MatX lands $500M-plus to build LLM-specific chips
MatX, a startup founded by former Google chip engineers, has raised more than $500 million to develop silicon designed to compete with Nvidia in the fast-growing market for AI compute, according to Bloomberg.
The round was led by Jane Street and Situational Awareness, an investment firm started by former OpenAI researcher Leopold Aschenbrenner. Additional backers include Marvell Technology, venture firms NFDG and Spark Capital, and Stripe co-founders Patrick Collison and John Collison.
Manufacturing capacity and scarce components
Mike Gunter said the financing is intended to help the company secure manufacturing space and critical parts—especially memory, which remains in tight supply across the semiconductor industry. “It lets us compete on kind of equal grounds with the largest companies in the way that they can scale very quickly,” Gunter said, adding the round brings the company closer to the resource levels of major incumbents.
Founded by Google alumni, targeting AI labs
MatX was started by Reiner Pope and Mike Gunter, who left Google in 2022 to design a chip optimized for large language models (LLMs), the technology behind today’s AI chatbots. The company plans to finalize its chip design this year and begin shipping in 2027, with manufacturing expected to run through TSMC. The startup currently employs about 100 people and is hiring primarily for engineering roles.
Rather than building a broad sales organization, MatX plans to sell to a small set of leading AI labs as developers increasingly diversify chip suppliers and cloud providers. The company claims internal tests show its proposed chip can outperform Nvidia’s forthcoming Rubin Ultra on compute performance per square millimeter, a key efficiency metric.
Pope noted that competing with Nvidia requires matching incumbents across performance, reliability and software compatibility while staying ahead on at least one major dimension—an approach he says many chip startups have struggled to execute.
Related
