Orbit Post Sitemap
Over the past two years, the capital market has believed in one thing: the more money AI spends, the more money it will earn in the future. But yesterday, this logic was challenged for the first time. Nvidia's stock price plunged nearly 5%, not only because of valuation but also because capital began to worry: AI infrastructure is entering a "circular financing" phase. The latest news shows that Nvidia is discussing guarantees for OpenAI's approximately $250 billion data center financing, and may also support about $350 billion in GPU procurement financing. In other words, chip manufacturers begin helping customers with financing, and then customers purchase their own chips. More noteworthy is that the CDS (Credit Default Swaps) of tech giants like Nvidia, Alphabet, and Meta have recently risen in tandem. CDS does not reflect profit, but rather the market's pricing of corporate credit risk. Capital is no longer asking: Can AI change the world? Rather: Who will ultimately pay for these investments? AI has moved from the tech story to the cash flow test for the first time. $NVDA NVIDIA | AI GPU Leader $AMD Chaowei | AI Accelerator Alternative $MU Micron Technology | HBM Storage Chip $INTC Intel | Server Chip $MRVL Mywell | Network Chip $SKHYNIX SK Hynix | Seven Giants in Storage HBM AI Internet & Large Model Technology $META Meta | Open-Source Large Model and AI Computing Power Investment $MSFT Microsoft | OpenAI CollaborationMeta's Quarterly EPS Declines, but First-Year EPS Grows: How to Break Down Tax Benefits and Period Differences
Meta's Q2 diluted EPS was $6.18, down 13% from $7.14 in the same period last year; However, diluted EPS for the first half of 2026 was $16.62, up from $13.56 in the same period last year, representing an increase of about 23%. The comparison of the same table with opposite directions is not due to data inconsistency, but rather because taxes, expenses, and investment items from Q1 and Q2 overlap over a half-year period.
Q2 revenue was $60.801 billion, operating profit was $18.775 billion, and net profit was $15.848 billion; In the same period last year, these were 47.516 billion, 20.441 billion, and 18.337 billion respectively. This quarter's revenue increased, but both operating profit and net profit declined, proving that costs and expenses are the main pressures. The effective tax rate was raised from 11% to 16%, which also weakened the net profit after tax.
First-half revenue was $117.111 billion, operating profit was $41.647 billion, and net profit was $42.621 billion; In the first half of last year, the figures were 89.83 billion, 37.997 billion, and 34.981 billion respectively. It is worth noting that net profit in the first half of the year was higher than operating profit, mainly because the core business profit was better, as half-year tax revenue was $2.113 billion, compared to $3.935 billion in tax expenses in the same period last year.
Quarterly taxes are completely different. Q2 pre-tax profit was $18.756 billion, with income tax expense of $2.908 billion; Pre-tax profit for the first half of the year was 40.508 billion NTD, but a tax gain of 2.113 billion NTD was recorded. This means the first quarter includes tax impacts sufficient to change the direction of the previous six months. If the release only cites half-year EPS growth, it will misrepresent the tax benefits of the first quarter as operational improvements in the second quarter.
The number of shares also has an impact, but not the main direction. Q2 diluted weighted average shares of 2.566 billion shares, compared to 2.57 billion last year; The first half of the year was 2.565 billion, compared to 2.58 billion last year. A slight decrease in the number of shares helped with earnings per share, but since there were no common share repurchases in the company's cash flow statement this quarter, the decline may reflect the cumulative results of past buybacks, share issuances, and incentive settlements, and cannot be attributed to new buybacks this quarter.
The cash flow caliber also shows differences over the period. Q2 operating cash flow was 31.862 billion NTD, free cash flow 784 million NTD; In the first half, operating cash flow was 64.088 billion NTD, and free cash flow was 13.17 billion NTD. Free cash flow in the first half of the year remained positive and large, but lower than last year's 18.883 billion NTD. Nearly zero quarterly results reflect concentrated capital expenditure in Q2, making it impossible to directly infer that full-year free cash flow remains at the same level.
So when reading Meta's quarterly EPS, you should at least keep the columns for single quarter, first half, pre-tax, and after-tax. It can be confirmed that both Q2 operating profit margin and EPS are under pressure, and it is also confirmed that net profit continued to grow in the first half; You can't write the first-half EPS growth as Q2 EPS growth, nor treat the first quarter's tax gains as sustainable operating income. Next quarter's comparison should first look at operating profit, then compare tax rates, number of shares, and non-operating items.#财报观察员: Microsoft's cloud revenue surpasses 100 billion, but Meta's guidance is disappointing—Is the AI story diverging? In late July 2026, the tech giants' earnings season once again unfolds in dramatic contrasts. Microsoft announced revenue of $90 billion for the fourth quarter of fiscal year 2026, up 18% year-over-year; Microsoft's cloud business reached $59.3 billion, up 27%, Azure grew 43%, and full-year Azure revenue surpassed $100 billion for the first time. Over 30 million paying users of Microsoft 365 Copilot. Full-year revenue was $331.8 billion, up 18%; Operating profit was $155.2 billion, up 21%; Net profit was $133.7 billion, up 31%; Earnings per share were $17.95, up 32%. Commercial Residual Performance Obligations (RPO) climbed to $678 billion, up 84% year-over-year. The stock price rose 7% to 10% in after-hours trading. Almost simultaneously, Meta announced Q2 2026 revenue of $60.8 billion, up 28% year-over-year, with advertising revenue at $59.4 billion, both showing impressive growth. However, earnings per share were $6.18, down 13% year-on-year and below expectations, operating expenses surged 55% to $42 billion, and operating margin was compressed to 31%. Full-year capital expenditure guidance narrowed to $130 billion to $145 billion, Reality LMETA $META BOUGHT BACK ZERO STOCK LAST QUARTER AND BORROWED $24.9 BILLION INSTEAD
A year ago the same quarter had $10.2B of buybacks and no new debt.
Long-term debt now sits at $83.7B, up from $58.7B at year end. Property and equipment on the books hit $225.7B, up nearly $50B in six months.
The dividend survived at $1.35B. The buyback did not.
Every spare dollar, plus $25B of borrowed ones, is going into data centers now.AI Earnings: A Tale of Two Extremes — Microsoft Turns AI into Contracts, Meta Still Drawing Promises
Last night after the US stock market closed, the tech world witnessed an annual drama — Microsoft and Meta released earnings simultaneously, one surged 9%, the other dropped 7%.
This isn’t just earnings season; it’s clearly the "final exam results" for AI. Microsoft scored 90 points and went home for a feast, Meta scored 60 and still needs to catch up.
Microsoft: Turning AI into Business
How strong is Microsoft’s earnings? Revenue hit $90.01 billion, up 18% year-over-year; adjusted EPS was $4.74, while market expectations were only $4.24. The highlight was Azure cloud business — up 43% year-over-year, with accelerating growth, and annual revenue surpassed $100 billion for the first time. Microsoft 365 Copilot paid users exceeded 30 million, up from 20 million last quarter.
What excited the market even more was Microsoft lowering its FY2027 capital expenditure guidance from $190 billion to $175 billion. They’re still spending, but less aggressively, and the money spent is starting to come back — commercial remaining performance obligations reached $678 billion, twice the annual revenue.
In short: Microsoft’s AI has paying customers.
Meta: Revenue Is Not Bad, But Where Did the Money Go?
Meta’s situation is a bit awkward. Q2 revenue was $60.8 billion, up 28% year-over-year, which isn’t bad. But net profit was $15.85 billion, down 14% year-over-year, far below the market expectation of $18.5 billion.
Where’s the problem? Costs and expenses surged 55% year-over-year, with R&D accounting for one-third of revenue. What scared the market more was Meta raising its full-year capital expenditure floor from $125 billion to $130 billion, while the ceiling remained $145 billion. Free cash flow collapsed to $784 million, a four-year low.
Q3 revenue guidance midpoint is $62.5 billion, below the market expectation of $63.2 billion.
Zuckerberg said on the call: "We won’t sell computing power for short-term gains." Translation: I’ll keep drawing the pie, keep burning money, just wait a bit longer.
Market Reaction
With two tech giants reporting on the same day, the market voted with real money — Microsoft’s shares jumped over 9% after hours, Meta’s dropped over 10%. Nasdaq 100 futures rose 0.5% accordingly.
The AI investment logic is undergoing a fundamental change. The market no longer asks "How many GPUs do you have?" but "Can your GPUs make money?" Microsoft answered with 43% Azure growth and $100 billion cloud revenue, Meta is still answering with "Better future ahead."
Trading Direction and Strategy
Short term: The post-earnings gap up for Microsoft will likely form short-term support. After-hours rose to around $425, today’s open will likely be higher; chasing the rally requires caution, but a pullback to the $415-$420 range is a good opportunity.
For Meta: It has already fallen 10 consecutive days, down over 14% cumulatively. Earnings added another hit. Short-term sentiment is extremely pessimistic, but note — revenue is still growing, ad business remains strong, AI recommendations have improved ad efficiency. Panic selling is often followed by a corrective rebound, but bottom fishing requires patience and waiting for stabilization signals.
Mid term: The AI sector is shifting from "hype" to "performance." Microsoft represents the "AI monetization success" paradigm, Meta represents the "AI still in investment phase" paradigm. Capital will continue to flow from the latter to the former. A hedge strategy of going long Microsoft and short Meta remains effective under AI divergence logic.
Position management: Volatility is intense after earnings, heavy one-sided positions carry great risk. It’s recommended to build positions in batches with stop-losses; for Microsoft, consider a stop-loss at $410, for Meta, at $520.
---
Trading Insights
Some practical points.
First, earnings trading is about "expectation gaps," not "good or bad." Meta beat revenue expectations but the stock still crashed — because the market priced in not just revenue but profit, guidance, and cash flow. Microsoft beat revenue expectations and the stock soared — because the market didn’t expect Azure to accelerate or capital expenditure to be cut. When trading earnings, you must first understand "what the market has already priced in."
Second, the AI sector has entered the "verification stage." In the past two years, whoever told the story rose. Now, whoever has real contracts with real money rises. Microsoft’s $678 billion remaining performance obligations are the best moat — customers have locked in money for years to come. Where is Meta’s AI revenue? Still in the "future." The market’s patience is limited.
Third, don’t fight the trend. Microsoft rose 9% after hours; if you short thinking "it’s overbought and will pull back" — this mindset is the easiest way to lose money in momentum-driven earnings moves. When the trend comes, go with it first, wait for momentum to fade before thinking about reversal.
