Architect Labs Designed AI Chip in Two Weeks

The startup used an AI system to build its Redwood inference chip and completed a redesign in 48 hours.

Updated on Sept. 28, 2026 in Semiconductors

Bold flat-color editorial illustration of a silicon wafer circuit, representing advanced hardware design. — Architect Labs has unveiled its Redwood inference chip, which the startup designed in under two weeks using an AI-driven hardware generation process. AI Illustration. Upload story photo >

Live Poll

Would you trust critical hardware components designed entirely by AI systems?

In August 2026, San Francisco-based Architect Labs unveiled the Redwood inference chip after completing its initial design process in under two weeks. The startup utilized an AI system to handle hardware description, verification, and firmware generation from a written specification.

Why it matters

By drastically reducing design cycle times, Architect Labs aims to synchronize hardware development with the rapid evolution of AI model architectures. This approach potentially minimizes the time between conceptual design and prototype testing for specialized silicon.

The Redwood chip achieved a 3.4x performance-per-watt gain over the Nvidia Jetson Orin Nano during FPGA-based performance modeling. The design system maintained 95% functional coverage with zero initial bugs in the first RTL drop.

The players

Architect Labs

A San Francisco-based startup with 25 employees that focuses on utilizing AI systems to automate the design of inference chips.

Nvidia

A leading multinational technology company that manufactures the Jetson Orin Nano, the unit used as a performance benchmark for the Redwood chip.

The details

The design workflow allowed the team to adjust specifications and generate new hardware descriptions in just 48 hours. While the company claims throughput gains of 1.75 times and a 1.9 times reduction in power draw compared to the Nvidia Jetson Orin Nano, these metrics are based on FPGA performance rather than physical silicon.

Architect Labs was founded in July 2025.
The company secured $24 million in seed funding in June 2026.
The firm published its paper on the AI-driven design process in August 2026.

The Tech Race

Architect Labs is attempting to disrupt the traditional silicon design cycle by replacing months of manual engineering with AI-driven automation. This shift positions the startup against industry leaders by targeting shorter development timelines to keep pace with rapidly changing AI requirements.

If successful, this development could accelerate the availability of more power-efficient edge AI devices for consumers and businesses. Users may see faster iterations of specialized AI hardware as firms move from design to validation more quickly.

The takeaway

The move toward AI-generated hardware design could fundamentally shift how semiconductor companies respond to emerging model architectures. Future success for the startup will depend on translating these FPGA-verified designs into reliable physical silicon.