F report.
#三星钱包将接入稳定币，支付场景继续扩展 #霍尔木兹海峡通航协议未落地，油价风险升温Life inside the system is really tough now; many people have overdue loans.
The bank president talked about finance, saying that many of the overdue loans he has are from people inside the system, and this data is growing rapidly.
I said, aren't the clients with housing provident funds your premium customers?
He said yes, precisely because they were premium, loans were previously given very liberally.
He said, do you know what they mostly do? The majority speculate on real estate,
next are those involved in investments; many projects are backed by them,
because they have networks and resources, controlling many good projects.
They invest themselves and find someone to hold the shares on their behalf, but who knows?
Projects that were good before are good projects, but given today's situation,
there are no absolutely good projects anymore. It used to be easy for them to get loans,
many people loved to stir things up, but now it’s getting difficult.
There are also some who have problems and need money to settle them, generally three types of problems:
issues with women, disputes with business owners, and dirty dealings that could be suppressed before but now clearly cannot be.
So the last wave of defaults will hit these people.
Ordinary people were already wiped out two years ago; these people have better conditions and can hold on longer.
Actually, what was the previous dividend period inside the system? It’s not that the system salaries were high,
they weren’t high. Even if you were on the coast, compared to local wealthy small business owners’ incomes, it wasn’t high.
But why did they have money, or get rich?
It’s not really about corruption; corruption is only for top leaders or those in real power positions who have the qualifications.
Others can’t get much money; bosses won’t give money to people who don’t have decisive influence.
They had easy access to credit, easy real estate speculation, easy information, familiarity with policies, and side projects.
They didn’t get rich by just earning wages or corruption. Which old comrade doesn’t have three to five houses?
Even someone who used to work at a state-owned restaurant could have three houses: two old and rundown, one new and improved.
Now it’s not like that anymore; it’s hard to make money.The seabed of the Strait of Hormuz is trembling—not due to geological movement, but because the geopolitical load-bearing wall is groaning under stress fractures. I stare at the set of oil price data—WTI standing at 82.13, Brent touching 87.72—like watching a supertall building record its sway during a typhoon. The wind tunnel test report shows the damper is fully loaded, yet the tower crane keeps rotating.
The "transit fee" on the negotiation table remains undecided, which to me is equivalent to the absence of a pile foundation inspection report. Any structural engineer knows that the bearing capacity calculation methods for friction piles and end-bearing piles are completely different, yet Iran and Oman haven't even confirmed the friction coefficient. This bridge called "stable navigation" dares to leave construction joints in the narrowest channel of Hormuz. Trump extended the Jones Act waiver for another 90 days, but that's just like adding an early-strength agent to concrete—on the surface, formwork can be removed in 3 days, but the accumulated heat of hydration is enough to cause internal cracking in large-volume concrete.
The real killer is not today's oil price gap, but the architectural core issue of "long-term expandability." The sanctions framework is the outer curtain wall, insurance clauses are the fire compartments, and shipping regulations are the evacuation routes—if these three cannot mesh tightly like steel structure nodes in a CBD, then every LNG ship passing through the strait is undergoing a load test without redundancy. I've seen too many skyscrapers in white papers, with renderings piercing the clouds, yet encountering karst caves during the piling stage.
The market's risk premium on Hormuz is essentially a discount on the untrustworthy "completion date" promise. Oil tankers rerouting around the Cape of Good Hope is like shifting the core tube by three degrees, exponentially increasing the building's eccentricity. Assets like XQQQ still maintain correlation when the panic index hits the greed zone of the FearAndGreedIndex, which itself indicates that capital is welding the "vertical transportation" of petrodollars and tech stocks together—if seawater enters the elevator shaft, no matter how expensive the penthouse is, it’s just the backdrop of a disaster movie.
Now all parties claim they can manage the risks, but architectural history never acknowledges strength levels spoken aloud. The "live load" of the Strait of Hormuz changes daily with the draft of oil tankers, yet the "structural health monitoring" system still lacks a tripartite signed acceptance report. I can draw the most beautiful suspension bridge plans, but I cannot convince myself to step onto a bridge deck that has never undergone wind tunnel testing.
