AI infrastructure startup Infinity has raised $15 million in seed funding to build a software layer that, in its words, makes any AI chip "inference-ready."
According to TechCrunch, the round values the company at $100 million and was announced on Monday. Investors include Touring Capital and Principal VC, along with individual researchers from companies such as OpenAI and Anthropic.
The company describes itself as Infinity.inc, or the Infinity Artificial Intelligence Institute, an early-stage AI infrastructure research firm. Per a Business Wire release, its goal is to build the software layer that makes any AI chip inference-ready.
Inference is the stage where a trained AI model actually runs to answer questions or generate output — as opposed to the earlier, costlier training phase. Today much of that work is tied to specific hardware. According to Techmeme, citing TechCrunch's Dominic-Madori Davis, Infinity is building an inference library designed to run on all chips.
The startup was founded by Jeremy Nixon, whom Techmeme identifies as the creator of AGI House, a hacker network community. Ukrainian outlet Межа framed the effort as a bid to build a "universal AI inference library."
Why it matters: if software can let AI models run efficiently across many different chips rather than being locked to one vendor, it could loosen the industry's heavy dependence on a small number of hardware suppliers and give buyers more choice.