Fourth, Meta’s lesson: Good company ≠ good stock. Meta’s fundamentals remain strong, with ad revenue of $59.3 billion, up 27%. But the stock has fallen so much from its peak — because the market is repricing its AI investment return cycle. Trading requires distinguishing "Is the company good?" from "Will the stock go up?" — they are sometimes two different things.
Fifth, keep some cash, don’t go all in. Earnings season isn’t over; Apple, Amazon, Nvidia are still to come. The AI divergence trade is just beginning, there are many opportunities, don’t hang yourself on one tree.
---
The above is a personal trading review note and does not constitute investment advice. The market has risks; trade cautiously.
$SNDK $META $BTC
#微软逆势下调资本开支，盘后涨8.5%
#美联储三票主张加息，今晚PCE成新看点
#财报观察员：微软云收入破千亿，Meta却指引拉胯——AI故事分化了？ #微软逆势下调资本开支，盘后涨8.5%
Microsoft cuts capital expenditure against the trend, after-hours up 8.5%
Why did Microsoft’s capital expenditure cut cause a surge, while Meta’s increase was slammed: what exactly is the market rewarding?
This generation of Wall Street is no longer the sucker for AI. Both are aggressively buying GPUs, but Microsoft’s lowered capital expenditure forecast rose 8%, while Meta’s increased spending guidance was hammered down 8%.
Many friends find it strange—aren’t they both in an arms race? Why is the treatment so vastly different?
The reason is simple: the market no longer pays for unlimited dreams; it rewards capital discipline that can self-sustain and punishes money-burning black holes with no visible investment returns.
Some might ask, Meta’s Q2 revenue of $60.8 billion also hit a record high, so why is it not favored?
Just take a look at its cash flow statement and you’ll understand. Meta’s Q2 capital expenditure surged to $31.08 billion, nearly doubling year-over-year, directly causing its quarterly free cash flow to shrink to only $784 million. Free cash flow plummeted 91% year-over-year—what’s the difference from throwing money directly into a furnace? Even more frightening for investors, Zuckerberg raised the full-year capital expenditure guidance to a range of $130 billion to $145 billion.
Buying GPUs can build core barriers, but when your free cash flow is nearly drained, Wall Street’s patience runs out.
What about Microsoft? Microsoft’s Q2 capital expenditure was $41 billion, even larger than Meta’s. Why does the market cut it some slack?
Because Microsoft played a smart accounting trick and delivered solid results.
Microsoft extended the lifespan of its data centers from 15 to 25 years, effectively lowering its full-year capital expenditure guidance from $190 billion to $175 billion on paper. Although it’s just a depreciation period adjustment, it sent a strong signal to the market: I’m starting to enforce capital discipline.
More importantly, Microsoft’s AI is genuinely generating revenue. Azure cloud business growth accelerated to 43% this quarter, and Copilot’s paid enterprise seats surpassed 30 million.
I did the math myself—I happily pay $20 a month to renew Microsoft’s productivity tools because they are essential. But although Meta’s Llama model is technically impressive and open source, where in daily life can you actually find an entry point to pay for it?
Without a consumer-level paid loop, the high AI infrastructure spending is just a catalyst for on-balance-sheet liabilities.
So, this round of earnings collisions gives us a clear warning. The AI narrative bubble period is over; next, the market will ruthlessly clean out projects that only burn money without generating cash flow. The same logic applies in the crypto market—projects riding the AI halo but whose protocol revenues don’t cover market maker and token distribution costs are risky to even glance at.
Next, I will focus on tracking the actual capital expenditure realization rates disclosed tonight by Amazon and Google. Only by confirming that the capital return slope of cloud giants hasn’t collapsed can the left-side valuation bottom of the entire tech and AI sector be truly solidified.📈 Bitcoin Reclaims $65K, Straight Back Into Resistance
Post-Fed bounce. Bitcoin recovered from $62,383 to push above $65,000, now at $64,905, up 1.47% on the day. The Fed held rates as expected, but price is right back at the resistance ceiling that has capped every rally this month.
📈 Where it stands:
Price: ~$64,905 (up 1.47%)
Bounced from: $62,383 after the Fed decision
Capped by: the $65,500 to $66,500 supply zone
Fear and Greed: 28, still fear
The Fed held at 3.5% to 3.75%, a sixth straight pause, but three policymakers favored a hike and the tone stayed cautious on inflation. That's a hawkish hold, exactly what the market feared: no relief, just confirmation that rate cuts aren't coming soon. Bitcoin bounced anyway on the removal of uncertainty, but it's a relief rally into resistance, not a trend change. On the 1D, price is still trapped under the descending trendline and the $66,500 cap that has rejected it four times.
Three fresh risks now sit overhead, and they're real. US-Iran fighting reignited, oil surged 6.6% reviving inflation fear, and the CLARITY Act got delayed again. Each one pressures risk assets. Without renewed ETF inflows, this move leans on short covering and the $9.6 billion options expiry tomorrow, where max pain sits at $64,000.
What to watch:
Close above $65,200 and hold through the expiry, and $66,500 then $68K open up.
Reject here and lose $62,500, and $60,000 comes back into play.
A hawkish hold plus a bounce into resistance is not a breakout to chase. Let $65K confirm or fail after tomorrow's expiry, then act.
Relief rally clears $66K, or another rejection back toward $60K?
Not financial advice. $BTC $ETH $SOL The market is flat at 64k, with both Meme and AI stocks lying down. Only SNDK has a strong long-short ratio #Fed calls for a three-vote rate hike, PCE becomes a new highlight tonight. @BTC Xingchen @OKX Chinese
Looking at OKX's late report, the market showed no major fluctuations, BTC currently at 64,253, up 0.44%; ETH at 1913, up 1.11%; SOL was at 73.96, up 1.13%. The overall market volume shrank, with a total market value of 2.27 trillion yuan and a 24-hour trading volume of 64.4 billion yuan.
(1) Sector performance: Funds are watching and waiting, no one is taking a heavy stance
The Meme sector fell 1.05%, with a turnover of over 41 million yuan. The artificial intelligence sector fell 0.34%, with a turnover of 15.93 million yuan. The new currency sector edged up 0.57%, with a turnover of 349 million yuan. All three sectors are flat, indicating that funds are in a wait-and-see mode before the interest rate meeting and PCE data release, with no clear main theme hotspots.
(2) Contract long-short ratio: Retail investors are still bullish, but SNDK's data is off
BTC long-short ratio is 1.61, ETH long-short ratio 1.48; retail investors are generally bullish, but not extreme. What really stands out is SNDK's long-short ratio of 6.81—SanDisk's contract long positions are nearly seven times longer than short positions. Even as SanDisk plummeted, retail investors were still frantically bottom-fishing and going long. This is not a sign of improved mood, but a tipping point of greed and panic. If SanDisk continues to fall, these long positions will be lambs waiting to be slaughtered.
(3) Announcement interpretation
The Federal Reserve has kept its benchmark interest rate unchanged for five consecutive times. After the FOMC statement, the probability of a rate hike in September decreased. Previously, three officials supported rate hikes, but after the announcement, the market instead lowered its expectations. Robinhood's Q2 event contract revenue surpassed crypto trading, reaching $156 million. Entertainment trading is eating into the flow of cryptocurrency. Robinhood earns more from sports betting than crypto trading, indicating that the market now needs to "gamble" rather than "buy."
(4) Tonight's PCE data remains the most critical variable
Although the probability of a rate hike has decreased, the PCE data at 8:30 PM is the real deal. If PCE exceeds expectations, all analyses must be redone. If the condition meets or falls short of expectations, short-term risk appetite may rebound.
(5) Summary
The market is moving sideways, sectors are lying flat, the long-short ratio is bullish but not extreme. SNDK's clustering of bulls is a risk signal.$ETH Guys, is CZ going to personally play memes? They just publicly supported all meme coins and even threatened that in the coming weeks, they might buy some selling points themselves to test — does this wave of enthusiasm need to ignite first?
But the real drama will be tomorrow: Deribit's $10.3 billion options expire, BTC's biggest pain point is at 64,000, ETH's pain point is at 1800, and both bulls and bears will be on edge at this level. $SHIB
On-chain is even more unsettled: Solana just printed 250 million USDC, 3,000 BTC quietly transferred to Kraken, Bitfinex outflowed another 127 million USDT—big money is flowing undercurrents, see for yourself. $DOGE
Additionally, DCG invested $45 million to expand the Zcash mining farm, Malaysia set up a power theft mining farm, and the Korean stock market rebounded violently, with related ETF tokens soaring by 11%.
Sensitive nodes all depend on option delivery and CZ's next move. Do you feel upward or downward? #美联储三票主张加息, tonight's PCE becomes a new highlight. #财报观察员: Microsoft's cloud revenue surpasses 100 billion, while Meta's guidance is lacking—Is the AI story diverging? So, $META was hit by a $3.58B set of one of expenses…
- Legal fees
- Severance packages
Without this, it gets a little better:
- EPS: +3% YoY
- Operating Margin: 37%
- Operating Income: +9.4% YoY
I’m not concerned long term. Yeah, earnings will be compressed short-term, but could offer compelling opportunity to accumulate.
$METAAt 8 p.m. Beijing time, there was still an hour and a half left before the official opening of the U.S. stock market,
Throughout the pre-market session, market sentiment was repeatedly pulled back, with intense bullish and bearish tug-of-war.
The three major stock index futures showed a clear divergence,
Dow futures edged slightly higher, with gains holding around 0.25%,
S&P 500 futures edged up 0.03%, with very weak gains,
Nasdaq futures as a whole weakened under pressure, falling by 0.81%, while selling pressure on the technology growth sector remained unresolved.
Internal trends in large tech stocks are split and divided,
Microsoft's pre-market rise of over 7% thanks to better-than-expected cloud business earnings reports, supporting the Nasdaq's slight base,
Meta's cash flow shrank sharply in Q2, plunging over 8% in pre-market trading, dragging down the overall tech sentiment,
Nvidia and Apple only experienced slight positive fluctuations, with investors showing strong cautious sentiment.
The tightening pressure on the U.S. debt side has not eased throughout,
The 30-year Treasury yield remained steadily above 5.23%, marking its highest level since 2007,
Long-term bonds continue to be sold off by the market, indicating that risk-free returns keep rising,
The valuation compression pressure on high-valuation tech cyclical stocks will persist according to Sina Finance.
