AI infrastructure startup Infinity has successfully completed a $15 million funding round, elevating its valuation to $100 million. The funding was led by Touring Capital and Principal VC, with participation from researchers at OpenAI and Anthropic. Infinity aims to develop a CUDA alternative through its universal inference software library, which can operate across various hardware architectures, reducing reliance on NVIDIA's ecosystem. Infinity's flagship product, Ignition, is designed to automatically generate low-level Kernel code for non-NVIDIA chips, featuring capabilities such as self-testing, debugging, and continuous optimization. This innovation significantly shortens the development cycle from months to mere hours or days. Founded by former Google Brain researcher Jeremy Nixon, Infinity employs 26 staff members and has partnered with AI chip manufacturer D-Matrix. The company is also in discussions with several major chip and cloud computing firms, operating on a performance revenue-sharing model without upfront licensing fees.