Stonks: Prefer low starting multiple + margin expansion + robust topline
Joined April 2020
- Tweets13.9K
- Following1K
- Followers8.3K
- Likes15.4K
Is this true?
Imo the single most bearish chart for the AI trade here
Inference in only ~37% of compute usage
Source @SemiAnalysis_
$SHVA.TA, single digit EV/EBIT on 2027 numbers, as moaty a business as any, double digit revenue growth there after + operating leverage. Too cheap?
Ask yourself what happens next:
-> where will AI build out related debt trade?
-> are ANT/OAI likely to increase or slow down $ burn on training?
-> will VCs pile another 0.5T into AI startups or pull back?
-> as the economy pulls into recession, will firms increase token spend?
Good thing nobody took the unfunded spending commitments seriously!
No one - and I mean literally no one in the AI supply chain - takes the ridiculous $1 trillion of unfunded spending “commitments” from OpenAI and Anthropic even vaguely seriously.
Hence the TSM executives dismissing Sam Altman as a nontechnical “podcast bro” after he came to them to make a commitment.
Just listen to the answers that public company executives give about these unfunded “commitments.” They are functionally irrelevant and this should be obvious to everyone.
Hey AI maxis/math challenged folk: Good news! Think of all the "cheap shares" you'll be able to buy!
So these guys were basically on the money again.
Market now has the "all clear" signal to crash the AI trade.
AI math went from virtually impossible to impossible.
Current AI use is artificially high:
-> 0.5 trilly in non lab AI funding
-> Subsidized AI use
-> Unaffordable frontier AI lab training
S-curve supply assumptions meet airpocket of demand.
Character.AI said on Monday it had filed for Chapter 11 bankruptcy in the US and agreed to sell some of its assets to Walt Disney for $20 million. reuters.com/legal/litigation…
"Commercial bookings increased 23% in constant currency driven by the previously large Azure commitment from OpenAI, reflects multiyear demand needs as well as the previously announced Anthropic commitment from November"
$MSFT Q2-2026 call, oops?
The hyperscalers did not and do not take the labs spending commitments seriously.
If they could get out of the contracts they signed with OpenAI and Anthropic, they would as open-weight inference is more profitable for them than the below market prices currently being paid by OpenAI and Anthropic.
And open-weight models are the only thing keeping most VCs afloat.
You are *exactly* wrong.
Good to know. Now we just need to blow up the 11th and 13th largest companies on earth with 1T in spending commitments and see if anyone minds.
Wonder if the half trillion in non lab VC AI money goosing demand will continue at the same pace, hmm.
Open-weight models that decrease margins at the model layer are positive for AI infra, exhibit 1001.2
Since the @silicondata “token index,” which captures both price and mix, peaked on May 31:
B300 prices are up 38%
B200 prices are up 10%
Other GPU indices are up *much* more.
BetterIRR retweeted
The AI boom is increasingly shifting from parabolic to linear growth. Most investors are totally unprepared for that shift. This week, free on the blog app.
bobeunlimited.substack.com/p…
When OpenAI goes to raise another round and no one is interested, ultimately I think the hyperscalers + nvidia will have to push the can down the road one more time, but it's just throwing good money after bad at that point and they will be punished by the market.
Was fun while it lasted
The most likely long-term outcome of the AI boom will be massive capital destruction in the US, with market share going to cheaper open-source Chinese models, according to Jefferies. bloomberg.com/news/articles/…
Global VC investment in AI reached $430 billion in the first half of 2026 alone.
The problem is not just that frontier models ARR are stalling out, a large chunk of their demand is from unprofitable startups!
Air pocket like 2021 but much bigger.
irishtechnews.ie/global-ai-i…
BetterIRR retweeted
Rising rates and a collapse VC funding could threaten 25% of frontier AI labs' recurring revenue.
That could then trigger a sell off in the broader AI trade.
@WarrenPies unpacks the bear case.
The AI trade is accelerating on record earnings, massive compute demand, and high rental rates for chips.
I sat down with @WarrenPies to discuss the under-the-radar indicators he's tracking in the market right now, his favorite sectors and opportunities for the year ahead, and his latest views on the macro setup.
This conversation will make you think critically about the AI trade.
TIMESTAMPS:
0:00 - Overweight equities
3:30 - Tech-heavy market and AI leadership
5:29 - Macro shocks
11:05 - Low correlations and pair trades
13:37 - Semis and broadening trade
17:35 - GPU availability and rental rates
21:22 - Compute indicators flashing bullish
25:08 - Frontier lab ARR
30:08 - Hyperscalers leadership
32:34 - AI CapEx and the Fed
36:38 - What ends the bull run?
40:32 - Favorite sectors
41:46 - @3F_Research
BetterIRR retweeted
OAI September ARR ~$50B (not ~$70B as previously reported a few weeks ago).
Yet in Ramp data OAI and ANT revenues are similar 🤔
See slides that suggest slowdown in ANT growth since July.
Amazing, unprecendented numbers for new companies - AI adoption incredibly fast by historical standards. But another ~10x over next few years (doubling every year?) seems challenging...
ft.com/content/b66a9858-f8fb…
AI Boom or Bust? – Manifold #122
Are we in an AI bubble?
Analysis of AI CapEx, revenues, OAI+ANT burn rates.
AI revenue has to reach $ TRILLIONS per annum within the next few years (an increase of > 10x; compare total global software revenue of $1.5T today), or else the current CapEx spend will turn out to be a huge misallocation.
Separately, even if total AI revenue reaches these levels, OAI and ANT could still be overvalued due to revenue/margin compression from competition with other US or Chinese labs.
Chapters:
(00:00) - Conference Takeaways
(03:21) - AI Capability Leap
(07:36) - Both Things True
(08:46) - CapEx Debt Risks
(10:31) - Trillion Revenue Hurdle
(18:32) - Spend Plateau Signals
(31:45) - IPO and Bubble Fallout