The memory chip sector reversed its intraday trend back and forth,
In early trading, the market rallied collectively on the oversold recovery rally, with SanDisk's pre-market peak nearing 6%.
Near the opening, bullish support weakened, funds cashed out and exited, and the market turned to weaken,
As of this point, SanDisk's pre-market rebound was slight, with gains of less than 1%.
Micron, Western Digital, and SK Hynix have all narrowed their gains, with some stocks turning negative and declining.
The core logic behind it is very clear,
The early rebound was merely a technical recovery after consecutive sharp declines,
There is no fundamental support for a trend reversal.
The Fed's hawkish tone has been fully priced in by the market, and the three committee members are signaling a commitment to raising rates.
This has caused expectations for rate cuts this year to keep being postponed, making the window for liquidity easing look far off.
Negative expectations of capacity expansions by major Japanese and Korean storage giants and long-term overcapacity,
It still firmly suppresses the entire storage sector. After a brief rebound, uneven selling pressure will continue to emerge.
During the next hour and a half before market opening, it is highly likely to maintain a narrow range of fluctuations,
It is difficult for a strong rally to rally again.
Once the US stock market officially opens, SanDisk is most likely to show a pattern of surges and pullbacks,
After a slight higher open, it was suppressed by selling pressure, continuing the trend of consolidating and bottoming out.
Do you plan to take advantage of this slight rebound to reduce your position and hedge risk-taking, or continue holding and wait for the bottom to stabilize? #美联储三票主张加息, PCE becomes a new highlight tonight. #微软逆势下调资本开支, up 8.5% in after-hours trading. #财报观察员: Microsoft Cloud revenue surpasses 100 billion, but Meta's guidance is disappointing—Is the AI story diverging? $SNDK $SKHYNIX $BTC Trivia: The P/E ratio of QQQ has decreased rather than increased from the beginning of the year until now
QQQ tracked P/E ratio: Early 2026 32.3x → Now about 30x
NDX expected P/E ratio: Early 2026 27.4x → Now about 23x
In other words
"This rally is driven by earnings growth, not P/E expansion inflating stock prices"
🔴 While the P/E ratio is contracting, revenue is still growing significantly. The latest quarterly actual revenue of QQQ component stocks:
・Nvidia Data Center 75.2B, +92% → +36B (Q1)
・Alphabet Total Revenue 119.8B, +24% → +23.2B (Q2)
・Microsoft AI Business ARR 37B, +123% → +20.4B
・Meta Total Revenue 60.8B, +28% → +13.3B (Q2)
・Google Cloud 24.8B, +82% → +11.2B (Q2)
・AWS 37.6B, +28% → +8.2B (Q1)
This is the confidence behind achieving new highs without P/E expansion. If the P/E ratio expands driven by sentiment, there is potential to go even higher.
🟠 So where will this correction go? Currently, several points worth watching:
1⃣ Margin balance 1.53 trillion, up 51.5% year-over-year
2⃣ Net credit balance −1.06 trillion, historical low (cash buffer depleted)
3⃣ Leveraged ETFs 218 billion, 4.5 times that of 2020
4⃣ CoreWeave CDS 855bp, included in QQQ since 6/22
5⃣ Four cases after peak earnings growth rate, three dropped ≥15%
If QQQ falls to $600 within a year, that’s a 19.5% drop from the peak, with historical baseline probability around 40%, and the above five points increase the likelihood somewhat. The chance of falling to 600 is higher than expected.
🟢 Looking longer term, global AI sales (Q1 2026) are 25 billion, with AI depreciation at 21 billion in the same period, leaving a buffer of only 4 billion. The top four QQQ companies (Microsoft, Alphabet, Amazon, Meta) have purchased 433.9 billion in equipment over the past four quarters, which must be amortized over several years:
・Over 6 years → an additional 65–75 billion per year
・Over 4 years → an additional 95–110 billion per year
Current annual depreciation is 149 billion; adding this batch will bring it to 220–260 billion annually, an increase of 40–70% at once, and this only counts already purchased equipment—more will be bought next year.
Meanwhile, revenue is also increasing, and at a higher rate. Alphabet adds about 93 billion annually, Meta about 53 billion; just these two companies add twice the new depreciation. But some companies are already diverging:
Alphabet: Revenue +24%, Operating profit +30%, revenue outpacing costs.
Meta: Revenue +28%, Operating profit −8%, costs outpacing revenue.
Both are spending heavily; one is still profitable, the other is already being eaten up. The future depends on whether revenue or depreciation grows faster.
$QQQ The Korean stock market has nearly spiraled out of control.
In just 40 days, the index has pulled back nearly 40% from its peak, with over $2 trillion in market value evaporating into thin air.
Foreign capital flight, leveraged capital stampede, and a large number of forced liquidations have poured in—multiple pressures have accumulated. This is no longer just an ordinary adjustment, but a sign of a market structure fracture.
This round of market activity began at a historic high in June, KOSPI fell rapidly, and foreign investors' net selling in July reached 18.5 trillion Korean won. Various leveraged ETFs, margin accounts, and algorithmic trading continued to be passively dumped, amplifying the downward momentum.
Ironically, Samsung and SK Hynix just announced record quarterly profits, yet their stock prices continue to decline.
This is enough to prove that funds are no longer trading corporate fundamentals, the core conflict has shifted to concentrated leverage clearing, liquidity continues to dry up, and market confidence has completely collapsed.
A conventional pullback cannot erase $2 trillion in market value in a short time.
The biggest problem with the Korean stock market at this stage is that the original price discovery mechanism has failed.
Many traders always feel that extreme market conditions are far from them.
When market sentiment and liquidity shift, no matter how good the performance, it is hard to withstand valuation contraction and a chain of stampedes. Use leverage cautiously during downward cycles; do not blindly buy bottom-fishing on the left side. #美联储三票主张加息, tonight's PCE is a new highlight Just saw K33 data—BTC spot daily trading volume dropped to $2.2 billion, the lowest point since November 2023. Volume shrank throughout July. But in this uncontested market, BTC stubbornly jumped from 64,568 to 64,938, up nearly $400 in half an hour. RSI hit 88.91, BOLL is on the upper band of 64, and 967 is right above the head.
In contrast, the SNDK girl opened at 1,182 and held on, holding out to 1,007 with a 44% floating loss, still asking if it's time to cut off. Some people on CORE are complaining about losing 79%. On one hand, BTC is shrinking and climbing close to 65K; on the other, fake counterfeit traders are in a bloodbath.
I know this script too well—it's not that the market is out of money, it's that money is concentrated in one place. Only when retail investors are worn down in the altcoins will BTC reveal the real direction.
$BTC $ETH $SOL美股盘前有点意思
不是全面修复
而是 AI 和存储主线在硬拉
微软 MSFT 财报后盘前大涨
英伟达 NVDA 跟着受益
市场重新买回 AI CapEx 逻辑
存储这边也要盯紧
Micron MU
Western Digital WDC
Seagate STX
SanDisk SNDK
如果 AI 资本开支继续上修
HBM DRAM NAND SSD 都会被重新定价
美股七雄里
MSFT NVDA AAPL AMZN GOOGL META TSLA
今晚重点看苹果 AAPL 和亚马逊 AMZN 财报
一个看消费电子
一个看 AWS 和 AI 支出
但 Meta META 走弱
说明资金不是无脑买科技
这更像财报驱动的结构性反弹
不是所有票一起涨
以上内容仅为市场分析，不构成投资建议
投资有风险，入市需谨慎$SNDK $ETH $BTC These are my insights and reflections from reading and studying Abu's "Price Behavior," recorded for my review and reflections in crypto trading. The writing is lively and colloquial, suitable for friends with some background in price behavior studies to discuss and learn together. Please do not repost. A few days ago, I reviewed three market structures: breakout (narrow channel summarized as breakout), channel, and trading range. The market continuously cycles within these three structures. Sometimes the evolution seems orderly, but more often the market is disorderly. Structural changes come suddenly—it could be a sudden news event triggering panic or good news, a deliberate sell-off by a dog farm, or even some technical issue at the exchange. But honestly, as a long-time insider, I believe that coins like Dabing and Estee, which have good depth and strong consensus (Suozi and other fake dogs are not included), have participants including ancient whales, various institutions, American and Asian capital, exchanges, market makers, etc. It's very difficult to maintain long-term strong market control. During normal trading hours (often scammed on weekends), the price movement of each U is the result of thorough negotiation among all parties. As long as there is thorough competition, there will definitely be a gradual development process on candlesticks ranging from minutes to hours to daily and weekly levels. In other words, the balance of power between bulls and bears is key to how to find, hypothesize, and confirm signal K from macro to micro levels within the critical conversion zone, in order to plan trades correctly. Trading Background: Refers to all the current candlestick movements to the left of the candlestick. It reflects whether the current market is in a breakout or trading range, how many times the current structure has driven or fluctuated, and how long it has fluctuated, revealing where liquidity is (Microsoft's latest financial report for Q4 of fiscal year 2026 is: this performance is very solid and highly valuable, completely dispelling previous market concerns that AI investment would burn money and be hard to monetize.
This quarter, Microsoft's revenue reached $90 billion, up 18% year-on-year, significantly exceeding the market expectation of around $87.7 billion; Net profit was $35.8 billion, a year-on-year increase of 31%. It is clear that profit growth far outpaced revenue growth, indicating that the company is not relying on scale to boost performance, but is boosting overall profitability through high-margin AI and cloud businesses, with operational quality continuously improving.
In my view, the biggest highlight of this financial report is the Azure cloud business, with full-year revenue surpassing $100 billion for the first time and quarterly growth maintaining a high level above 40%. This made me realize that Microsoft's AI commercialization has moved from the initial investment stage to a phase of large-scale monetization. Combined with data from over 30 million paid seats in the M365 office AI assistant, it is clear that its AI ecosystem is no longer just a concept, but a core business that generates stable revenue.
At the same time, I found that Microsoft's business structure is particularly healthy. Traditional businesses such as Office and enterprise software provide stable backing and ample cash flow; Rapid growth in cloud computing + AI new business drives incremental growth, with dual support from both old and new businesses, providing strong risk resistance.