Structural failure never happens instantly; it is only visible to the naked eye at the very end. #hormuzdealunresolved📊 $CL Contract Liquidation Express (August 16)
According to liquidation data, this wave of shorts was brutally crushed by the dog whales...
Time Total Liquidation Long Liquidation Short Liquidation
1 hour $5,888.01 $87.09 $5,800.92
4 hours $34,200 $3,522.65 $30,700
12 hours $153,500 $15,100 $138,300
24 hours $1,686,500 $85,300 $1,601,200
From the $CL liquidation data, short liquidations crushed longs in 1 hour, with short liquidations 66 times that of longs, a short squeeze blitzkrieg starting with nuclear-level intensity; the 4-hour short advantage continued, ratio about 8.7 times, a full-scale short squeeze outbreak; at 12 hours shorts still far ahead, ratio about 9.1 times, short squeeze spanning short to mid cycles; at 24 hours short liquidations soared to $1.6012 million, 18.8 times that of longs. The dog whales completed a full-cycle slaughter of shorts on CL—shorts across short, mid, and long cycles were comprehensively targeted and blasted, longs’ only resistance was futile, with cumulative liquidations exceeding $1.68 million. Shorts bled heavily, the short squeeze momentum unstoppable. Everyone, control your positions and avoid being repeatedly harvested.
🔥 Market Wind Vane | August 16
Today's three hot topics point to the same theme: the capital feast of AI infrastructure is facing a brutal test from "burning money" to "making money".
🏗️ AI Infrastructure Earnings Relay: The Market Only Recognizes "Cold Hard Cash"
In Q2 earnings season, Wall Street’s logic has completely shifted. Amazon AWS revenue surged 37% year-over-year, Microsoft Azure soared 43%, and the three major cloud businesses combined grew 48%. What truly propelled Amazon into the $3 trillion market cap club was AWS’s highest growth rate in 18 quarters.
However, the market does not "buy just because it’s AI." Meta delivered better-than-expected earnings but fell after hours because AI investments have yet to generate independent revenue; Nvidia rose only about 2% for the week, with investors holding their breath for the August 26 earnings. The market now rewards not "who invests more," but "who profits faster."
📊 CPI Released Tonight: The Scale for September Rate Hike Hangs in the Balance
At 20:30 Beijing time on August 12, the US July CPI will be released. The Cleveland Fed forecasts July overall CPI to rise slightly by 0.09% month-over-month, but core CPI is expected to increase 0.21% month-over-month, rebounding compared to June’s flat month-over-month.
Current market pricing for a September rate hike is about 44%-55%. If tonight’s data beats expectations, the hawkish camp will quickly expand; if moderate, rate hike expectations may further fade. This data will be the first domino deciding the direction of the September FOMC meeting.
💰 Nvidia Drives $500 Billion AI Infrastructure Financing: GPUs Become "Investable Assets"
On August 10, Nvidia announced cooperation with Apollo, BlackRock, Blackstone, Goldman Sachs, and KKR to establish an independent computing power financing platform, aiming to leverage over $500 billion in third-party capital. Jensen Huang personally went to Wall Street; all six institutions were present.
However, on the day the news was announced, Nvidia’s stock price fell about 2.8%, wiping out over $70 billion in market value. Michael Burry, the prototype from "The Big Short," publicly warned that the "circular financing" model might repeat the lending chaos before the 2000 internet bubble burst. Jensen Huang emphasized that AI computing power is now "a new era essential infrastructure equivalent to electricity and the internet."
💎 Summary
The AI infrastructure earnings season proves one thing: the market no longer pays for "stories," only for "returns." Nvidia’s $500 billion financing plan is the climax of this capital game—and the biggest bet. Every basis point of tonight’s CPI could determine the macro tone of this gamble. #财报观察员：AI基建财报接力登场
#本周三CPI公布，9月加息定价会改写吗？
#英伟达推动5000亿美元AI基建融资 Never would have expected
Shorting $SPCX got schooled by Elon
Sure enough, there are no guaranteed opportunities in this world
Everyone thought the unlock would cause a crash
But the final result was a pump
The market always moves opposite to what most people expect
However, I still don't have a positive outlook on $SPCX's performance within this year
The August 6 unlock was just an appetizer
There will be staggered unlocks through December this year
Elon won't just accept it all
Selling pressure still exists. #财报观察员：AI基建财报接力登场 #英伟达推动5000亿美元AI基建融资 #本周三CPI公布，9月加息定价会改写吗？ The valuation divergence in the Chinese concept e-commerce sector is narrowing, with buying support for supply chain assets and profit concerns from price wars intersecting at specific price levels.
