Breaking

The Information Sets Ricursive Founders Interview on AI-Designed Chips and Recursive Self-Improvement

September 28, 2026 at 14:33 EDT

Infra & Chips
AI Agents
Foundation Models

The Information has scheduled Episode 3 of AI Deep Dive for September 28, 2026, at 12 p.m. PT / 3 p.m. ET, featuring Ricursive Intelligence co-founders Anna Goldie and Azalia Mirhoseini on recursive self-improvement - the idea that AI can help design better chips, which then help build stronger AI systems. The appearance puts a fast-rising AI hardware startup at the center of a debate that has moved from theory into production chip design.

Sep. 28, 2026 · Ricursive Intelligence

AI designs chips. Chips train AI.

Ricursive Intelligence is advancing a recursive self-improvement thesis: automate chip design, build better silicon, then use that hardware to train stronger AI systems.

$335M

total backing listed by the company

$4B

post-money valuation after Series A

<6h

reported AlphaChip placement time

rollout phases described publicly

Funding ramp, drawn to scale

Each block represents about $35 million. The Series A column shows how quickly the financing expanded.

$35M

seed round

$300M

Series A

The recursive loop

AI model proposes chip designs

→

Custom silicon improves workloads

→

Better hardware trains stronger AI

Why builders care

AlphaChip reached production hardware, and workload-specific silicon could help high-scale inference, scientific discovery and power-sensitive applications.

What remains open

Ricursive has not released a generally available product, and real commercial chip data remains more representative than open-source examples.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.