Overall, I believe this financial report fully proves that Microsoft, relying on a complete ecosystem of cloud computing power + AI applications, has firmly established itself as a global tech leader. Its fundamentals are solid and its growth logic is clear, and the commercialization of AI has achieved impressive results. #FinancialReportObserver #FinancialReportObserver: Microsoft's cloud revenue surpasses 100 billion, but Meta's guidance is lacking—Is the AI story diverging? #SpaceX获 $1.6B US military contract, stock price plunges and sparks controversy. On July 29, the US Space Force announced it had awarded SpaceX a $1.6 billion contract, planning to carry out 18 Falcon 9 rocket launches by the end of 2027 to deploy military satellites for detecting and tracking aerial targets. However, on the day the news was released, SpaceX's stock price fell instead of rising, closing at $112.55, down 3.32%, halved from the June 16 high of $225.64, with its market value evaporating by about $1.18 trillion. The market's attitude toward this major order is clearly divided.
Bullish view: Military orders are solidifying the foundation
Supporters believe this contract further solidifies SpaceX's core position within the U.S. defense system. Since the beginning of this year, SpaceX has secured at least $7 billion in Pentagon contracts, covering key projects such as the satellite tracking system for the Golden Dome missile defense program and space data networks. Some analysts point out that SpaceX is reconstructing the U.S. military's space communications, reconnaissance, and launch infrastructure from the ground up, transforming into an indispensable "central nervous system" for global operations. The high certainty and long-term nature of government orders will provide companies with stable cash flow support.
Bearish: Lifting the ban selling pressure and cost concerns
Bears focus on multiple risks. First is the pressure of unlocking equity locks—after SpaceX released its first earnings report, about 911.5 million shares of insider and employee shares will be unlocked, accounting for 20% of the unlocked shares, creating huge potential selling pressure. Second, to meet U.S. defense compliance requirements, SpaceX is promoting "de-Sinicization" of its supply chain, which may lead to increased capital expenditures and pressure on gross margins. In addition, the deep ties between government and business have sparked controversy: 10 federal officials hold shares in SpaceX, regulatory approvals are seen as "green light" all the way, and the risk of forced public pension management to take over cannot be ignored.
Conclusion
SpaceX is caught in the paradox of "orders overwhelming" and "stock prices keep falling." In the short term, equity unlocking and financial report verification will be key variables for stock prices; In the long run, whether its military business moat is sufficient to support valuation return remains to be tested by the market. For investors, at a juncture where good news and negative news intertwine, it may be even more important to be wary of excessive emotional reactions.Bitcoin holds firm amid pressure from the Fed and geopolitical tensions The crypto market continues to face a lot of pressure as the Fed signals that it can still raise interest rates, while escalating Middle East tensions caused oil prices to rise 8% overnight. Even so, Bitcoin remained around the $64,000 mark, even as the Dow Jones fell 2.2% and the Nasdaq fell to a three-month low. In my opinion, the fact that BTC retains an important price range in the context of many unfavorable factors shows the resilience of the market.The South Korean stock market is rebounding, and the real protagonist is not Samsung, but the AI semiconductor cycle is being repriced.
Samsung Electronics' latest earnings report sends a strong signal:
Demand for AI infrastructure remains robust, and high-bandwidth memory (HBM) is becoming the core resource in the global computing power race.
Previously, the market doubted the AI investment cycle due to the semiconductor sector's sharp decline, but Samsung's performance proves:
AI is not just a story; it is turning into real orders and profits.
The biggest change in the South Korean stock market is that capital is starting to refocus on:
HBM → AI servers → data centers → computing power infrastructure
This industry chain is still in an expansion phase.
Samsung and SK Hynix, as global storage giants, are becoming the "shovel sellers" of the AI era.
In the short term, the market will still trade based on valuation, capital expenditure, and competitive pressure.
But in the long term, the global AI arms race is just beginning.
In the coming years, what may truly be scarce is not AI models, but the chips, storage, power, and infrastructure that support AI operations.
This rebound in the South Korean stock market is essentially trading on an expectation:
The AI cycle may not be over; it is just reshuffling.Last night the Fed's drama ended, let's talk about the market.
Bitcoin is now hovering around $64,000-$64,400, Ethereum just above $1,900. After the decision was announced last night, it first dropped about 1%, hitting a low of 63,890, then slowly recovered some.
First, the biggest event last night — the FOMC.
The result: 9 votes in favor of keeping interest rates unchanged, 3 votes against, advocating a rate hike. The Cleveland, Minneapolis, and Dallas Fed presidents jointly voted for a rate hike. This is the first time since 2016 that three dissenting votes aligned.
Nominally no rate hike, but the signal from the 3 opposing votes is even more hawkish than a hike. The market's expectation for a September rate hike has surged to 82%. Previously, the probability priced in for keeping rates steady was about 65%, so the market volatility after the result wasn't large, but tightening expectations remain strong.
Geopolitical issues are also adding chaos.
After the Iran attack, the US and Saudi Arabia launched a joint retaliatory strike against Iraqi armed forces. Oil prices jumped nearly $4 to $83. Inflation pressure remains high, so the probability of a September rate hike stays elevated.
ETF fund flows continue to diverge.
Bitcoin ETFs have seen net outflows for the fourth consecutive day, with BlackRock's IBIT withdrawing $54.83 million in one day. Over the past 7 days, Bitcoin ETFs have collectively outflowed 3,170 BTC, roughly $200 million.
Ethereum is the complete opposite. On July 29, another 2,000 ETH (about $3.8 million) flowed in. Over the past 7 trading days, Ethereum ETFs have accumulated inflows of 20,277 ETH, worth $38.49 million. BlackRock's ETHA is the main buyer. The trend of money moving from Bitcoin to Ethereum continues.
From a technical perspective, Bitcoin faces short-term resistance at 64,700, with support at the lower Bollinger band around 63,500, and further down at 62,500. Ethereum is slightly weaker than Bitcoin, with clear resistance at 1,935 and strong support between 1,885-1,890.
In short, Bitcoin is stuck hovering around 64,000 — no rate hike from the FOMC means short-term negative factors are mostly out, but the 3 opposing votes imply a real possibility of a September hike, so funds are cautious. On the ETF side, Bitcoin is flowing out while Ethereum is flowing in; institutions are rebalancing, not fleeing.
Those with positions should watch 63,500 closely; if it breaks, look for 62,500. For those looking to buy the dip, wait for Ethereum to pull back near 1,900 or try light positions on Bitcoin around 63,800-64,000. There's no need to chase highs at this point; wait for clearer direction. $BTC $ETH#财报观察员: Microsoft's cloud revenue breaks 100 billion, but Meta's guidance disappoints—Has the AI story diverged? The AI story hasn't cooled off but has shifted from a "broad growth narrative" to a "cash flow and return validation period." Microsoft is the AI "shovel seller": strictly speaking, the 100 billion refers to Azure's annual revenue. In the latest quarter, Azure grew 43%, Microsoft Cloud revenue reached $59.3 billion, up 27%, and Copilot paid seats exceeded 30 million. This indicates that massive AI capital expenditures have already converted into observable cloud revenue, orders, and profits. Microsoft earnings report
Meta's issue is not weak demand but a mismatch between input and output: quarterly revenue grew 28% to $60.8 billion, with both ad volume and price performing well; however, expenses surged 55%, and free cash flow dropped to $784 million. The Q3 revenue guidance midpoint is $62.5 billion, below the market expectation of $63.14 billion, while full-year capital expenditures remain high at $130–145 billion. Meta earnings report and market expectations comparison
Core divergence: Microsoft's AI revenue path is "compute → cloud services → enterprise subscriptions," a short monetization chain; Meta currently follows more of a "compute → models/recommendations → ad efficiency and future new products" path, with indirect returns and a longer realization cycle.
However, Meta's profit decline also includes legal fees and layoff costs, so it cannot be entirely attributed to AI. A more accurate conclusion is: the market now rewards companies that have proven AI can generate profits and punishes those that still rely on long-term commitments to justify capital expenditures. Microsoft leans toward certainty, Meta leans toward high elasticity and high execution risk. #OKX星球话题来啦 $XMETA Amid the bear market, Bitcoin's native Layer 1 smart contract narrative continues to break through. This article provides a comprehensive analysis of the underlying logic, mechanism details, ecosystem competition, bull-bear logic, and core risks of Alkanes protocol's official flagship token $DIESEL, providing a comprehensive in-depth investment study suitable for direct collection, review, and industry benchmarking. Disclaimer: This report is based on publicly available on-chain data, official documents, and industry media information for analysis and compilation. It is for blockchain industry research reference only and does not constitute any investment advice. Crypto assets are highly volatile and projects are in early iteration stages. All participants are advised to make independent judgments, make prudent decisions, and bear their own profits and losses. 1. One-sentence core summary: Alkanes is a BTC L1 native smart contract meta-protocol officially launched at Bitcoin block height 880,000 (March 18, 2025). Without relying on any Layer 2, it can directly deploy WASM smart contracts, composable tokens, and native DApps on the Bitcoin mainnet. It is the next-generation technical upgrade for Bitcoin programmability after BRC-20 and Runes. $DIESEL As the official launch of Alkanes and its flagship token aligned with the Bitcoin halving model, it replicates BTC's scarcity output cycle and is defined by the industry as the "native crude oil asset of the Bitcoin alkane ecosystem." Its technical narrative is hardcore and top-tier capital background, but it has clear shortcomings such as monopoly on minting mechanisms, lack of ecosystem empowerment, cooling of the overall market cycle, and community voice diversion, making it an early stage of "technology first, rich narrative, and implementation yet to be verified."The U.S. Central Command officially announced that the large-scale strike against Iran's Revolutionary Guard Corps targets has been completed, and has confirmed that this round of military operations has officially concluded. After the news confirms, short-term emotional turbulence is inevitable. Geopolitical warming will temporarily stimulate safe-haven demand, and Bitcoin is very likely to see a short-term rally. But one key point to focus on is that the U.S. has already put an end to it. Market panic over a full-scale escalation of the US-Iran conflict will quickly subside, and the trend is very likely to play out the classic scenario of "safe-haven good news fulfilled as bad news," with gain-taking suppressed after a rally.