$JD stock price hovers around $33.47, with a single-day trading volume maintained at $7.65 million, showing an overall narrow consolidation pattern.
Against an annualized revenue of $1.32 trillion, its circulating market value of $45.2 billion is significantly lower than that of peers Alibaba and Pinduoduo.
Whether the gap between scale and valuation can be converted into a recovery momentum remains to be confirmed, with the core issue being whether the gross margin matches capital expenditure.
If order growth recovers and gross margin stabilizes, there is a 20% to 50% expansion potential in the price-to-sales ratio; however, if rising macro interest rates suppress the overall valuation baseline, this upward logic will be invalidated.
If price wars intensify squeezing profits and AI capital expenditure falls short of expectations, the market value risks converging toward a halving position of $45.2 billion; if subsequent buyback amounts significantly increase, the downward trend will be interrupted.
Changes in institutional holdings in the latest reporting period and the backlog of orders will directly falsify or confirm the current valuation bottoming judgment.
The most worthwhile variables to continuously track in the coming days are the actual quarterly performance of gross margin and revenue growth.
#标普收盘再创新高，8000点预期升温 #CLARITY表决推迟至9月，监管窗口后移 #伯克希尔结束净卖出，重启大额配置New Large Short Position Opened On-Chain: SKHX Faces Whale-Level $4.54M Short Order Breakdown
The SKHX order book was just hit by consecutive broken orders pushing the price down, with a new $4.54M short order rapidly filling.
The placing address 0x66f4...8836 is a regular on the 7-day PnL leaderboard and whale equity rankings, holding $20.36M in equity, $1.75M in historical profits, a 33.3% win rate, and mostly long swing trades out of 58 transactions.
This time, however, it opened 4.49K short contracts on SKHX at an average price of 1012.36, completed through 1119 fragmented orders. This is not an add-on to an existing profitable position but a pure new position.
A whale who has been predominantly long for a long time suddenly turning short is definitely worth a closer look. Currently, this account holds no positions in the same direction; if it continues to add shorts, it indicates a changed outlook on the current market trend.
If you like my sharing, please consider following.#Spot ETF capital divergence, BTC selling pressure remains Last week, the combined net inflow of $BTC and $ETH spot ETFs was about $1.1 billion. Institutional capital returning was originally a positive signal, but the latest data shows divergence again: BTC ETFs have reverted to net outflows, while ETH still maintains a slight net inflow.
At the same time, on-chain whales and miners are still transferring out BTC. So I think what really deserves attention in the short term is not "whether ETFs are inflowing or outflowing today," but:
Whether new buying can continuously absorb the selling pressure from old holdings. ETFs have indeed changed BTC's capital structure, but they cannot eliminate the cycle. Miners, early whales, and long-term holders cashing out at high levels mean that as long as selling pressure persists, new capital must continuously take over.
ETH's current relative strength is also worth observing. If ETFs continue net inflows while BTC capital keeps flowing out, there may be signs of short-term capital rotation from BTC to ETH.
But if I had to choose between the two now, I still favor $BTC.
The reason is simple: when the macro environment is not yet fully clear, BTC remains the core asset for institutional capital entering Crypto.
I will focus on three things next: whether ETFs can sustain net inflows again, whether whale selling pressure decreases, and whether BTC spot trading volume can expand.
Only if all three improve simultaneously will I believe the next rally truly has a capital foundation. Ultimately, the market trend is not about who shouts the loudest, but whether there is enough money in the market to absorb the selling pressure.Uh... The retail investor subscription multiple for Unitree's IPO exceeded 8,000 times.
Is the demand for pure humanoid robot companies absolutely huge?
I am personally a member of the US Agile Robotics Club (SoftBank, $NVDA, $AMZN, Foxconn, etc.) and have obtained a $2.5 billion pre-financing valuation through $CCXI.
