Ricursive: AI for Designing Chips in Just Weeks

Key Takeaways

1Ricursive aims to reduce chip design time from 2–3 years to just a few weeks

2The startup has raised $335 million at a valuation of $4 billion, with Nvidia as an investor

3Goldie and Mirhoseini will detail their AI-hardware loop at Disrupt 2026

Why it matters : Accelerating design could, according to the founders, pave the way for new chip architectures and more efficient AI.

Ricursive Intelligence aims to reduce design cycles currently measured in years to just a few weeks. The startup, funded with $335 million at a valuation of $4 billion, promises a self-improving AI-hardware loop that its founders will detail in San Francisco at Disrupt 2026.

Disrupt 2026 gathers over 10,000 participants from October 13 to 15

The session featuring Anna Goldie and Azalia Mirhoseini is part of more than 200 sessions spread across six stages, including panel discussions and workshops, from October 13 to 15 at Moscone West in San Francisco. More than 10,000 founders, investors, operators, and executives are expected, alongside over 250 speakers and more than 300 exhibiting startups. In addition to the agenda, networking, partnership creation, and negotiation opportunities are announced. Today is presented as the last day to save up to $200 on a pass, with purchases before midnight PT and the possibility of obtaining a second pass at 50% off on certain tickets.

Ricursive aims to reduce chip design cycles to a few weeks

Currently, designing a chip can take two to three years. Ricursive wants to bring this timeline down to a few weeks by further automating steps such as component placement and design verification, while learning from each project. AI models are advancing rapidly while hardware follows a different pace; the company believes that AI can help bridge this gap.

A self-learning AI-hardware loop driven by an experienced team

Ricursive is developing an AI that designs chips and learns from the process to improve subsequent iterations. The goal is to automate and accelerate design, with a system designed to capitalize from one chip to the next. This approach is envisioned as a powerful loop: the AI designs better hardware, which in turn enables more efficient AI systems.

Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisissez votre rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

Background and funding: from Google DeepMind to a $4 billion valuation

The founders co-led AlphaChip at Google, a system capable of producing schematics in a matter of hours, and their work contributed to several generations of Google’s TPUs. They also co-founded the ML for Systems team at Google, were early employees at Anthropic, and served as senior research scientists at Google DeepMind. Anna Goldie is the CEO, holds a PhD in computer science from Stanford, and is a recipient of the 35 Innovators Under 35 award; Azalia Mirhoseini is the CTO, an assistant professor at Stanford, and the founder of the Scaling Intelligence Lab. Launched in late 2025, Ricursive raised $335 million in four months for a valuation of $4 billion, including a $300 million Series A, with Nvidia at the table.

What will be discussed on stage: closing the loop between AI and design to alleviate a bottleneck

Goldie and Mirhoseini, who have helped demonstrate that AI can design real chips, will detail how to close the loop between AI and chip development and how to build self-improving systems to address a bottleneck in the industry. According to them, accelerating design could pave the way for new architectures and ultimately lead to more capable and efficient AI. Their talk is presented as an opportunity for executives, founders, and investors to understand the role that the speed of chip development can play in the next stages of AI.