This memory Intel & SK is working on is going to absolutely mogg. The world is not ready for what’s coming.

🎯

Everyday with Oil >$100 we are one step closer to the mother of all TACO. Still think right play is to be max long - feels like valuations bottoming in some spaces. AI fundamentals fine, everyone launching agents at the same time as well. Maybe not the best 2nd leg of demand (vs. coders) but could be all we need to get us through YE.

$MSFT probably a really good r/r here. Stake in OAI worth $30/sh. $465 for rest of business, ~$20 in GAAP EPS this FY, growing EPS ~20% a year. Azure accelerating, but importantly, Copilot adoption accelerating and shifting to consumption based pricing.

How do we know Intel Foundry is barely in the price of Intel? If you asked Mag7 + AI Labs if they could own their own chip manufacturing supply chain for $400B they would jump at the opportunity 1 Intel Fab is 8 gigawatts a year for $30B in CAPEX Intel Foundry OPEX added to Google would be a 15% bump to R&D, 10% for Amazon etc. If they wanted to invest to produce memory in the future it would be an additional 7.5% and 5% bump. R&D cost is peanuts for them. The margin capture in design and manufacturing of semiconductors is a huge expense for upstream customers. The reason Mag7+AI labs can't enter semiconductor manufacturing is Fabs are an intractable moat. Intel ex product design biz values the fabs at a 1 P/B. Arguably almost 0 if DCAI grows as expected next year. $INTC

Blow-off top in AI losers through next week, then it’s time to wake up

With TSMC’s COUPE CPO slipping, NV is working with $INTC on an all INTC Rubin CPO variant(18A,Intel OCI, EMIB), a 4-die version that would outperform TSMC’s 2 dies

Hey @jimcramer you sold your $BTC just in time please for the love of fucking Baby Jesus sell $INTC and throw a fit over it on live television thank you

🤔 #Intel CEO Lip-Bu Tan says the company won’t miss the next memory wave, eyeing techs like cross-batch memory (XBM) and Z-angle memory (ZAM). But Team Blue knows memory well—it launched the world’s first commercial DRAM over 30 years ago.💡More: pse.is/9hf4xv 🔗

Round and round we go

Chase Coleman’s Tiger Global just doubled down on $INTC Intel stake. Q1: Started with 1.64M shares Q2: More than doubled to 4.3M shares (~$594M) All while trimming $NVDA, $GOOGL and $TSM. Tiger Global is a $50B+ NY growth hedge fund. Big conviction on the Intel turnaround.

UBS is the first wallstreet firm to nail Intel's CAPEX ramp up $27/$30B in 2027 (intra year could increase to $35B IMO) With 28/29 in $40B range However they are off (sorta a given) on 28/29 revenue ramp because then they would have to admit their AMD/ARM share models are wrong. Intel gained revenue share in x86 server (as Intel's mix is heavy lower core counts that are now shifting to higher core counts, unit growth is harmed until mix shift finishes) So UBS ends up being quite bullish on Foundry growth. But then misses that most of Foundry volume is actually IDM CPU for 2027,2028. AMD and ARM will struggle to gain share. That leaves them being bearish on Intel product They also have an weird valuation model. It's simply much easier to look at Intels cleanroom footprint, back into max WSPM in 2029. Then make some assumptions on IDM vs Foundry mix. Then get your EPS, which is well north of $10 $INTC

$INTC bofa with $145 PT and CEO buying at these level 5x oversubscribed upped from $15B to $20B offering is this bullish ? $SPX $SPY

$META is to 26’ what $GOOG was to 25’ Feels like when Gemini 2.5 was hot and Gemini 3 was cooking, and the street had no fucking clue what was going on Meta Watermelon will be interesting

Lot of hype posts about TeraFab Cheapest way to buy into the revenue stream of this Fab is still Intel Put it another way. Intel earns a 15% royalty fee on the value of each logic wafer 38K × 100K WSPM × 15% = $6.8B Intel earns 6.8B per 100K WSPM, 100% gross margin. Each new technology node has a wafer ASP uplift. 7A node could be 75K, so $13.5B per 100K WSPM If Elon runs 1M WSPM of 14A Intel earns $68.5B in profit. 1M of 7A is $135B $INTC $SPCX

OpenAI and Anthropic currently have more demand than compute. Since GPU capacity is limited, they need to maximise how much revenue and gross profit they earn from every GPU-hour. If they can maintain high margins, they can achieve higher ARR and can justify trillion-dollar IPO valuations to raise more equity, secure even more compute and extend their frontier lead. Meta can disrupt this by selling near-frontier intelligence at a much lower price. Because compute is already constrained, OpenAI and Anthropic cannot simply cut prices and make up for it with unlimited additional volume. They would either have to keep prices high and lose market share to Meta, or cut prices and weaken their ARR growth, gross margins and IPO narrative. Meta can afford to tolerate low margins on its models because its advertising business generates enough cash to fund the datacentres. It does not necessarily need to make large profits from AI immediately. So Meta’s strategy may be to start a price war, weaken the labs’ economics and slow down their ability to raise capital and expand compute. This prevents OpenAI and Anthropic from becoming economically unassailable and gives Meta more time to close the frontier gap. $META

i never thought i’d see a meta ai model reach parity with fable and gpt 5.6 at coding but here we are muse spark 1.2 beats gemini 3.6 flash and grok 4.5 in deepSWE. this is metas second model release in <4 weeks. they truly pulled off a miracle and the single common denominator is compute zuck aggressively invested in data centers and its paid off. well done.

Some thoughts post tonight’s EPS calls 1) The comments around $/GW monetization of Rubin GPUs are wild. If they comes to pass, available power (to the extent it can be delivered) should accrue more rent, especially anyone with 2027 capacity to sell or existing compute they can re-price (neoclouds). The Texas audit stuff, while likely zzzz, is short-term chef’s kiss for anyone with approval already or capacity outside the lone star. Ask your Claude to find you those companies! (That one’s for you @GavinSBaker 😝) 2) Elon highlighted memory as a *compounding problem*. I’ve been thinking bit-demand growth of 40-50% vs 20% supply. If we haircut Mr musk and say 75% demand growth, we’ve got a monster glut. Chinese supply can’t help (p.s. they’re getting shorter over there too). And, yes, there could be some tech breakthrough that overwhelms Jevons for memory. But it’s also curious that the engineer with the biggest imagination in history seems to think that’s not a reliablem med term solve. Translation: longer cycle than even bulls think, more FCF => buybacks => lower vol => higher multiple. Aka a beautifully reflexive loop. Greed also has a SHORT memory, so the ANTS will be back… $DDRAM. The extra D is for double :) 3) The collective Lab ARR is likely to surprise in the 2H, limited only by compute deployed. As an example, Anthropic is supposedly guiding folks softly to $145Bn at YE. BUT they just did high teens in July (~$83bn total 😉) and they are bringing on more capacity b/w now and then. Imagine the next 5 months are 20, 20, 25, 25, 30 ==> $200bn ARR. OAI w/ a similar ramp could hit 150Bn. All this makes one want to go full leotard…