Geopolitical conflicts are essentially sudden short-term variables that are difficult to shake the core logic of the market. Currently, the medium- to long-term market is still driven by Fed policy, with delayed rate cut expectations and US Treasury yields remaining elevated. Bitcoin's direction has always followed global risk appetite; this geopolitical event mainly creates localized fluctuations and does not form a sustained trend. The focus will return to the evening's U.S. inflation and GDP data; only economic data can determine the final direction of the current range. In terms of trading, avoid chasing news impulses; beware of rapid spikes and pullbacks. Once the price confirms it holds a key resistance level or effectively breaks through support, then follow the trend. $BTC Chengdu was pouring rain 🌧 today, so I didn't go out to deliver food. Reality was a mess—the girlfriend left in the BMW owner's car 🥀. The market was equally brutal; at 1200 HKD, SanDisk was immediately liquidated 💥. Unwilling to give up, he continued to buy the dip 📉 using the meager salary from Planet Settlement. Wind and rain and market conditions both pressed down, hoping only to cross to the other side. Let's talk about several reasons for SanDisk's recent continuous plunge: 1. The previous gains were too wild, rising from a low point all the way up to above 2000, with massive profit-taking orders fleeing and triggering heavy sell-offs and stampedees 📊; 2. The market has begun to question the story of unlimited expansion of AI computing power, worrying that cloud providers' capital expenditure growth is slowing and storage demand expectations are cooling; 3. South Korean memory giants are planning large-scale capacity expansion, with capital preemptively betting on future chip oversupply, signaling a turning point in the flash memory price hike cycle; 4. Changxin Technology's IPO raises market concerns about the rise of domestic storage, which will long squeeze overseas manufacturers' profit margins; 5. Delayed rate cut expectations have led to collective valuation sell-offs in high-valuation technology sectors, with the high-end storage sector bearing the brunt. Everyone is waiting for the August 5th earnings report to set the direction ⏳. The main force is pushing the market, so just bear with it! If the main players are selling down, just endure it! Institutional whales are trading back and forth between long and short positions, so we can only endure it! They say profits are borne by carrying them, but even when carrying orders, you have to set clear boundaries ⚠️ for yourself. The U money saved from working in the wind and sun while delivering food has already lost quite a bit, so you can't go all in without holding back. Hopefully, this time, the market will no longer let down those who are unwilling to give up. $SNDK $BTC $SKHYNIX Aave's recent gradual shutdown of low-adoption assets and networks should not be interpreted as a view of any L1 or L2. The goal is simple: to reduce Aave's operational, technical, and economic risks, so it can focus on higher-impact priorities, such as developing existing high-value markets and expanding into securities financing. L2s remain crucial to Ethereum's user experience. For example, Aave App's stable vault uses L2 as its accounting layer, which will guide mainstream users into DeFi. Similarly, networks like Avalanche play a key role in bringing RWAs on-chain through institutional business development, an area Ethereum has long lacked.
Previous reports stated that Aave will phase out 50 low-adoption asset reserves and shut down 6 chain deployments.---
**Step 1: Macro Tone — This Is Not an Ordinary Pullback, It's a Structural Risk-Off**
From the data: Among 5 tokenized US stock ETFs, **1 up and 4 down**, with an average decline of **-1.93%**, total trading volume $154 million.
On the surface, it looks like a decline, but look at the structure:
- $XSPY: **-1.03%, but trading volume 87.86 million, more than 1.5 times the recent average**. Volume-driven drop = panic spreading, not retail selling, but institutions rebalancing.
- $XSOXL: **-8.01%, volatility 24.2%**. This is not a slow decline, it's a stampede. Semiconductor ETF crashed from 112 to 86 in one day, meaning someone is using algorithmic trading to hammer the price.
- $XSNDK: The 3x short NASDAQ ETF, which should have risen, actually fell 2.46%. **Even the shorting tool dropped, indicating shorts are also being squeezed** — this market defies normal logic.
A key RSS headline says: **"Bitcoin seesaws as hit to risk sentiment and surging oil offsets Fed rate hold"**
In plain terms: oil prices are rising, the Fed is on hold, geopolitical tensions (Iran) are flaring, and risk appetite is shattered.
**This is not a problem with tech stocks themselves. This is global capital repricing risk.**
---
**Step 2: Capital Flow — Smart Money Shifts from Indexes to Innovation, But Hasn't Returned to Bottom-Fishing Yet**
Looking at the gain/loss structure:
- Biggest declines: $XSOXL (-8.01%), $XSNDK (-2.46%), $XSPY (-1.03%)
- The only gainer: $XSKHY (+2.45%, ARK Innovation ETF)
This data tells you: **money is fleeing large-cap indexes and high-leverage ETFs, flowing into the battered innovation sector**.
$XSKHY volatility is 12%, volume 5.52 million — volume isn't large, but the direction is clear: someone is tentatively bottom-fishing tech growth stocks.
But note: **volume hasn't exploded**. Total volume $154 million, down from over $200 million in previous days.
What does this mean? Big money hasn't entered yet. **Right now it's retail playing, smart money is still watching.**
---
**Step 3: Crypto Positioning — Is BTC a Safe Haven or a Follower?**
BTC at $64,790, 24h +0.44%.
Compare: XSOXL down 8%, BTC barely moved. XSNDK down 2.46%, BTC still steady.
**This is the key signal — crypto did not plunge alongside tech stocks this time.**
RSS headline says: **"Bitcoin flat at $64.6k amid rates, Iran jitters"**
During geopolitical tension and rate uncertainty, BTC didn't crash, indicating someone is supporting it.
Another headline: **"Bitcoin ETF inflows return as Ether funds slip into outflows"**
ETF money is flowing back into BTC, while ETH is seeing outflows. **Money is concentrating in BTC, not chasing gains but seeking safety.**
BTC's current position: **a gray zone between risk asset and safe haven**. When tech stocks crash, it becomes the "less bad" choice.
---
**Step 4: Resonant Conclusion — Not All Declines Are Opportunities; Some Pits Are for Cleanup**
When XSOXL drops 8% in one day, some in the group will definitely shout "bottom-fishing semiconductors."
But data tells us: **volume-driven drop + 24% volatility + shorting tools also falling = the market is undergoing a structural cleansing**.
Smart money is shifting from indexes to innovation but hasn't fully entered. BTC is playing dead, but ETF money is flowing back.
**The current market question is not "to buy or not to buy," but "to survive until tomorrow or not."**
When all signals say "run," if you rush in — you’re either going to make a killing or suffer a heavy loss.
**My judgment: wait until XSOXL stabilizes with shrinking volume, volatility drops below 10%, and shorting tools start rising before making a move.**
Until then, weld your hands to the chair.$BTC July 30, Deribit data shows that crypto options with a notional value of $10.3 billion will expire and be settled in concentrated settlement tomorrow, marking a relatively large options window in the near term. Overview of Core Holdings Data: · BTC options have a nominal value of $9.5 billion, a put/call ratio (PCR) of 0.28, and a maximum pain point of 64,000 USDT; ETH options have a nominal value of $819 million, a put/call ratio of 0.59, and a maximum pain point of 1,800 USDT. 1. Brief Explanation of Key Terms 1. PCR Put-to-Call Ratio Value<1 = More Call Option Position. BTC 0.28, ETH 0.59, indicating that mid- to long-term long positions are dominating options market participants. ⚠️ Key point: Concentrated long positions do not necessarily mean the market will rise; Large bullish positions also easily become targets for major shakeouts. 2. Biggest Pain Max Pain When the settlement price reaches this level, the vast majority of option buyers lose money while option sellers maximize their profits. Under the capital competition, the pre-settlement market is pulling closer to the pain point price; But this is only a probability reference, not an inevitable outcome, and cannot be taken directly as a target price. 2. Objective Market Prediction 1. BTC Current Status The biggest pain point is 64,000, with the current price oscillating within a range for a long time. If the price continues to approach 64,000 near closing, it is a pull caused by option funds hedging; If it stays away from the pain point, it means bullish or bearishIran currently feels a bit "overwhelmed" with my feeling. Iran's recent toughness is betting that Trump won't dare to get stuck in Iran's geopolitical quagmire before the midterm elections
The direction of the gamble is correct, but personally, I feel the timing is a bit lacking. Real negotiations require learning to be superficial, to encircle three and leave one behind, letting the opponent follow your own pace, but not to push the opponent too hard
But given that Iran's attack on the U.S. base in Jordan this week feels very abrupt, and while the U.S. has maintained relative restraint in retaliatory strikes, Iran has threatened to continue its strikes?
This is a deliberate attempt to escalate conflicts and embarrass the U.S., and I believe there are only two possible reasons for doing so
1. Iran's backing supports this move, with someone dragging things down, so naturally Iran becomes tougher.
2. Iran is betting on its own interests, believing that as long as it remains tough, the U.S. will keep softening because Trump dares not launch a large-scale attack
But regardless of which direction is involved, in the short term, it will increase geopolitical risks for the US and Iran, which is not very favorable for global energy prices, inflationary pressures, and your investment accounts! #美军空袭伊朗, oil prices surged and then retreated 🚨 Markets Are Repricing the AI Trade The latest move in equities shows that investors may be shifting their priorities in AI. Microsoft’s roughly 8.5% after-hours move wasn’t driven by an aggressive spending increase. Instead, the market responded positively to expectations around lower capex. That’s an important change in mindset: From “spend aggressively to win AI” ➡️ to “prove the spending is translating into efficiency and returns.” That shift also fits with what credit markets have been si$SNDK SanDisk has been halved since July. If you're the one still holding on, this is written for you.
Falling 50% from the peak, this is no longer a "pullback"—it's a complete crash. Your account may have gone from floating profits to deep traps, and every day you just glance at your phone and want to delete the app.
But have you ever thought about a question? What changes have occurred in SanDisk's fundamentals during the period of its axling?
Last quarter's revenue grew 251% year-on-year. NAND contract prices are still rising. It is a pure flash memory company spun off from Western Digital, and the logic of market repricing after independence is not yet complete.
So why was it cut in half?
First, the previous rally was too fierce, doubling 26 times from the bottom, and profit-taking could pierce through the market. Second, the emotional contagion of SK Hynix's miss, the storage sector was struck down with a single blow, regardless of good or bad. Third, news about Burry shorting storage was repeatedly amplified on Twitter, causing retail investors to panic and stamp on it.
None of the three reasons were that "SanDisk's products were no longer selling."
I'm not telling you to hold out in this position. If your position is beyond what you can handle, reduce when necessary—mindset is more important than position. But if your reason for reducing your position is "a 50% drop is too scary," then you need to think carefully about what the person on the other side seeing at the moment you sell.
He saw a company with 251% revenue growth being cut in half. He's picking up things you threw away with trembling hands from fright.
Historically, every ticket that multiplies several times has experienced at least one halving. Nvidia fell 66% in 2022. Tesla fell 75% in 2022. SanDisk is now down 50%.
