Infinity raises $15M: startup builds a CUDA alternative for AI chips

Infinity has raised $15 million at a $100 million valuation, with backing from Touring Capital, Principal VC, and researchers from OpenAI and Anthropic.

The company is building a universal inference layer designed to run across different chip architectures and reduce dependence on CUDA.

Its agent, Ignition, writes, tests, and refines low-level code for model deployment, then improves results through training.

Already serving customers such as D-Matrix, the company ties monetization to a share of performance gains and cost savings.

If successful, the approach could give AI teams a more flexible way to move models across hardware platforms.