Key facts
- Infinity raised $15 million at a $100 million valuation.
- Investors include Touring Capital, Principal VC, and researchers from OpenAI and Anthropic.
- The company is developing software to enable AI models to run on various chips.
- Infinity's goal is to create a universal inference library that challenges Nvidia's CUDA platform.
- The startup's AI agent, Ignition, automates code generation and optimization for AI hardware.
AI infrastructure company Infinity has secured $15 million in funding at a $100 million valuation, with participation from investors including Touring Capital, Principal VC, and researchers from OpenAI and Anthropic. The startup is focused on developing software that allows AI models to run efficiently on a variety of AI chips, aiming to provide an alternative to Nvidia's dominant CUDA platform.
Infinity's core offering is a universal inference library designed to automate the process of running state-of-the-art AI research results across different hardware architectures. Their AI research agent, Ignition, writes, tests, debugs, and optimizes low-level code for AI inference, adapting to various chip designs and continuously improving performance. This approach seeks to reduce the significant resources and expertise typically required to port AI applications to different hardware.
The company was founded by Jeremy Nixon, a former researcher at Google Brain and creator of the AGI House community. Nixon's motivation stems from a belief in "automated invention," envisioning AI systems as a meta-technology capable of generating not only algorithms but also the necessary low-level code for hardware optimization.
Infinity's current customers include AI chip maker D-Matrix, and the company is in discussions with other major chip and cloud providers. The startup operates on a performance-based revenue model, taking a share of cost savings and performance gains rather than upfront license fees. Infinity currently employs 26 people across design, operations, and engineering.