Are you holding another Nvidia or another LeEco? This question is worth answering with your brain, not your emotions.
Flipping through financial reports. Look at the order. Look at the contract price. If these things haven't changed, then only the price and your mindset will change. The price will recover, but if your mindset collapses, it's really over.🚨 Same AI Revolution. Very Different Investor Reactions.
Why did Microsoft get rewarded while Meta faced a tougher market response?
The answer may come down to one thing: visible monetization.
Microsoft has been able to demonstrate strong AI demand through committed contracts and a substantial backlog, giving investors clearer evidence that the spending is translating into revenue.
Meta, meanwhile, continues to grow rapidly, but the market is focusing more heavily on the enormous investment required to build its AI infrastructure—and how long it may take before that spending produces comparable cash returns.
The message from Wall Street is becoming clearer:
AI potential gets attention.
AI revenue gets rewarded.
Microsoft currently has the advantage because its AI opportunity is backed by contracted business and measurable demand.
Meta could change that equation if it eventually turns its AI infrastructure into a major external revenue stream.
Until then, the market appears to be favoring proven AI economics over future promises.
#DailyOrbit
#Fed3Dissents
#MSFTCutsCapex
#AIStoryDiverges
$BTC
$ETH
$SNDK 🧠SKHY's best ecosystem isn't the "contract market," but HBM chips directly on-chain
Don't just treat SKHY as a perpetual stock on OKX (SKHY/USDT on 7/10 16:10 UTC, 24h trading, 8h funding fee, USDT margin).
Behind it is SK Hynix = 56%+ global HBM market share, main HBM3E supply, HBM4 leading the way to Nvidia's Vera Rubin, and AI memory oligarchs whose production capacity will be sold out early in 2026.
This is how the ecological closed loop works:
• Physical layer: For every additional GPU server in an AI data center→ SK Hynix's HBM → company profits anchor AI capital expenditures
• Traditional financial layer: 7/10 Nasdaq ADR (SKHY) listed, ICE group funds + global institutions can allocate directly
• Crypto layer: OKX makes ADR prices perpetual 24h, with unified accounts using BTC/ETH/OKB as margin, allowing sideways trading to generate interest and offset fees; Going forward, XSKHY-type tokenized stocks can also be withdrawn via X Layer and integrate with OKX Pay/Exchange OS
• RWA layer: Platforms like Ondo launched SKHYon tracking ADR in the same week, allowing on-chain portfolios to rebalance "AI infrastructure exposure" with stablecoins and US Treasuries in one wallet
Previously, if you wanted to bet on AI water sellers, you had to open a US account + exchange currency + wait for the market to open; Now, using USDT to open SKHY perpetual in your OKX account is equivalent to buying a "AI hash power utility receipt" using crypto-native methods. $SKHY #财报观察员: Microsoft's cloud revenue surpasses 100 billion, but Meta's guidance is disappointing—Is the AI story diverging?
The JPMorgan trading team has officially issued a buy signal for the S&P 500. The U.S. tactical position monitoring model shows that current U.S. stock positions have dropped to levels with inverse positioning value.
The historical data is straightforward: after similar position changes, the S&P 500 rose about 3% on average over the next 20 trading days, while the average gain across all periods was only 1%.
What is the logic supporting this signal?
U.S. Treasury yields retreated, the dollar weakened, corporate earnings remained solid, Middle East tensions eased, and the Federal Reserve kept rates unchanged. The decline in oil prices eased inflation concerns and opened a breathing room for risk assets. Nearly 80% of companies during earnings season have exceeded expectations, with fundamentals supporting the market.
But this signal has two obvious flaws.
First, semiconductor sector positions are too crowded—after the AI narrative loosens, the most crowded areas are often the first to be hit hard. Second, the US-Iran conflict continues to recur. Trump just said he would "strike Iran," and a ceasefire could turn hostile at any moment.
JPMorgan itself admitted in its report that the market no longer believes that increased AI investment will automatically benefit chip manufacturers; investors are now more closely examining whether large investments can be converted into actual revenue.
One detail worth noting: this is just a tactical signal, not a strategic shift—based on lighter positions, not changes in profit expectations. It's basically saying, "The market has dropped too much, so it's okay to rebound in the short term," not "US stocks are starting a new bull market."
But the bigger problem is: as soon as the signal is given, the market may have to verify it.
Last night, the FOMC kept rates unchanged, but three votes were against it. The 30-year U.S. Treasury yield jumped to 5.228%, the highest since 2007. The S&P 500 fell 1.52%, and the Nasdaq entered a technical correction zone. On the first trading day after the buy signal was issued, the market was declining.
If historical patterns hold, the S&P 500 should have risen 3% over the next 20 trading days. But a shovel company is the first to dig gold in the mine—the ones who sell the signal are often the ones who shout the loudest.
$SPY $QQQ $SNDK 🚨 Markets Are Repricing the AI Trade
The latest move in equities shows that investors may be shifting their priorities in AI.
Microsoft’s roughly 8.5% after-hours move wasn’t driven by an aggressive spending increase. Instead, the market responded positively to expectations around lower capex.
That’s an important change in mindset:
From “spend aggressively to win AI”
➡️ to “prove the spending is translating into efficiency and returns.”
That shift also fits with what credit markets have been signaling, as AI-related bond spreads have reflected growing expectations that the sector is entering a more mature phase—with risk being priced more carefully.
Meanwhile, $BTC and $ETH remain relatively quiet.
Three Fed officials pushing for a rate hike, combined with the upcoming PCE inflation data, is keeping macro uncertainty elevated.
Crypto doesn’t look like it’s preparing for an immediate breakout.
It looks like it’s waiting for confirmation.
The structure isn’t necessarily damaged—the market simply needs the next macro signal before choosing a direction.
Don’t guess the move. Let the data confirm it.
DYOR.
#OKXOrbit
#Fed3Dissents
#MSFTCutsCapex
#AIStoryDiverges
$BTC
$ETH
$SNDK Wall Street's attitude toward crypto assets has finally unified: you can choose not to believe, but you can't avoid paying fees.
Morgan Stanley's Ethereum ETP (MSSE) and Solana ETP (MSOL) have started trading on NYSE Arca, both with an annual fee rate of 0.14%, and staking has been included from day one. The service provider disclosed that about 95% of staking rewards are expected to be passed on to holders.
What really matters is not "two more tickers," but the sales channels. The coins remain the same, but once wrapped in brokerage accounts, advisory systems, and compliance packaging, the buyers change.
Of course, packaging won't eliminate volatility. Staking also brings additional risks such as lock-up, operations, slashing, and taxation. ETPs solve the "how to buy" problem, not "whether the price will drop after buying."
The most interesting part is: in the past, Wall Street researched crypto assets to judge if they had value; now they research more seriously because they can finally charge management fees on them.
$ETH $SOL #MorganStanley launches spot ETPs for ETH and SOL After the FOMC was implemented, the market began to digest Walsh's "new way of communicating."
Core changes:
The Federal Reserve kept interest rates unchanged
Wash made it clear:
The market should look at the "ball" (economy, inflation, yields), not the "referee" (Fed)
He reduces reliance on forward-looking guidance and prefers to let the bond market set its own financial conditions
The results were:
Long-term US Treasury yields rose significantly (30-year yields approached 5.20%)
U.S. stocks came under pressure and declined
Gold has found some short-term support but remains constrained by the high interest rate environment
Currently, gold is still fluctuating above 4100, and market attention has shifted from "whether to raise interest rates" to "how Walsh will reshape policy communications."
Cross-market logic:
Walsh's more "let-go" stance → long-term yields rising on their own→ passive tightening of financial conditions→ putting pressure on risk assets.
For gold:
Rising yields are putting pressure on you
However, geopolitical risk + safe-haven demand provide support
The US dollar has not strengthened unilaterally, leaving some room for gold
Operating Approach:
Still treating as range-bound fluctuations, no chase orders
Focus on whether key support holds up
If yields continue to rise, gold's upside will be limited
If geopolitical conditions heat up or yields fall, gold is more likely to find support
What is more important now is to observe whether the bond market is tightening for the Fed, rather than just looking at the rate decision itself.
Do you think Walsh's approach of reducing guidance is good or negative for gold in the long run?$KAITO This wave feels like the last dance.
Yesterday, it was still surging wildly, with market sentiment overwhelmingly bullish, but today, after waking up, the trend has completely changed.
Many people were still immersed in the fantasy of a rise, not expecting the market to quietly begin to turn.
This rapid rally followed by a pullback is actually the most critical moment for judgment.
I originally planned to wait for the pullback to finish and then push further, but seeing the trend clearly weaken, I decisively went short following the trend.
But the market didn't give it a chance to resist and kept dropping.
Looking back now, those who chased long positions at high levels may have been trapped halfway up the mountain.
That's the rhythm of Yaobi: when it rises, it feels like it could double, but when it falls, it gives you no time to react.
At present, $KAITO's bearish trend is not yet over.
A rebound may happen, but before the trend changes, a rally is more like giving bears a chance to re-enter.
For now, hold onto this move and wait for the market to make a clear change before considering a break.
$KAITO Is this wave a brief correction, or is it the official start of a downward cycle?
What do the Air Force brothers think?
#美联储三票主张加息, PCE becomes a new highlight tonight. #微软逆势下调资本开支, up 8.5% in after-hours trading. #财报观察员: Microsoft Cloud revenue surpasses 100 billion, but Meta's guidance is disappointing—Is the AI story diverging? #财报观察员：微软云收入破千亿，Meta却指引拉胯——AI故事分化了？
Damn! Microsoft has already turned computing power directly into a money printing machine!
Meta is still burning shareholders' hard-earned money like firewood! On the same day of earnings release, the market gave you no face at all, directly tearing the “AI story” to shreds with the stock price: on one side is the ruthless player already vomiting cash wildly, on the other side is the empty shell still drawing big promises waiting for three to five years.
Capital stopped buying the PPT spiel long ago, now it only recognizes one hard truth: every dollar you pour in, can you immediately spit out more real cash? If you can't, then get ready to be ground down repeatedly.
This quarter Microsoft directly showed its cards: cloud business will exceed $100 billion for the year, Azure growth at 43%, and they dare to call 45% next quarter. Copilot paid users exceed 30 million, holding $678 billion in orders, enterprise customers are not just watching, they are already paying and using.
