AI Infrastructure Startup Infinity Completes $15 Million Financing
AI infrastructure startup Infinity announced the completion of $15 million in financing, with the company's valuation reaching $100 million.
This round of financing was led by Touring Capital, Principal VC, and researchers from OpenAI and Anthropic. The company is dedicated to developing general AI inference software that supports various AI chips to run large models, reducing reliance on NVIDIA's CUDA ecosystem. Founded by former Google Brain researcher Jeremy Nixon, Infinity currently has 26 employees and has served clients such as D-Matrix.
Capital continues to flow into startups challenging NVIDIA's software moat, with early VCs and AI lab researchers becoming buyers, putting pressure on those reliant on NVIDIA's CUDA. Infinity benefits from a performance-based revenue-sharing model that capitalizes on performance improvements.
Source: Public Information
ABAB AI Insight
Jeremy Nixon, as a former Google Brain researcher and founder of AGI House, previously focused on automated invention research. The founding of Infinity continues his historical path of optimizing underlying code for AI infrastructure, similar to early open-source framework contributors turning to commercial entrepreneurship.
On the capital front, Infinity's $15 million financing and resources from investors with OpenAI/Anthropic backgrounds focus on the Ignition agent for automatically generating kernel code, concentrating R&D resources on cross-hardware software stack development. The motivation is to capture the infrastructure gap amid the diversification trend of AI chips, using a pay-for-performance model to lower the initial barriers for clients.
Similar to how CUDA initially locked the hardware market through its software ecosystem, or how frameworks like PyTorch promote hardware adaptation, Infinity is currently in the early expansion phase of transitioning AI software from NVIDIA's dominance to a multi-architecture openness.
Structural Judgment Essentially belongs to technological substitution and industrial chain reconstruction: Emerging software stacks like Infinity attempt to replace CUDA's monopoly position in the AI inference layer, reconstructing the hardware-software adaptation chain through automated code generation. This mechanism is driven by the diversification of AI chips and cost pressures, prompting developers to seek more flexible infrastructure options.