MatX, founded by two former Google chip engineers, has raised more than $500 million in new funding to develop hardware designed to compete directly with NVIDIA.
According to Bloomberg, the financing round was led by Jane Street and Situational Awareness, the investment firm started by former OpenAI researcher Leopold Aschenbrenner. Other investors include Marvell Technology, venture firms NFDG and Spark Capital, and Stripe co-founders Patrick and John Collison.
The new funding will allow the startup to secure manufacturing space and obtain important parts, especially memory, which is currently in short supply in 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. “This round puts us almost on the same footing as the players who have a huge amount of money.”
Founded by Google Chip Veterans
MatX was started by Reiner Pope and Mike Gunter, both alumni of Google’s semiconductor unit. The two left Google in 2022 with a focused ambition: design a chip built specifically for large language models, the technology behind today’s AI chatbots.
MatX is part of a growing wave of startups trying to capture a slice of the AI chip market, which has long been dominated by Nvidia’s graphics processing units. It designs hardware specifically for top AI models, focusing on maximising performance for large-scale applications.
Their technology provides 10 times the computing power, enabling AI labs to significantly enhance models. This hardware can train advanced models like GPT-4 and operate applications like ChatGPT, making them accessible even to small startups.
MatX aims to make intelligence accessible to everyone, regardless of background, by offering specialised AI services such as medical advice, tutoring, coaching, and assistance.
What’s next?
The company plans to finalise the chip design this year and begin shipping in 2027. MatX intends to manufacture its chips through Taiwan Semiconductor Manufacturing Co.
At present, the company employs around 100 people and is hiring aggressively for engineering roles. The company is less focused on building a large sales organisation, as its strategy centres on selling to a small group of leading AI labs.
Developers such as OpenAI and Anthropic are increasingly relying on multiple chip suppliers and cloud providers, potentially opening the door for new entrants.
MatX claims internal testing shows its proposed chip can outperform Nvidia’s upcoming Rubin Ultra product on computing performance per square millimetre, a key efficiency metric.
As Bloomberg reported, competing with Nvidia requires more than building a fast chip. Companies must anticipate how AI models evolve, match incumbents across performance, reliability and software compatibility, and scale manufacturing in a market where critical components are scarce.
“You need to match what is in the market on all of maybe five different important aspects, and you need to be far ahead on at least one of them,” Pope said. He added that the typical startup approach of focusing on one advantage while neglecting other areas has not worked well in this sector.