They are not reinventing the wheel or launching some revolution, but forcefully integrating AI into existing cloud and office software, just charging customers a bit more subscription fee. The efficiency is visible, and budgets are willing to pay.
Analysts on X directly named this as the strongest enterprise AI demand signal this quarter; traders are also complaining: one side has already turned investment into auditable revenue, the other is still shouting “wait a bit longer.”
Looking at Meta, revenue growth of 28% looks shiny on the surface, but free cash flow has crashed to a four-year low, some calculated a 90% plunge. Next year they plan to pour 130 to 145 billion into infrastructure, Zuckerberg still stubbornly insists “selling computing power short-term is shortsighted.”
Ads are indeed tuned more precisely with AI, but advertisers only look at final ROI, they won’t pay extra just for AI. 3 billion daily active users, no matter how big, can’t turn into immediate cash flow. Investors on X directly mocked the near-zero buybacks this quarter, heavy borrowing to build data centers, debt soaring; investment banks cut target prices citing uncontrolled spending and unclear returns.
The core difference is here: Microsoft is adding value on existing business with high certainty, orders have locked in most of the next one to two years. Meta is betting that infrastructure and algorithms can bring bigger entry points, which is almost impossible to quantify short-term, with terrifyingly high trust costs.
The market now only asks three things: how much real money has AI made? Has profit margin improved? Can future cash flow be realized? Whoever can’t answer will be abandoned by capital.
Voices on X are even harsher: some say investors have finally separated “AI revenue” from “AI show”; some are willing to bet small positions on Meta’s attention pool, in case AI social or assistant actually works someday, the space could indeed be bigger.
Stop fantasizing that everyone can take off with AI. Capital will no longer spread evenly, it will only pour into those who have already turned investment into a cash flow closed loop.
Microsoft has handed in the exam paper, Meta is still doing homework. Anyone still clinging to the old logic of “more investment means stock rise” is waiting to be taught a lesson by reality. The real selection phase has just begun.＄SNDKUSDT — The Bounce That Lost Its Legs
Bounces look good on screenshots. They feel good in portfolios — until they stop. And this one is stopping.
At $1,053.66**, SNDK is up **+2.90%** in 24 hours — but the high of **$1,125.29 is a distant memory, and the low of $972.00 keeps the bulls nervous. This isn't a breakout. It's a dead cat wearing a bull costume.
________________________________________
The Rotation Has Passed This Sector
Memory chips had their moment. AI narratives are exhausting. Capital is flowing toward assets with clearer regulatory clarity and mature ecosystems — not speculative semiconductor proxies.
The 8 that still matter:
$SOL, $ETH, $KAITO, $ZAMA, $SOON, $ALLO, $ZEC, $XAU
The 92 laggards:
$BEAT, $LAB, $SHIB, $WLD, $UB, $FIL, $LINK, $ONDO, $HYPE, $DOGE, $XRP, $ADA, and every low-liquidity altcoin whose narrative has expired.
SNDK is clinging to relevance by a thread. That thread is $1,043.13 (EMA5). If it breaks, the selling accelerates.
________________________________________
Technicals — Weak Bounce, Weaker Conviction
Price sits above EMA5/10/20 (1,043.13/1,036.17/1,029.67) and SAR at 1,024.65 — bullish posture, but the posture is slouching.
• MACD: DIF 6.17 > DEA 2.36,柱状线 +7.62 — positive but decelerating.
• RSI6: 74.04 — overbought. This bounce is running on fumes.
• KDJ: K 84.28 < D 85.62 — bearish crossover confirmed. The warning is flashing.
Volume: 2.50M SNDK, $2.63B USDT — heavy, but price can't hold the highs. That's distribution disguised as demand.
________________________________________
News Flow — Headline Noise, Not Substance
"Flash memory sector sell-off" headlines are doing the heavy lifting for this bounce — but headlines don't create sustainable trends. The macro pressure and AI fatigue are real. This bounce is reactive, not proactive.
________________________________________
Outlook — Topping Process in Motion
• **Break above $1,100** targets $1,125.29 — but that's a 50-point fight through heavy seller congestion.
• Break below EMA5 at $1,043.13Last night, a friend asked me: SanDisk dropped from 2354 to 1015, should we go long or short now?
I said, don't rush to pick sides.
Because at this level, it's not about "going long or short," but about "who gives the signal first."
NAND fundamentals have not collapsed — AI server eSSD long-term contracts locked to 2028, original manufacturer inventory only 4 weeks, Q3 contract prices still rising. But the stock price has already weakened the bulls, with MACD -132 and RSI 19, indicating a purely bearish alignment.
Go long? and the right side. Shorting? No evidence from fundamentals.
I only look at five signals that can truly take action:
(1) The daily chart closed back to 1300, with increased volume
(2) RSI bullish divergence + MACD golden cross
(3) The Philadelphia Semiconductor Index reclaimed the 60-day moving average, signaling Micron's decline
(4) Spot 512Gb TLC prices have stopped falling
(5) Original factory inventory ≤ 6 weeks, no downward revision in Q4 guidance
How many of the five are shining now? Zero.
Dropping too much is never a reason to buy at the bottom. Five resonances are the true essence.
When you just cut your hair, the most valuable thing isn't the next opportunity, but the 70% principal you hold.
Wait until the 8/5 earnings report + 8/13 investor day to pick the direction, and you'll have a 20% higher chance of earning than now by guessing bulls or bears. $SNDK $SKHYNIX Over the past year, the biggest change in the crypto industry is not a sudden explosion in a particular sector, but a shift in how capital views projects. Previously, the funding market was more willing to pay for the future. As long as a project had a big enough story and a grand enough vision, it could attract funding attention. Next-generation public chains, on-chain social, GameFi, the metaverse, NFT—these directions once attracted massive capital due to market imagination about the future. But now, it's clearly different. The money hasn't left Crypto; it has just become more cautious. What institutions truly care about now has shifted from "Is your story big enough?" to: Do you have real users? Do you have real revenue? Do you have capital and asset accumulation? Can you connect with institutional clients? Can you enter real financial scenarios like payments, trading, custody, and settlement? This is also why the funding market has seen a very obvious change in the past year: Capital is shifting from "investing in narratives" to "investing in entry points." In the past, everyone competed to create bigger spaces for imagination. Now, the competition is about who can become the channel for capital, assets, and users to enter the chain. The market's money hasn't disappeared; it has just started seeking certainty. If you only look at the secondary market, many might think Crypto is still in a phase of rapid hot spot rotation. Meme can attract massive traffic in a short time. AI Agent can quickly become the market focus. Certain on-chain applications can also generate huge trading volumes. The AI industry is now increasingly like real estate: shovel sellers are starting to fingerprint loan contracts for those who buy them.
According to media such as Axios, Nvidia is discussing about $250 billion in financing support for OpenAI's data center project in Ohio, and may also arrange up to $350 billion in chip procurement financing. First, clarify the boundaries: the terms are still under negotiation, not finalized, and may fall through.
This does not mean demand is fake. Supplier financing allows data centers to start construction faster, and Nvidia's profits and cash flow indeed qualify for the guarantee table.
But the risk structure has changed. Previously, NVIDIA only sold GPUs and could see orders; In the future, it may simultaneously be a supplier, investor, and credit supporter. If a client's rent falls short of expectations, the impact is not just on an order, but also on accounts receivable, credit exposure, and market valuation.
In the past, people asked, "Who bought so many cards?" Now you have to ask: "Who paid for the cards?" ”
$NVDA# Nvidia and Google provide massive guarantees for AI data center debt 🚨SOL ETFs have accumulated $1.1 billion in accumulation, but their price has dropped 5.8%—where did the money go in, and who did it dump?
This morning, the FOMC voted three votes to raise interest rates, causing risk assets to shiver collectively. SOL weekly price was -5.87%, but on the US spot SOL ETF side:
• Cumulative net inflow of $1.13 billion across 5 spot ETFs (May alone: $115 million with no outflows)
• Morgan Stanley MSOL amendment 0.14% rate (lowest globally), SEC approval expected at 90%
• Institutional 13F holdings of approximately $540 million (including Goldman Sachs and Electric Capital)
• At the same time: FTX's bankruptcy legacy unlocked monthly with ongoing selling pressure, and the Genesis wallet's 180,000 SOL converted to ETH was marked by ZachXBT
The conclusion is quite conflicted: institutions are slowly buying in through ETFs, while FTX's legacy + short-term leveraged long positions are being dumped. The ETF buys Solana 6–12 months later, while the futures market trades tonight's PCE.
At the $74 level, the ETF channel is open, technical upgrades haven't stopped, and selling pressure is a clear sign. Rather than the "SOL reset to zero" narrative, it feels more like "institutions cutting losses to retail investors."
$SOL Today, the platform token $OKB saw a rare sharp drop, breaking its recent steady decline and becoming the only platform coin among the top ten losers. OKB has long relied on exchange buybacks and burns, platform rights, and compliance expectations to produce a steady and slow bull run, giving it strong defensive attributes in a volatile market. The core logic of this round of decline is indiscriminate follow-up drops amid widespread market panic. Today, US stocks crashed, the crypto world weakened across the board, and the market's selling sentiment spread to the extreme. All coins were slashed, and even the most stable platform coins couldn't remain unaffected. Second, recent platform benefits have been continuously overdrawn, with compliance expectations and traditional financial cooperation benefits all factored into prices in advance, with no new catalyst to further upward movement. When market sentiment weakens, profit-taking is concentrated and triggers a pullback. The stable trend in the market was broken in the short term, with increased volume leading to a decline, and a temporary high level was established. However, compared to altcoins, OKB's decline is limited, with strong support and stable holdings, which is a normal correction and not a trend reversal. Market outlook: OKB is only passively catching up in the short term, not a bearish trend. Once the market stabilizes, platform coins will recover first and remain the most stable defensive asset in a volatile market. In the short term, it avoids the risk of a pullback and remains valuable for allocation even after stabilizing.Today I came across $AEON, and at first glance, it really looked tempting. I chased the rise from Sesame and bought a little bit. As soon as I got in, I got hit 😂😂 No fate to be a small dog, but it’s okay, it’s only 300u total, just like buying a lottery ticket!
To be honest, the AI intelligent agent economy, A2A settlement underlying infrastructure, and backing from well-known institutions like IDG and HashKey make the concept packaging very cutting-edge and easy to get excited about.
But I habitually checked the basic data first and immediately saw a key risk: the circulating market cap accounts for only 18.3%.
