Hot French startup ZML releases free product to speed inference across lots of AI chips

French AI startup ZML released a free inference server, ZML/LLMD, that enables open-source large language models to run at peak performance across multiple chip types including Nvidia, AMD, Google TPU, Apple Metal, and Intel Arc. The company aims to break vendor lock-in and reduce AI costs by allowing enterprises to mix chips, potentially disrupting Nvidia's dominance in the inference market.

The days of Nvidia’s unparalleled market dominance aren’t over, but challengers and choices are arising from all directions. ZML https://zml.ai/ , a hot French AI startup endorsed https://x.com/ylecun/status/1836030233796874244 by Turing Award winner Yann LeCun, has released inference https://techcrunch.com/2026/07/03/artificial-intelligence-definition-glossary-hallucinations-guide-to-common-ai-terms/ inference -performance software that allows a variety of open-source large language models https://x.com/steeve/status/2073166739668345255 to run on a variety of chips https://x.com/steeve/status/2058815921997680841 — including Nvidia’s, AMD’s, Google’s TPU, Apple Metal and Intel Arc. With ZML/LLMD https://zml.ai/llmd/ , the newly launched LLM inference server, the company’s ambition is to break existing silos and make different chips available for AI use cases at their maximum available speed, and sometimes faster, ZML founder Steeve Morin told TechCrunch. As AI becomes integrated into our work and everyday lives, optimizing inference — aka, the processing of prompts — has been outpacing model raining https://techcrunch.com/2026/07/03/artificial-intelligence-definition-glossary-hallucinations-guide-to-common-ai-terms/ training in importance, but often feels patchy behind the scenes, with software and architecture barriers that lead to vendor lock-in, Morin said. The promise of achieving peak performance across a variety of chips is a technological feat, but it could also be a market disruptor, amid mounting fears over AI-related costs. ZML hopes to provide enterprises and clouds with the option to use a mix of chips, some of which might be less costly or consume less energy. “The idea is to give people back the power to create their own system and achieve real efficiency gains that allow AI to be disseminated,” Morin said. Such a software assist may help novel AI chipmakers, many of which happen to be from Europe, Morin observed, citing Axelera https://axelera.ai/ , Fractile https://www.fractile.ai/ , Kalray https://www.kalrayinc.com/ , OLIX https://olix.com/ , Q.ANT https://qant.com/ , SiPearl https://sipearl.com/ , SpiNNcloud https://spinncloud.com/ , and VSORA https://vsora.com/ . But more than their region of origin, what matters to him is that ZML can work with them on “things that haven’t been done before anywhere in the world.” That doesn’t mean Morin is bearish on Nvidia. He’s not https://www.thetwentyminutevc.com/steeve-morin , in part because of its existing supply. He told TechCrunch that ZML has a good relationship with the AI chip giant, which has been gearing up https://techcrunch.com/2024/11/20/nvidias-ceo-defends-his-moat-as-ai-labs-change-how-they-improve-their-ai-models/ for the rise of inference. Inference has been an area of such intense investment, that the trend has been hailed the “ inference gold rush https://thenextweb.com/news/baseten-1-5bn-round-13bn-valuation-ai-inference .” So ZML has competition such as Baseten https://techcrunch.com/2026/06/18/ai-inference-startup-baseten-reportedly-raising-1-5b-months-after-its-last-mega-round/ , recently valued at $13 billion; Inferact https://techcrunch.com/2026/01/22/inference-startup-inferact-lands-150m-to-commercialize-vllm/ , from the creators of open source project vLLM https://vllm.ai/ ; as well as RadixArk https://www.businesswire.com/news/home/20260505077157/en/RadixArk-Launches-with-%24100-Million-in-Seed-Funding-Led-by-Accel-to-Grow-SGLang-and-Democratize-Frontier-AI-Infrastructure , the commercial company behind SGLang https://techcrunch.com/2026/01/21/sources-project-sglang-spins-out-as-radixark-with-400m-valuation-as-inference-market-explodes/ . Both vLLM and SGLang partially compete with LLMD, but Morin’s ambitions for ZML cover a broader spectrum. “We have reached the point where we are co-designing silicon,” he said. He further credited ZML’s lean team of 20 people as the reason why the Paris-based startup has been able to move fast, with more releases in the plans. It also helped that this small team is well funded for its size. Thanks to his track record as VP of engineering of Zenly https://techcrunch.com/2017/06/21/snapchat-buys-zenly/ , which Snapchat acquired for nine figures in 2017 https://techcrunch.com/2017/06/21/snapchat-buys-zenly/ , Morin raised $20 million from venture firms including Harry Stebbings’ 20VC, commit, AALVC, Drysdale Ventures, Xavier Niel’s Kima Ventures, Kindred Capital, LocalGlobe, and Puzzle Ventures. Unlike ZML’s first public project, the inference-focused ML framework http://github.com/zml/zml released in 2024 and updated in March https://zml.ai/posts/zml-v2/ , ZML/LLMD is not open source. But it is launching as a free product with the goal of learning about usage. “I’d rather measure and then generate revenue where it is most effective without hindering my growth stupidly because I have been too greedy from the get-go,” Morin said. It is too early to tell when ZML/LLMD might become a paid product, and what its adoption will look like. But the startup’s cap table confirms that other founders are paying attention, including Dagger and Docker founder Solomon Hykes, Clément Delangue and Julien Chaumond from Hugging Face, as well LeCun, now with AMI Labs https://techcrunch.com/2026/03/09/yann-lecuns-ami-labs-raises-1-03-billion-to-build-world-models/ . This also builds the case that Europe’s AI startups can now build from home https://techcrunch.com/2026/07/06/station-f-ramps-up-as-a-launchpad-for-europes-hottest-ai-startups/ . “I couldn’t do ZML anywhere but in Paris,” Morin said.