Ricursive Intelligence is betting on AI-designed chips to help solve the compute crisis
- ID
- 35642
- Status
- summarized
- Published
- 15 Oct 2026, 9:01 AM
- Fetched
- 15 Oct 2026, 9:49 AM
- Provider
- TechCrunch
- Category
- technology
- Original URL
- https://techcrunch.com/2026/10/14/ricursive-intelligence-is-betting-on-ai-designed-chips-to-help-solve-the-compute-crisis/
- Source URL
- https://techcrunch.com/feed/
Summary
- Score
- 3.0
- Created
- 15 Oct 2026, 9:49 AM
- Tags
- Audience
- developersai_ml_learnerssaas_startup_founders
What happened
At TechCrunch Disrupt, Ricursive Intelligence founders Anna Goldie and Azalia Mirhoseini described building AI that designs chips, learns from each iteration, and improves subsequent designs to automate and accelerate chip design. Goldie said compute demand could grow at least 10,000x over the next five years—which she called potentially conservative—and argued the response should be more efficient chips rather than 10,000x more data centers. They cited AlphaChip, which they previously developed, as producing an unconventional donut-shaped chip layout that outperformed traditional layouts.
Why it matters
The article gives no product, pricing, availability date, benchmark, or Malaysia/SEA angle, so most builders should not change roadmaps from this. The only concrete signal is the stated 10,000x compute-demand assumption and AI-designed chip layouts as a long-term supply-side lever; if you model AI infrastructure costs, treat it as a thesis to watch, not a procurement input.
Discussion angle
The claim is that AI can discover unconventional chip layouts like a donut shape; what benchmarks, yield data, or fab adoption would prove this is more than a long-horizon bet, and would cheaper compute actually change your next product decision?