BlockBeats News, July 20th, AI infrastructure startup Infinity announced the completion of a $15 million financing round, with the company's valuation reaching $1 billion. This round of financing was led by Touring Capital, Principal VC, and researchers from OpenAI and Anthropic.

Infinity is dedicated to developing universal AI inference software with the aim of helping various AI chips run large models, reducing developers' reliance on the NVIDIA CUDA ecosystem.

Currently, one of the key reasons NVIDIA dominates the AI chip market is its CUDA software platform. CUDA enables GPUs, originally designed for graphics processing, to perform general computing tasks. It has become the foundation of mainstream AI development frameworks such as PyTorch and TensorFlow, allowing developers to run on NVIDIA hardware without additional adaptation.

Infinity aims to build an alternative software stack similar to CUDA, supporting various hardware architectures including GPUs, SRAM, mobile chips, and Systolic Array, enabling AI applications to run across different chips.

Founded by former Google Brain researcher and AGI House founder Jeremy Nixon last year, Infinity is driven by Nixon's research interest in "automated invention." The company seeks to leverage AI systems to automatically generate the low-level code required to optimize chip operations.

The company's AI research agent, Ignition, can automatically write the low-level kernel code needed for AI inference, continuously optimize the code through testing, debugging, and performance evaluation to adapt to different chip architectures. Infinity claims that the system can continuously learn and automatically improve performance, aiming to build a software infrastructure on par with CUDA.

Currently, Infinity's clients include AI chip company D-Matrix, and the company is in discussions with other major chip manufacturers and cloud computing companies for collaboration. The company operates on a performance-based pricing model, taking a share of the performance improvement and cost savings instead of upfront licensing fees.

Infinity currently has 26 employees and is becoming one of the emerging startups challenging NVIDIA's AI software ecosystem.