My view is that Unitree's listing valuation may exceed $30 billion (from pre-IPO investors), which could bring more attention from leading US investors within a week or two.
Anyway, this is simply a ridiculous demand.
#财报观察员：AI基建财报接力登场 #本周三CPI公布，9月加息定价会改写吗？ #英伟达推动5000亿美元AI基建融资 Grayscale has submitted Form RW to the SEC, officially withdrawing the ETF applications for Cardano, Hedera, and Polkadot. No shares were issued; the applications were directly canceled.
After the news broke, ADA, HBAR, and DOT dropped more than 2% in 24 hours. HBAR fared worse, having fallen about 30% in the past two months, now around $0.068. Market expectations for the "altcoin ETF channel" are cooling down.
This does not mean Grayscale is bearish on altcoins, but under the triple pressure of regulatory windows, liquidity, and institutional demand, preserving the mainstream (BTC/ETH) first is the key.
Do you still hold $ADA, $HBAR, or $DOT? Will this withdrawal affect your position decisions?
#现货ETF资金分化，BTC卖压仍在 #本周三CPI公布, will the pricing for a rate hike in September be rewritten? MicroStrategy has been selling coins recently, and I see some friends are worried about MicroStrategy crashing and dragging down $BTC
Actually, there’s no need to be afraid at all:
1. It’s normal for MicroStrategy to sell coins; they are an investment institution, and their profit model is buying low and selling high. The initial claim of "never selling" was just to rally more people to support the price, so don’t fully trust it. Even if they don’t sell this cycle, they will sell in the next.
Moreover, they sold coins in the previous cycle a few years ago; this is not their first time selling.
2. This cycle, MicroStrategy won’t crash because they won’t be troubled by the annual 1.76 billion in interest.
In fact, MicroStrategy has many options. First, issuing common stock. Their common stock issuance capacity is about 24.6 billion. Second, issuing perpetual preferred stock. Their perpetual preferred stock issuance capacity is 25 billion. Third, selling $BTC , whose total BTC value is 58.5 billion.
Also, when they really stop caring about goodwill, the interest they must repay annually is only 2 trillion. Not paying other interest won’t be illegal; their contract terms are very flexible.
3. Even if MicroStrategy sells coins to repay debt, BTC’s current market cap is 1.2 trillion, with a daily spot trading volume of 30 billion and a daily derivatives trading volume of 150 billion. Such huge trading volume won’t be crashed by an annual 1.76 billion sell-off.
#Strategy再卖1690枚BTC，企业财库出现分化 Long $BTC positions
Mainly because I believe the CPI data should be fine, at least not exceeding expectations.
Regarding oil prices, although they rose in July, the average price was not much higher than in June.
Regarding wages, last week's wage data showed both annual and monthly rates lower than previous values, especially the monthly rate which was significantly lower. Wages are an important component of the cost of goods and services.
Regarding consumption, the July consumer confidence index (preliminary) was lower than the previous value, indicating no demand growth to push prices up.
Regarding the "pubic hair theory," we previously discussed that the non-farm payroll data was a stepping stone for the Fed, and the Labor Department's CPI data should be similar. The goal is to prevent the Fed from raising interest rates in the near term. Tomorrow at 8:30 AM ET, the Bureau of Labor Statistics drops the July inflation read. Most traders will stare at the headline figure. The ones who profit will be watching what happens after the first five minutes of chaos. Here is what the tape is actually telling us right now: The Macro Backdrop June headline inflation came in at 3.5% year-over-year, down from 4.2% in May, with the month-over-month figure falling 0.4% seasonally adjusted. Core inflation held flat for the month. The 10-year Tre#现货ETF资金分化，BTC卖压仍在
I frankly say: ETF funds flowing in and out is a normal market phenomenon; there is no need to panic and turn bearish on the market just because of a few days of outflows. In the short term, BTC selling pressure has indeed accumulated, but the long-term foundation remains unshaken.
Recently, BTC spot ETFs have shifted from net inflows to net outflows, while Ethereum ETFs are still slightly attracting funds. Coupled with rumors circulating in the market about whales and miners continuously transferring out and selling, many traders have panicked and declared that the market has turned downward. In my view, this is an overreaction.
