“when we first invested in OpenAI... we invested originally at the $125 billion round. So I don't think people were entirely sold on LLMs as a business model.…”
OpenAI's do-everything strategy risks mastering nothing, but its consumer lead and coming ads engine make it a durable winner
“And the pattern recognition, to answer your question, for me on Anthropic was just reading Dario's essays and listening to him on podcasts.”
Anthropic is winning AI by going all-in on enterprise coding, now the market leader where it once looked like the No. 2 player
“I am more confident in the thesis that the hyperscalers are a worse business model going forward.”
Hyperscalers face shrinking margins as AI labs insource their own compute and customer base concentrates to a few players
“I certainly would not make the case that they are fantastic businesses, but I don't think they're going away like people thought because I think they're…”
Neoclouds survive and matter because they run GPU clusters and inference better than traditional hyperscalers
“there haven't been any shorts in AI... there was like basically no shorts prior to 2026, really... In our letter, I said like there are going to be a lot of…”
Software stocks become a worse business and the first real AI short as vibe-coding erodes everything but systems of record
“Japan is probably going to become a military power at some point in the future again”
ITA is the iShares US Aerospace & Defense ETF, the most liquid defense expression. Sundheim cites Japan remilitarizing and European defense spending as structural tailwinds. US defense contractors benefit from [allied defense procurement programs](https://www.defense.gov/News/Releases/). ITA up 61% over the last year, +0.77% since source date.
“I'm most optimistic in economic growth. It has to be the case that if you believe in scaling laws and you believe in AI, that economic growth will be very…”
AI drives a productivity boom that delivers strong growth with disinflation, the nirvana setup for markets
“I really believe that scale in EVs is going to be important, which is why Tesla is kind of one of the reasons why Tesla's won.”
Tesla wins EVs on manufacturing scale, the moat Rivian never reached, while autos stay a brutal business
“I think there's really interesting things happening in Asia... Japan's probably going to become a military power at some point in the future again.”
Japan and Korea hard-asset and engineering companies are underpriced as Asia rearms and defense spending rises
“I think we are on a collision course with China over semiconductors... to the extent that we don't figure that out, I think, you know, we're going to have…”
A US-China clash over Taiwan semiconductors is a depression-grade tail risk that AI only raises the stakes on