Infinity, an AI infrastructure startup, has raised $15 million in funding at a $100 million valuation. The round was led by Touring Capital and included researchers from OpenAI and Anthropic, aiming to challenge Nvidia's dominance in AI chip software.
TL;DR
- Infinity raises $15M to build CUDA-alternative software for AI chips, backed by notable investors and researchers.
- The startup aims to reduce reliance on Nvidia's CUDA by developing a universal inference library for various chip architectures.
- Infinity's AI research agent, Ignition, automates the creation and optimization of low-level code for AI inference, potentially speeding up development processes.
What happened
Infinity, an AI infrastructure company, announced a $15 million funding round on Monday, achieving a $100 million valuation. The round was led by Touring Capital and included Principal VC, as well as researchers from OpenAI and Anthropic. The startup is focused on developing software that makes it easier for AI chips to run AI models, aiming to challenge Nvidia's dominance in the market.
Infinity is building a CUDA-alternative kernel software that works with various chip types, including SRAM, GPUs, phone chips, and Systolic Arrays. The company's goal is to create a universal inference library that can run on all chips, automating the replication of state-of-the-art research results. This approach is part of a broader effort by startups to chip away at Nvidia's market dominance.
Infinity was launched last year by Jeremy Nixon, a former researcher at Google Brain and creator of the hacker network community AGI House. Nixon's interest in 'automated invention' led him to develop Infinity's AI research agent, Ignition. This agent writes, tests, debugs, and measures low-level code for AI inference on Nvidia-alternative chips, continuously learning and improving itself. The system adapts to different chip architectures and claims to offer a CUDA-level software stack.
Why it matters
Infinity's success could democratize AI chip development, reducing reliance on Nvidia's CUDA and enabling developers to use a wider range of chip architectures. This could lead to increased innovation and competition in the AI hardware market, potentially driving down costs and improving performance. For developers and startups, Infinity's software could simplify the process of optimizing AI models for different hardware, speeding up development and deployment.
For investors, Infinity's approach represents a significant opportunity to capitalize on the growing demand for AI infrastructure. The startup's focus on automated invention and its ability to attract high-profile investors and researchers signal strong potential for growth. However, the competitive landscape is challenging, with established players like Nvidia and other startups vying for market share. Infinity's success will depend on its ability to deliver on its promises and gain traction among AI developers and chip manufacturers.
Key facts
- Infinity raised $15 million in funding at a $100 million valuation.
- The funding round was led by Touring Capital and included Principal VC, as well as researchers from OpenAI and Anthropic.
- Infinity aims to build a CUDA-alternative kernel software that works with various chip types.
- The startup's AI research agent, Ignition, automates the creation and optimization of low-level code for AI inference.
- Infinity was launched by Jeremy Nixon, a former researcher at Google Brain and creator of AGI House.
- The startup has 26 employees, including those in design, operations, and engineering.
- Infinity's customers include AI chip maker D-Matrix, and the startup is in talks with other big chip and cloud companies.
- Infinity charges a cut of performance gains and cost savings, rather than an upfront license fee.
Context
Infinity is part of a new wave of startups attempting to challenge Nvidia's dominance in the AI chip market. Nvidia's CUDA software has been a key factor in the company's success, enabling developers to write apps in popular languages like Python and use major AI frameworks. However, this reliance on CUDA has also created a barrier to entry for other chip manufacturers and limited the flexibility of AI developers.
Infinity's approach to automated invention and its focus on building a universal inference library could help level the playing field. By enabling AI models to run on a wider range of chip architectures, Infinity's software could drive innovation and competition in the AI hardware market. This could lead to improved performance, lower costs, and greater accessibility for developers and startups.
The success of Infinity and other startups in this space will depend on their ability to deliver on their promises and gain traction among AI developers and chip manufacturers. As the demand for AI infrastructure continues to grow, the competitive landscape is likely to become even more intense, with established players and new entrants vying for market share.