There are two main reasons:
First, whether institutional ETFs or corporate treasuries, they do not only buy without selling. When there were continuous net inflows before, no one loudly proclaimed a bull market was established; yet after just a few days of outflows, some declare a market top, which is being led by short-term emotions. Funds flowing in and out is a healthy market norm.
Second, the divergence between BTC and ETH is not a withdrawal of funds from the crypto market, but rather a rotation of funds within mainstream assets. Under a zero-sum game, this seesaw effect of one rising while the other falls is common, and it does not mean either side is completely weakening.
My own strategy remains conservative: the long-term base position remains firm, neither aggressively chasing highs due to a few days of net inflows nor panicking to cut losses due to short-term outflows. From a short-term market perspective, the overhead supply and selling pressure are indeed heavy, making a quick breakthrough of previous highs difficult; however, the support below is also strong, limiting the downside space. If adding to positions, it is prudent to wait for a pullback to stabilize and clear signals before acting.
In cultivation, one must avoid restlessness; in trading, one must avoid chasing highs and selling lows. Let short-term funds flow in and out as they will The headline $1.1B weekly inflow masks a more useful signal: marginal demand is beginning to separate across the two assets. On Aug 10, Bitcoin ETFs saw roughly $91M of net outflows while Ether ETFs added about $5.3M.
For BTC, that matters alongside reported whale sales and miner transfers to Binance. ETF demand does not need to disappear for the balance to weaken; it only needs to absorb less of the available supply. CPI may support risk appetite, but sustained flow divergence would argue for watching market depth, not just cycle narratives. Not advice, just analysis.
#BTCETHETFFlowsDiverge#财报观察员：AI基建财报接力登场
The first batch of SpaceX lock-up expirations didn't crash; instead, it bounced back to the IPO price. I actually think we need to be more cautious here. Before the unlock, the market was unanimously bearish, so the first wave might have been a short squeeze; on August 20, about 7% more restricted shares will be unlocked. Whether the second round of chips can be held after the short squeeze ends will be the real stress test.
What’s more worth watching this week are the AI industry chain earnings reports. Lumentum and Coherent focus on 800G/1.6T optical communication demand, CoreWeave looks at GPU leasing and AI computing power demand, Applied Materials examines wafer fab capital expenditures, and Cisco can verify whether enterprise AI network upgrades have truly started spending.
Putting these earnings reports together just happens to answer one question: whether the huge AI Capex spent by Microsoft, Google, and Meta has actually translated into real orders along the chain from computing power → optical communication → networking → semiconductor equipment.
If all these lines continue to grow simultaneously, the current high valuation of AI still has earnings support; if order slowdowns and margin declines begin, the market will start to reprice the AI Capex cycle.One piece of data makes me very emotional: the average cost basis for BTC ETF holders is $83,000.
And now BTC is at $65,000.
That means most institutional investors who bought BTC through ETFs are currently at a loss. The average unrealized loss is 22%.
Do you know what this means? They have a strong incentive to sell when the price rebounds to around $80,000—because they finally break even.
This is the so-called "selling pressure from breaking even."
Between $83,000 and $65,000, there is a huge potential sell wall. Every time BTC rises a bit, a group of people say, "Finally breaking even, selling now."
This is why BTC struggled to rise after rebounding from $58,000 to $65,000. It's not that there is no buying interest; the buying is absorbed by the selling pressure from those breaking even.
My judgment: for BTC to truly start a new upward rally, it needs to effectively break through $70,000. Because above $70,000 is where most ETF holders actually start making money—above that, selling pressure will actually decrease.
But $70,000? I don't see any catalysts in the short term. Last week, Strategy "sold at a loss" 1,690 BTC at an average price of $64,262, using all proceeds to repurchase STRC preferred shares, while also raising $650 million net by issuing 6.58 million new MSTR common shares (six times the funds raised from selling BTC); this seemingly abnormal loss operation is actually a capital maneuver for the company’s transformation from "pure BTC holding" to a "digital credit framework," aiming to bolster its USD reserves up to $4.65 billion to ensure payment of preferred dividends and debt interest, thereby maintaining overall financial stability.