DeepSeek can't buy more compute — everything is spoken for, says analyst
teortaxesTex · x · 2026-09-11
An analyst argues DeepSeek cannot spend more on compute because there simply isn't more to buy — everything is spoken for — and its data spending headroom may also be limited. By contrast Meta and peers can scale both. The quoted take: if data is marginally more useful than compute, data-moat companies benefit, a mildly bearish signal for AI compute capex.
More from Companies & People
- OpenAI FDE lead on building the team: make it work first, scale second — colintjarvis · 2026-09-12
- Jeremy Ashkenas, creator of CoffeeScript, joins Anthropic after 4.5 years at NYT — threepointone · 2026-09-12
- Meta gives every user a free VM for agents, analyst says it could own consumer AI — signulll · 2026-09-12
- David Sacks says Anthropic's IPO must be paused pending ex-researcher claims — JFPuget · 2026-09-12
- Meta Gives Every User a Free 2-vCPU VM With 100GB Storage, Compute Moat Even OpenAI Can't Match — signulll · 2026-09-12
- Microsoft Teams is the purest example of distribution beating product quality, says Ben Casnocha — deepakns · 2026-09-12