{"slug": "this-startup-raised-15m-to-break-nvidias-tightest-grip-on-ai-its-software", "title": "This startup raised $15M to break Nvidia’s tightest grip on AI, its software", "summary": "Infinity, a startup founded by former Google Brain researcher Jeremy Nixon, raised $15 million in seed funding at a $100 million valuation to develop its AI agent Ignition, which automates the writing of low-level chip code to break Nvidia's CUDA software lock-in. The company claims Ignition can achieve 92% of a new chip's peak performance within 10 hours and boost model output from 1,400 to over 20,000 tokens per second in a single day, though these benchmarks are self-reported and not independently verified. Backers include Touring Capital, Principal VC, and researchers from OpenAI and Anthropic.", "body_md": "The pitch is bold for a seed round. Infinity says it can strip away the one advantage that has kept Nvidia untouchable. It has just raised money to try.\n\nThe company closed a $15 million seed round at a $100 million valuation on Monday, [it announced](https://infinity.inc/press/seed-funding). Backers include Touring Capital, Principal VC, executives at major chip firms, and researchers from [OpenAI](https://thenextweb.com/news/gpt-red-openai-ai-hacker) and [Anthropic](https://thenextweb.com/news/anthropic-1-5-billion-copyright-settlement-approved).\n\n## The CUDA problem\n\nTo understand what Infinity is chasing, look at why Nvidia won. Its chips are quick, but its real moat is CUDA, the software layer it has built up over nearly two decades.\n\nThe big AI frameworks, PyTorch and TensorFlow, sit on top of CUDA. Write your app in Python and it runs on Nvidia hardware by default. That convenience is why [Nvidia](https://thenextweb.com/news/nvidia-halves-asia-buyer-list-china-crackdown) holds an estimated 80% of the data-centre AI accelerator market.\n\nRival chips from AMD, Qualcomm, and AWS often match Nvidia on raw power. What they lack is the software. Porting models to new silicon means writing kernels, the low-level code that drives a chip, and few teams can afford the effort.\n\n## An agent that writes the chip code\n\nInfinity wants to automate that work. Its AI agent, Ignition, generates, tests, and rewrites those kernels itself, and keeps tuning them as performance data comes back. Human engineers set the direction. The agent does the grind.\n\nThe results Infinity reports are striking, though they come from the company itself. Working with the chip maker d-Matrix, it says Ignition hit 92% of a new chip’s peak performance 10 hours after first touching the hardware. Within 10 days, three frontier models were running on it end to end.\n\nOn another test, Infinity says it lifted a model’s output from about 1,400 to more than 20,000 tokens a second in a single day. That would beat vLLM, a widely used open-source inference framework. The claims are not yet independently checked.\n\n## Automating invention\n\nThe founder gives the project its flavour. Jeremy Nixon is a former Google Brain researcher who created AGI House, a San Francisco hacker network that says it has spawned hundreds of startups.\n\nNixon told [TechCrunch](https://techcrunch.com/2026/07/20/inference-startup-infinity-raises-15m-from-touring-capital-openai-and-athropic-researchers/) he is obsessed with “automated invention,” the idea that AI can be a kind of meta technology. He once built an algorithm, Omega, that invented other machine-learning algorithms and scored them in a loop. Ignition applies the same instinct to hardware.\n\nHe announced the raise on [X](https://x.com/JvNixon/status/2079228475760865423) with a flourish, hailing the arrival of AI systems that “enable, optimize and invent” the next generation of AI systems.\n\n## A crowded race, and the caveats\n\nInfinity joins a wave of startups trying to [break the CUDA lock-in](https://thenextweb.com/news/alibaba-t-head-sail-open-source-nvidia-cuda-alternative). It says it already earns millions in annual recurring revenue and employs 26 people. Its business model takes a cut of the speed and cost gains it delivers, rather than a licence fee.\n\nThe caveats are real. This is a seed-stage firm with one public chip partner, and the headline benchmarks are self-reported. The prize, though, is large. As [inference](https://thenextweb.com/news/moonshot-kimi-k3-subscriptions-paused-gpu-capacity) races to become two-thirds of all AI compute spending this year, cheaper ways to run it are worth a fortune.\n\n“The next era of AI will be defined not just by who makes the best chip, but by who can make any chip run state-of-the-art models at blazing speeds,” Nixon said. If Ignition works as billed, the moat that made Nvidia untouchable starts to look shallower.\n\n## Get the TNW newsletter\n\nGet the most important tech news in your inbox each week.", "url": "https://wpnews.pro/news/this-startup-raised-15m-to-break-nvidias-tightest-grip-on-ai-its-software", "canonical_source": "https://thenextweb.com/news/infinity-seed-funding-any-chip-ai-inference", "published_at": "2026-07-21 13:31:34+00:00", "updated_at": "2026-07-21 13:49:34.665700+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-startups", "ai-tools", "ai-agents", "ai-chips"], "entities": ["Infinity", "Nvidia", "Jeremy Nixon", "Ignition", "Touring Capital", "Principal VC", "OpenAI", "Anthropic"], "alternates": {"html": "https://wpnews.pro/news/this-startup-raised-15m-to-break-nvidias-tightest-grip-on-ai-its-software", "markdown": "https://wpnews.pro/news/this-startup-raised-15m-to-break-nvidias-tightest-grip-on-ai-its-software.md", "text": "https://wpnews.pro/news/this-startup-raised-15m-to-break-nvidias-tightest-grip-on-ai-its-software.txt", "jsonld": "https://wpnews.pro/news/this-startup-raised-15m-to-break-nvidias-tightest-grip-on-ai-its-software.jsonld"}}