Chip stocks are down 24% from their June 22 peak. Nvidia is up 2.6% over the same stretch. The losses are somewhere else: $CRWV −48% $NBIS −42% $AMAT −37% $LRCX −36% $AVGO −15% $TSMC −11% The market is marking down chip-equipment makers and the debt-funded companies that rent out AI computing, the two groups hit first if AI spending slows or borrowing stays expensive with the 10-year yield above 5%. Expecting market to remain rangebound until mid term election. However, don't see much downside from here.

Hedge funds sold tech at the fastest pace in Goldman Sachs data (back to 2016) in the five days to Oct 15, led by chips. Tech is still 23.5% of their net exposure, above 90% of readings in the past five years. Coming week $META , $MSFT and $GOOGL will report earnings. Let's see whether the sale continues.

$MSFT has pulled out of Nscale's £2 billion AI data center in Loughton, Essex. Power for the site was due by the end of this year. The grid now says it won't arrive until the early-to-mid 2030s. Only last month, Microsoft still counted the site as part of its $30 billion UK investment. Nscale says it will keep building and look for other power sources, and its other Microsoft projects in Norway, Portugal and the US aren't affected. The chips and the customer were both ready. The project stalled because the grid connection is now about eight years late.

Is Japan dumping Treasuries? Its holdings fell $171B between February and August, to $1.07T. But $120B of that drop came from short-term bills. Net sales of longer-dated notes and bonds were just $13B. The timing matches Tokyo's defense of the yen: ¥27.1T (about $170B) of dollars sold between April 30 and May 6, and again in August. So far this is about supporting the yen, not Japan walking away from long-term US debt. Japan's own 10-year yield hit 3.1% in September, its highest since 1996, which gives Japanese investors more reason to keep money at home over time.

The next two weeks will shape how the AI trade looks going into 2027. Big Tech is expected to spend $1.1-1.2T USD on capex next year, and none of them has put a number yet. I'm watching for the first guidance. Here's how the next two weeks go: Tue 20: $SPCX Wed 21: $META, $LRCX, $TSLA, $TXN, $IBM Thu 22: $WDC Wed 28: $MSFT and $GOOGL, same day as the Fed decision, plus $KLAC Thu 29: $AMZN, $INTC What I'm watching: $META hasn't given a 2027 outlook yet. A first range on the 21st sets the tone. $GOOGL said capex will "increase significantly in 2027." Whether it puts a number on it. $LRCX and $KLAC on 2027 chip-equipment orders, which is where that spending lands first.

CXMT's Q3 revenue rose 15-20% QoQ to RMB 114.7-119.7B (about $17-18B) but its bit shipments barely moved. DRAM prices rose 10-20% in Q3. So almost all of CXMT's growth came from higher prices, not from selling more chips (which grew ~5% as per my estimate). Omdia had CXMT at 7.7% of global DRAM in Q4 2025. Bernstein sees about 12% by 2028. For Samsung, SK hynix and Micron, that's the pace to watch.

AI took 82.7% of all US venture capital through September, according to PitchBook and the NVCA. That's about $427 billion, more than all US startups raised in 2021, the previous record year ($358 billion). AI's share fell to 65.9% in the third quarter as the giant rounds for AI labs slowed, and total deal value dropped about 40%. Excluding the AI labs, US startups are raising roughly what they did ten years ago.

A day after the crash, Jensen stepped up for $CRWV. In Stockholm he said Nvidia doesn't need to backstop CoreWeave, and that its cost of capital is actually coming down. The stock liked it. It bounced about 17% off the intraday low, from $59.51 to a $69.69 close. And honestly, the debt side backs him up more than the headlines do. CoreWeave has two very different kinds of debt. The newest secured loan, $8.5B from March, is investment grade at around 5.9% because it's backed by customer contracts. Meanwhile the unsecured 2032 bonds trade under 80 cents, yielding over 15%. Lenders with contracts behind them are calm. Lenders with nothing behind them are nervous. EBITDA covers interest about 2.3x, so this isn't a near-term default story. What I'm watching is whether secured lenders keep the door open. As long as backlog keeps turning into contracts, CoreWeave can fund itself. If that slows, those 15% bonds will have been the early warning.

The Kospi closed down 7.6% on Thursday after a 20-minute circuit breaker. SK hynix fell 11% and Samsung fell 9.5%. Today, we have seen some recovery in the Korean market. KOSPI recovered ~3%, SK Hynix and Samsung both are up ~4%. What set it off: China's CXMT said it will launch server DDR5 on a denser 4F² cell design by year end, and a Korean newspaper column warned of memory oversupply in 2029. Elon Musk added: "China will be making a LOT of RAM in 2028." What I take from it: CXMT says its first 4F² server DDR5 ships by year end, with no volumes or yields given. Bernstein has it at 9% of global DRAM bits in 2026 and 12% in 2028, and Deutsche Bank still sees 2028 demand at 129% of supply. Even if every CXMT wafer got the full 4F² gain, that's ~6 extra points of supply, less than a year of ~22% bit growth. Buyers are signing longer. The same day, KeyBanc's John Vinh said $GOOGL is discussing memory supply deals through 2035, and HP and Dell are extending theirs into 2031. He expects DRAM prices up 15-20% in Q4 and HBM prices up 80-100% in Q1 2027. The thing I'd watch next is whether $MSFT, $AMZN and $META ask for terms like Google's.

