Infinity Aims to Disrupt Nvidia's AI Chip Monopoly

Infinity, an AI infrastructure startup, has secured a $15 million raise at a $100 million valuation, backed by investors such as Touring Capital and researchers from OpenAI and Anthropic. This significant funding will enable Infinity to further develop its software, which aims to challenge Nvidia's dominance in the AI chip market.

Building a CUDA-Alternative Kernel Software

The startup is working on a CUDA-level software stack that can operate on any type of chip, including SRAM, GPUs, phone chips, and Systolic Arrays. This is a crucial aspect of AI development, as most app-level startups lack the resources or expertise to write their own kernels and port their apps to other AI chips.

Infinity's software will enable developers to write their apps in popular languages like Python and use major AI frameworks such as PyTorch and TensorFlow. The startup's goal is to create a universal inference library that can run on all chips, allowing them to automate replicating state-of-the-art research results.

The Vision of 'Automated Invention'

Infinity was founded by Jeremy Nixon, a former researcher at Google Brain and creator of the AGI House hacker network community. Nixon's vision is to create 'automated invention' – the idea that AI systems can generate low-level code, like kernels and so forth, needed to run chips more effectively.

Ignition: The Self-Optimizing AI Research Agent

Infinity's AI research agent, Ignition, is designed to write the low-level code needed for AI inference on Nvidia-alternative chips. It tests, debugs, and measures how fast the hardware performs with the code, and automatically rewrites the code if needed to improve performance. The system is self-optimizing, meaning it continuously learns and improves itself.

Collaborations and Future Plans

Infinity is already working with AI chip maker D-Matrix and is in talks with other big chip and cloud companies. The startup doesn't charge an upfront license fee; instead, it takes a cut of performance gains and cost savings, measuring changes in tokens per second. Infinity currently has 26 employees, including those in design, operations, and engineering.