A company called Infinity has raised $15 million to build software that automatically writes the low-level code that AI chips need to function. The round values the company at $100 million and includes investment from Touring Capital, Principal VC, and individual researchers from OpenAI and Anthropic. TechCrunch

Founded in 2025 by Jeremy Nixon, a former Google Brain researcher and creator of the AGI House hacker network community, Infinity operates under the domain infinity.inc. Nixon told TechCrunch the company grew out of his interest in "automated invention" — the idea that AI systems can build and improve other technologies on their own. TechCrunch

Nixon previously built a machine learning algorithm called Omega, which created new ML algorithms and tested them automatically in a loop. Infinity is now applying that same idea to hardware. The company is building a software library designed to run on all chips, enabling automated replication of the latest research results. TechCrunch

Central to Infinity's approach is Ignition, an AI tool that writes the low-level code needed to run AI on chips that compete with Nvidia's. Ignition tests, debugs, measures performance, and rewrites code on its own to improve it. Infinity claims Ignition can match the capabilities of CUDA — Nvidia's proprietary software that developers currently rely on to program Nvidia chips. Think of CUDA as the language Nvidia chips speak. Infinity wants to build a similar language for every other chip maker. TechCrunch

AI chip maker D-Matrix, a would-be Nvidia challenger, is an Infinity customer. In a case study involving D-Matrix's Corsair platform, Ignition reduced a process that could have taken months or years down to hours or days. TechCrunch

Infinity does not charge upfront licensing fees. Instead, the company takes a percentage of the performance gains and cost savings it delivers, measured in tokens per second — essentially, how fast an AI model produces output. TechCrunch

Nixon said Infinity is in talks with other major chip and cloud companies, though he did not name them. As of July 2026, the company employs 26 people across design, operations, and engineering. The TechCrunch report is based on an interview with Nixon and does not cite an external press release or a posting on Infinity's own domain. TechCrunch

The broader context here is that Nvidia dominates the AI chip market not just because its hardware is powerful, but because its software ecosystem, CUDA, has been built up over years. Developers have created countless tools and optimizations specifically for Nvidia chips, which makes switching to a different chip maker expensive and slow. If Infinity's automated code generation can match what human engineers produce by hand for Nvidia, the cost of bringing a new AI chip to market drops significantly. TechCrunch

Worth flagging, the claims should be treated with caution. The D-Matrix case study is a single example, and the details come from Infinity's own documentation rather than an independent test. Automatically generating code for brand-new chip designs is a much harder problem than optimizing code for well-known hardware. The claim of matching CUDA's capabilities is ambitious and has not been verified by outside parties. TechCrunch

If Ignition works as described, it lowers the barrier for new chip makers. Chips from smaller companies could reach production-ready performance faster, relying on automated optimization instead of large engineering teams. That would not displace Nvidia overnight, but it could shorten the timeline for viable alternatives. The long-term implication is that Nvidia's software advantage narrows, and hardware performance and pricing become what buyers care about most. TechCrunch

For now, Infinity is a 26-person company with a working case study and a funding round backed by investors who understand this space. Whether Ignition works across many different chip designs is the open question, but the approach targets a real and expensive bottleneck in AI hardware development. TechCrunch