$TSMC Q3 results Net profit: $23.9 billion (NT$763.1 billion) vs $23.4 billion (NT$746.3 billion) consensus. Up 69% year on year. Revenue: $46.7 billion, above its own $44.6–45.8 billion guidance. Up 16% quarter on quarter and 51% year on year. Gross margin: 66.4%, down from 67.7% in Q2 because of the early costs of its new 2-nanometer chips. Operating margin was 58.7%, above the top of guidance. Q4 guidance: revenue of $52.3–54.3 billion, about 14% higher than Q3 and 58% higher than a year ago. Analysts were looking for growth of 10–15%. Gross margin is guided at 65–67%. Full year: TSMC now expects 2026 revenue to grow about 44% in dollars, up from "above 40%". Capital spending stays at $60–64 billion. CEO C.C. Wei said demand is extremely robust with no sign of slowdown, and that AI demand will probably grow faster than TSMC's earlier estimate of the mid-to-high 50s percent a year.

Anthropic told investors on Wednesday that it made $16.8B of revenue in Q3. That's up 45% from $11.6B in Q2. Total revenue in first 9 months stands at $33 vs full 2025 revenue of $4.6B. On an adjusted basis, gross margin was 85%, operating margin was above 11%. Free cash flow was positive. Per the FT, it made $1.70 of revenue for every $1 spent on compute. Per Reuters expecting $110B annual run rate by year-end. This is very bullish news for entire AI space. Anthropic has committed at least $518B to compute and cloud over about a decade.

$AMZN is expected to add 42 GW of compute capacity by 2030 and you are bearish?

Everyone's worried about "CoreWeave pays 15% interest." It doesn't. $35.6B of debt. Only one loan is at 15%: $1.3B left, due 2028, and shrinking every quarter. The bonds pay 8.5- 9.75%. The 15% is the secondary-market yield on the 2031 and 2032 bonds. So what's the real picture? Interest is covered. H1 EBITDA was $2.67B against $1.18B of interest, about 2.3x. Gross margin is around 66%, so renting GPUs makes money. Depreciation and interest are what turn it into a loss. Refer the table for 2027 interest coverage projection. The real problem is funding the growth. 2026 capex guidance is $35 to 39B. By my math, about $3 to 9B still has to be raised this year. Most of it can come from GPU-backed loans tied to customer contracts, not 15% bonds. The risks I'd actually watch: 1. Delays. Money is borrowed and spent months before a cluster goes live and starts paying. 2. Customer concentration. Two customers are 62% of revenue. 3. Contracts average about 3 years, some loans run 5. Renewals may come at lower GPU prices. 4. About $3 to 9B still has to be raised this year to fund the buildout. My read: no near-term solvency issue. $104B of backlog and no big bond maturities until 2030. The real risk isn't demand. It's timing. $CRWV

BlackRock's research arm cut US stocks to neutral on Monday. The posts going around say it's because of AI. That's half right. It closed a long-held US overweight "to manage potential volatility around the AI buildout." But inside US stocks it still prefers AI and tech. Part of the money goes to Treasuries, mostly short-dated. The 10-year yield is up 110 bps this year, to about 5.3%. It blames part of the jump in real rates on "intensifying competition for capital." It still favors AI-related companies and listed infrastructure, citing scarcity in critical infrastructure.

Sept CPI came in hotter than expected. Headline: 3.7% YoY vs 3.6% expected Core: +0.3% MoM vs 0.2%, 2.6% YoY My read: a hold is still the base case. One hot core print rarely forces a hike two weeks out. But the market is now taking a hike seriously.

$ASML reported a record Q3 earnings. Sales: $13.2 billion (€11.7 billion), in line with consensus. Up 56% year on year and 25% quarter on quarter. Earnings per share: $14.60 (€12.95) vs $12.13 (€10.58) consensus, a 22% beat. Gross margin 57.5%, above its own guidance. Q4 guidance: sales of $16.0 to $17.2 billion (€14.2 to 15.2 billion), above the $15.85 billion consensus. Full year 2026 is now €44 to 45 billion. The big line: ASML says its planned supply for 2027 and 2028 won't meet customer demand, and it's looking at adding capacity beyond the 30% increase already planned.

Foreigners dumped 4.53T won of Korean stocks yesterday. Biggest single day since July. And 3.72T of it was two stocks SK hynix and Samsung. Over the last 15 sessions they've sold 24T won. The won appreciating from 1,530 to 1,345 per dollar since June, it further incentivises foreign investors to take profit. So who's buying all that? Mostly the companies themselves. Buybacks soaked up roughly 19T. Retail added 2.9T, local institutions 1.7T. Samsung's buyback ended Oct 6, however the retail buyers are coming back which is a good sign. Nothing in the numbers broke. Samsung just reported a record 107T won quarterly profit. Memory supply is moving onto 3 to 5 year contracts. That cash funds the next round of buybacks. SK hynix reports Oct 27. With buyback resuming from SK Hynix and Samsung and momentum from retail investors, we may see the market recover in coming days.

The S&P 500 is 0.2% from its record high. But 78% of the stocks in it are at least 10% below their own highs, and 41% are down more than 30%. A handful of giant companies are holding up the index while most of the market is in a correction. This is not particularly negative rather this is how the US economy is moving. It's the tech that is growing and not the rest of the economy.

Broadcom's CEO just did the simplest math on AI I've seen. He estimates the world spends about $200 billion a year running AI models and gets about $150 billion back in revenue. His per-gigawatt version is more interesting. Anthropic makes about $30 billion a year per gigawatt of computing. That gigawatt costs about $10 billion a year in chips and data centers, so the model maker keeps the other $20 billion. The money is moving up to the model makers, not down to the hardware.