Architect Labs raises $24 million seed round to build AI tools for designing custom chips at scale

Architect Labs, a startup building artificial intelligence tools for custom chip design, said it has raised $24 million in seed funding as it looks to accelerate how companies create and modify application-specific integrated circuits and other custom silicon. The company’s focus is on making chip design faster and easier, aiming to reduce time-consuming steps that normally require specialized engineering teams and extended iteration cycles.

The announcement frames the startup’s effort as competitive with businesses currently serving the custom chip market at large semiconductor suppliers. In particular, the company said it intends to take on the custom-chip work associated with Broadcom and Marvell, two major players that supply chips for networking, infrastructure, and other enterprise and communications applications.

Architect Labs’ product concept centers on using AI to assist chip design. While the post describes the goal as speeding and easing the process, it does not spell out the specific workflows the technology will target, such as whether it will automate parts of hardware description generation, verification, physical design, or design-space exploration.

The company did not provide, in the information shared, any customer names, pilot programs, or production timelines that would indicate when its tools might be used commercially. It also did not disclose whether the platform is intended to be sold as software, offered as a service alongside design engineering, or delivered through partnerships with chipmakers or cloud infrastructure providers.

Broadcom and Marvell both operate large semiconductor portfolios, including components used in data centers and communications networks. Custom chips can be attractive to customers that need tailored performance, power, or integration characteristics that off-the-shelf components cannot match. That demand creates opportunity for specialized tooling and services that shorten the cycle from design intent to manufacturable silicon.

From a sector perspective, custom chip ecosystems are increasingly influenced by software. Even when the chip fabrication is outsourced to foundries, design, verification, and optimization are expensive and slow without automation. AI-based design assistance is therefore a natural pitch, because it targets the bottlenecks that keep custom silicon programs from moving faster.

What is still unclear is the scope of Architect Labs’ technology and how it fits into real customer design pipelines. The announcement, as published, does not provide technical benchmarks, details on model training data, integration requirements, pricing, or evidence of performance improvements such as reduced design time, lower power, or higher yield.

For the near term, the main items to watch are whether Architect Labs names any early customers or partners, publishes measurable results tied to specific design steps, and clarifies how it plans to compete in a market that is often governed by customer relationships, engineering trust, and long product timelines.