The European Fee is investigating Microsoft’s €15 million ($16.3 million) funding in French startup Mistral, which got here simply after the latter launched a big language mannequin to compete with OpenAI’s ChatGPT.
Based in 2023 by researchers from Google DeepMind and Meta, Mistral has raised over €385 million ($417.1 million) from traders comparable to Andreessen Horowitz and Lightspeed. The upstart is valued at round €1.8 billion ($2 billion) at this time.
On Monday, Mistral introduced its newest and strongest massive language mannequin, Mistral Giant, and launched an online app netizens can use to experiment with a chatbot powered by the mannequin. It additionally put out a smaller mannequin, Mistral Small, which – because the identify suggests – is optimized to be quicker and extra compact. Mistral has beforehand shared different fashions, comparable to its seven-billion parameter Mistral-7B in September.
To present builders extra entry to its know-how, Mistral additionally revealed it partnered with Microsoft to convey its open and business fashions to Azure.
In return, Microsoft will assist Mistral by offering it with {hardware} and infrastructure from its cloud computing platform to coach and run its future methods. Mistral co-founder and CEO Arthur Mensch supplied enthusiastic endorsement of the deal in a Microsoft assertion that outlined a shared enthusiasm for “impactful progress within the AI trade and delivering unparalleled worth to our clients and companions globally.”
It isn’t unusual for cloud suppliers to put money into AI startups and for these offers to contain entry to cloudy infrastructure. Microsoft has invested $13 billion into OpenAI, whereas Google and Amazon Internet Providers have invested closely into Anthropic.
However Mistral’s deal has attracted the eye of EU regulators
These offers are a win-win for each events. Internet hosting the most recent massive language fashions drives extra enterprise to cloud platforms, whereas startups want massive clusters of GPUs to develop new fashions. The European Fee, nevertheless, is cautious that AI mergers with Large Tech may consolidate energy and crush innovation, making it probably tougher to manage the know-how and for smaller corporations to compete.
“The fee is wanting into agreements which have been concluded between massive digital market gamers and generative AI builders and suppliers,” European Fee spokesperson Lea Zuber instructed Politico. “On this context, we now have acquired the talked about settlement, which we are going to analyze.”
Final month, EU regulators launched an inquiry into Microsoft’s stake in OpenAI. “We’re inviting companies and specialists to inform us about any competitors points that they could understand in these industries, while additionally carefully monitoring AI partnerships to make sure they don’t unduly distort market dynamics,” Margrethe Vestager, government vice-president accountable for competitors coverage for the European Fee, declared in a press release.
The fee has agreed on the wording of an AI Act – a sweeping plan to implement guidelines for AIs based mostly on the dangers every represents. Generative AI fashions like Mistral Giant or ChatGPT are thought-about as “high-impact general-purpose” methods which may pose “systemic danger”. Builders can be required to reveal the supply of their coaching knowledge and design guardrails to forestall them producing worrisome content material.
The Register has requested Microsoft and Mistral for remark. ®
