AI infrastructure company Infinity has secured $15 million in funding at a $100 million valuation. Investors include Touring Capital, Principal VC, and researchers from leading AI firms like OpenAI and Anthropic. Infinity is developing software to simplify the process of running AI models on various AI chips, aiming to challenge Nvidia's dominance in the market.

Similar to how Nvidia's CUDA software enabled its GPUs to run AI, Infinity is building an alternative kernel software that supports a wide range of chips, including SRAM, GPUs, and phone chips. This universal inference library allows developers to deploy AI applications across different hardware without custom kernel development.

Founded by Jeremy Nixon, formerly of Google Brain, Infinity's mission is "automated invention." The company's AI research agent, Ignition, automatically writes, tests, and optimizes low-level code for AI inference on non-Nvidia chips, significantly reducing development time. Infinity's business model involves taking a percentage of performance gains and cost savings from its customers, such as AI chip maker D-Matrix.