The 'ARR Ouroboros': why VC-funded AI startups selling to each other could trigger a death spiral
menhguin · x · 2026-10-21
The author lays out his genuine AI-bubble bear case: much of B2B AI revenue is VC-funded startups buying tools from each other and booking it as ARR — an "ARR Ouroboros."
- Example chain: company A sells sales agents, lab B sells tokens, company C sells RL environments; C spends seven figures on A's agents to sell to labs, A spends six figures on B's tokens. If C switches to B, A's revenue drops, cutting its token spend, shrinking B's RL-environment budget
- The uptrend is levered end to end: heavy AI usage concentrates in well-funded early adopters — other AI startups spending VC money. Enough players cutting spend could cause a death spiral
- A cited chart adds the key risk: the top 1% of companies drive 80% of AI spend, and 80% of those are venture-backed, so a downturn could mean large customers pulling back simultaneously
In a follow-up the author notes this isn't fraud or incompetence — these are legit positive-ROI use cases; heavy users are just a tiny demographic, and LLM versatility makes seemingly separate businesses a highly correlated ecosystem with overlapping funders, which is why labs are diversifying beyond SWE.
More from AGI Musings
- Margaret Mitchell: click-driven AI misinformation is causing real harm — mmitchell_ai · 2026-10-21
- Mathematician Daniel Litt concedes his AI bet, but says AI won't solve math — ziv_ravid · 2026-10-21
- AI SaaS monetization is cooked: consumers small, enterprises pay $20/month — menhguin · 2026-10-21
- Tim Dettmers: Academia is about to have a renaissance — the best AI work of the next decade happens in university labs — StasBekman · 2026-10-21
- AI saves scientists 7 hours a week, but validation is the new bottleneck — AnneliesGamble · 2026-10-21
- UW's Amy X. Zhang brings CSCW theory to human-agent collaboration in Stanford talk — amyxzh · 2026-10-21