In other words, less than 20% of the tokens are circulating and tradable in the market, while the vast majority are locked up by early investors and the project team. For this kind of token, it’s easy to pump the price without much capital; but once a large unlock happens later, ordinary investors simply can’t withstand the selling pressure.
It’s issued on the BNB chain BEP20, listed on only 9 exchanges. The hype looks lively during the hot market, but once the heat fades, large exits will face slippage.
The crypto world never lacks beautiful stories, and AI Agent is currently the hottest sector. Any new project can package a grand narrative. Everyone must distinguish: whether the project vision can be realized and whether the token can surge in the short term are two completely different things.
As for my operating principles, I only hold spot and firmly avoid leverage. For new narrative small tokens like AEON, I won’t bet heavily. My core holdings remain steady with market-tested assets like Ethereum and OKB.
At most, I allocate a very small amount to new projects for entertainment, definitely not as a long-term core allocation.
When you see new projects backed by big-name VCs and hot sectors, is your first reaction to want to invest or to calmly dig into the data first? #美联储三票主张加息，今晚PCE成新看点 $BTC $ETH After Microsoft $MSFT released this Q4 earnings report, based on the after-hours quotes recorded on Microsoft's investor relations page in the US East Coast, the stock price hovered around $425, about 8.8% higher than the regular close. But when I look at the data, the point that stops is actually another set of figures appearing side by side: this quarter's revenue is $90 billion, capital expenditure plus financing leases is $41 billion. One is the speed at which the company makes money, and the other is the speed at which the company continues to pour money into AI and data centers. In recent seasons, Microsoft's biggest question has not been whether it will do AI. The real market stuck: can the money spent be converted into revenue, rather than just staying stuck in the GPU, data center, and OpenAI relationship? Yesterday's financial report at least gave a tougher answer than in previous quarters. Azure and other cloud services revenue grew 43% year-over-year, surpassing the market previously targeting around 40%. Intelligent Cloud revenue was $39.3 billion, up 32% year-over-year. Microsoft Cloud revenue was $59.3 billion, a year-on-year increase of 27%. If you only look at the phrase "strong cloud business," these numbers are too flat. At first, I thought the sharp after-hours rally was mainly due to Azure's impressive growth, but when I checked the cash flow table, I realized the market was actually looking at both sides at the same time. Free cash flow was still $19.6 billion, but down 23% year-over-year; In the same quarter, capital expenditure and finance leasing jumped 69% year-over-year. This is not an easy growthLet's talk about when US stocks will stop falling
Can you still buy the bottom after storing everything?
Yesterday's closing session saw a surge in volume and a decline. From the news side, it appears inflation is intensifying, and expectations of interest rate hikes are starting again. In reality, it's still the US stock gains from March to June and the uncertainty of AI investment. When faced with uncertainty, the market tends to sell blindly.
Currently, there is no sign of stabilizing the decline. The Nasdaq has mostly fallen and risen more than gained, mainly supported by the Seven Sisters. The original logic was hard tech falling, software rising, NVIDIA rising, Apple rising, Google rising. Now, after clearly selling hard tech, funds have abandoned the Seven Sisters
When will this situation improve? The ultimate form of this improvement should be when Nvidia begins to stabilize, stabilize, or even form a strong trend. This requires not only explosive earnings reports but also technological breakthroughs and a new round of AI gaps. When Nvidia rises, semiconductors will recover, storage will stop falling and rebound, and US stocks will rebound.
But now, there's no need to pay more attention to everyone's advice. When I first bought US stocks, the pandemic hit right after I bought them. At the time, I was completely on guard at the summit, but when I checked afterwards, that spot was just a small pit
The advantage of spot trading and regular averaging is that you can endure. US stocks are highly volatile, and in the long run, QQQ still has an average return of 20%. I don't think AI will ever collapse, nor do I think we'll buy Nasdaq at its century-high now.
The market really needs to cool down. Once cooled, only a rational self can buy Google and SMH. Back then, buying and holding prices wasn't just to sell at a slight rise or unbearable with a slight drop.
There are too many people who want to get rich overnight in the US market, but Buffett has long said that speculating on US stock market prices is tantamount to suicide. I remain firmly optimistic about AI tech stocks. Maybe looking back in a few years, now is definitely a great opportunity 😊When people talk about @babylonlabs_io (BABY), the conversation usually starts with a number: how much BTC is staked. It's convenient to measure attention with this, but I don't think it's the most important thing.
For me, the real question is much simpler: Can #bitcoin help secure other networks without requiring people to abandon the principles that first led them to trust Bitcoin?
This is where Babylon becomes interesting.
Under suitable market conditions, anyone can attract capital. The more difficult challenge is building a system that can still operate as expected when conditions are not ideal. Markets change, validators make mistakes, incentives shift, and the network faces unexpected pressures. It is precisely at times like this that protocol design becomes more important than those impressive metrics.
I find what makes Babylon convincing not the idea of making #BTC more efficient. Instead, it aims to make Bitcoin contribute to network security while placing self-custody at the core. In this way, the discussion shifts from "What are users allowed to do?" What can the "transition" agreement actually enforce? ”
These two discussions are completely different.
True trust doesn't come from a fancy dashboard or a constantly growing TVL chart. True trust comes from rules that are transparent, verifiable, and difficult to bypass when problems arise.
This does not mean the model is without challenges. As more networks rely on shared security, complexity naturally increases. The real test is whether these safeguards can remain robust as the ecosystem evolves.
Perhaps a better way to assess Babylon is not to ask how much Bitcoin it attracts today, but to ask whether its security assumptions still hold when the market no longer makes things simple.
@babylonlabs_io
#baby
$BABY
{future} (BABYUSDT)
$BTC
{future} (BTCUSDT)Samsung is investing heavily in a Texas chip plant—who benefits it, and what risks are hidden in it? #英伟达. Google provides huge guarantees for AI data center debt $BTC
Many people only saw one piece of production expansion news: Samsung officially announced that construction of its second wafer fab in Taylor, Texas, USA, will begin construction at the end of this year, aiming for mass production by 2030.
The market simply interprets it as: Samsung ramping up AI to capture orders from US customers.
But few have broken down the three core truths: this is not simply capacity expansion, but the integration of Samsung's foundry breakthrough battle, geopolitical supply chain competition, and long-term AI chip competition; At the same time, this news hints at a divergence in the sector, with short-term sentiment speculation and medium- to long-term supply-demand logic being completely different.
1. First, clarify the core facts of the project
Taylor Plant 1 is progressing and will start production within the year, gradually ramping up to 2nm advanced processes;
Taylor Plant 2 will officially start construction at the end of the year and begin mass production in 2030, also planning to use 2nm processes with long-term compatibility with 1.4nm processes.
The entire Taylor campus reserves enough land to build 10 wafer fabs, and Samsung's overall investment plan in the U.S. exceeds $40 billion, which can receive massive subsidies from the U.S. CHIPS Act.
Key Differentiation: Two Texas factories positioned as [logic chip foundry], not producing DRAM, NAND, HBM storage; The core storage production line remains firmly in Pyeongtaek, South Korea.
This is the most common misconception in most interpretations online.
2. Samsung is investing heavily in building factories in the United States, with four deep strategic objectives
1. Facing direct competition from TSMC and seizing the major AI client in North America
TSMC's Arizona plant in the United States continues to advance, with a large number of orders secured from Nvidia, AMD, and Apple.
Leading North American tech companies are increasingly focusing on supply chain localization, prioritizing foundries with advanced process capacity locally.
The establishment of Samsung's Taylor dual factories sends a direct signal to Microsoft, Tesla, and AI startups: it can provide nearby 2nm foundry capacity, take on orders for AI computing chips and autonomous driving chips, fill the gap in foundry business, and narrow the gap with TSMC.
2. Secure subsidies from the CHIPS Act to balance geopolitical risks
To secure billions in U.S. subsidies, it is necessary to establish advanced domestic manufacturing capacity.
For Samsung, building factories is a ticket to gain policy benefits; At the same time, it has established a dual-base layout of "high-end storage in Korea + local logic foundry in the US," dispersing supply chain risks in a single region and meeting the mandatory requirements for diversified supply chains of U.S. government and enterprise clients.
3. Business Structure Transformation: Building a "Storage + Foundry" Dual Engine
Currently, Samsung's biggest growth pillar is HBM high-end storage, benefiting from AI computing power dividends;
However, storage is highly cyclical, and price fluctuations are unavoidable.
The goal is to vigorously expand overseas logic foundry capacity to reduce dependence on a single storage business. In the future, it will earn profits from the HBM cycle while relying on long-term AI chip foundry orders to secure stable cash flow and smooth out cyclical fluctuations.
4. Lock in long-term demand to address the shortage of advanced production capacity
Samsung made it clear during its earnings call: the pace of global expansion of advanced process production is not keeping pace with the growth in AI chip orders.
Currently, customer inquiries about 2nm and next-generation 1.4nm continue to surge. Securing land and starting construction ahead of schedule is a way to secure advanced production capacity for the next 7-10 years ahead of time.
3. Rarely mentioned online: The duality behind the news, opportunities and hidden dangers coexist
✅ Positive logic
1. Long-term perspective: Foundry business opens a second growth curve, no longer relying solely on storage cycles;
2. Close to core North American demand, making it easier to secure long-term AI chip supply agreements;
3. Improve global capacity layout, increase bargaining chips with customers, and strengthen barriers for comprehensive semiconductor giants.
⚠️ A major negative risk that cannot be ignored
1. The investment cycle is extremely long, and the realization is very distant
Factory No. 2 will start construction at the end of the year, with mass production starting in 2030. A four-year construction + yield ramp-up cycle cannot contribute revenue in the short term, continuously consuming huge capital expenditures and suppressing free cash flow.
2. The cost of building factories in the United States is extremely high
Labor, infrastructure, and energy costs far exceed those in South Korea, with the production cost of the same wafer 20%-35% higher. Even with government subsidies, long-term profitability remains uncertain.
3. There is a bottom line in the technology strategy: cutting-edge R&D remains local
Samsung will not place all next-generation core cutting-edge processes in the US. The Taylor factory will focus on mature advanced processes, while core R&D and HBM storage lines will remain in Korea, meaning the US factory will find it difficult to produce disruptive technological breakthroughs.
4. Intensified long-term competition
In the coming years, TSMC and Intel will simultaneously expand advanced capacity in North America. By 2030, when a large amount of new capacity is released, competition in the advanced foundry track will intensify.