Bill Gates leads $110M Series A for Neurophos optical AI chip to challenge Nvidia
Neurophos raised $110M led by Gates Frontier to build optical AI chips. Their OPU claims to be 50x more efficient than Nvidia's B200, arriving in 2028.
Bill Gates is placing a massive bet on the future of AI hardware. As labs struggle with skyrocketing power bills, an Austin-based startup called Neurophos claims it's found the solution—not in better transistors, but in the physics of light.
Neurophos optical AI chip Series A funding and technology
According to TechCrunch, Neurophos has raised $110 million in a Series A round led by Gates Frontier. The investor syndicate includes heavy hitters like Microsoft’s M12, Aramco Ventures, and Bosch Ventures. The capital will fuel the development of their Optical Processing Unit (OPU), which aims to redefine AI inferencing.
At the heart of their innovation is a "metasurface modulator." While optical computing has existed for years, components were often too bulky for mass production. Neurophos claims its metasurface is 10,000 times smaller than traditional optical transistors, allowing them to pack thousands of modulators onto a single chip using standard foundry tools.
Outperforming the Nvidia Blackwell Architecture
The performance claims are staggering. Neurophos' OPU is designed to run at 56 GHz, delivering a peak of 235 Peta Operations per Second (POPS) while consuming only 675 watts. In contrast, Nvidia’s flagship B200 (Blackwell) provides roughly 9 POPS at 1,000 watts. This represents a massive leap in energy efficiency and raw computational power.
CEO Patrick Bowen emphasizes that they aren't just following TSMC’s node roadmaps, which typically yield 15% efficiency gains every few years. Instead, they are starting with a 50x advantage over current silicon. The company expects its first chips to hit the market by mid-2028.
Authors
Related Articles
China's LightGen optical chip claims 100x+ gains over Nvidia. Is this a lab curiosity or China's strategic answer to crippling US tech sanctions?
Samsung Electronics' union branch revealed that 84% of survey respondents oppose the government's roughly $290 billion (₩400 trillion) semiconductor megaproject in Gwangju. Government, management, the union and shareholders are all reading the same national project in sharply different ways.
Nvidia shipped roughly a billion RISC-V cores in 2024, then announced it would run CUDA on the open standard. We break down how royalty-free instruction sets and open software stacks are trying to route around CUDA's lock-in. Part 2 of the Semiconductor Sovereignty series.
US AI-chip export controls split into three layers in the first half of 2026 — January easing, a May crackdown on circumvention, and a pending bill. Nvidia erased China from its guidance and still posted a record $81.6 billion quarter. A look at the export policy that both shields and cages it.
China's LightGen optical chip claims 100x+ gains over Nvidia. Is this a lab curiosity or China's strategic answer to crippling US tech sanctions?
Samsung Electronics' union branch revealed that 84% of survey respondents oppose the government's roughly $290 billion (₩400 trillion) semiconductor megaproject in Gwangju. Government, management, the union and shareholders are all reading the same national project in sharply different ways.
Nvidia shipped roughly a billion RISC-V cores in 2024, then announced it would run CUDA on the open standard. We break down how royalty-free instruction sets and open software stacks are trying to route around CUDA's lock-in. Part 2 of the Semiconductor Sovereignty series.
US AI-chip export controls split into three layers in the first half of 2026 — January easing, a May crackdown on circumvention, and a pending bill. Nvidia erased China from its guidance and still posted a record $81.6 billion quarter. A look at the export policy that both shields and cages it.
Thoughts
Share your thoughts on this article
Sign in to join the conversation