Microsoft has partnered with French AI business Mistral in a multiyear deal to commercialise its new language models.

Mistral’s open and commercial language models will now be accessible on Microsoft’s Azure AI platform. Microsoft’s partnership with Mistral, which has existed for ten months but is valued at around $2.1 billion, intends to prioritise the advancement and implementation of next-generation large language models.

It is Microsoft’s second major AI partnership after OpenAI, which was the first company to offer a commercial language model on Azure and in which Microsoft has invested around $13 billion over the past four years.

Eric Boyd, Corporate Vice President of Azure AI Platform at Microsoft, wrote in a blog post accompanying the news:

Today, we are announcing a multi-year partnership between Microsoft and Mistral AI, a recognized leader in generative artificial intelligence. Both companies are fueled by a steadfast dedication to innovation and practical applications, bridging the gap between pioneering research and real-world solutions.”

Boyd highlighted Mistral’s emphasis on open-source community development, but its Microsoft partnership will see the AI startup explore new commercial opportunities, be granted access to fresh global markets, and “foster ongoing research collaboration”.

“We are thrilled to embark on this partnership with Microsoft,” said Arthur Mensch, CEO at Mistral. “With Azure’s cutting-edge AI infrastructure, we are reaching a new milestone in our expansion, propelling our innovative research and practical applications to new customers everywhere. Together, we are committed to driving impactful progress in the AI industry and delivering unparalleled value to our customers and partners globally.”

Additionally, Mistral announced its Mistral Large AI model this week, with ambitions to seriously compete with OpenAI’s GPT-4 LLM. However, Mistral Large won’t be open source. Mistral Large can be accessed either through Mistral’s own infrastructure or Azure AI Studio and Azure Machine Learning.

The Financial Times also reported that Microsoft will take a minor stake in Mistral.

More Specifics On the Microsoft and Mistral Partnership

The first key area Boyd cites where the partnership will focus is supercomputing infrastructure. Microsoft aims to provide Mistral with Azure AI supercomputing infrastructure, ensuring elite performance and scalability for AI training and inference workloads, particularly for Mistral AI’s flagship models.

The second key area is “Scale to market”. Microsoft and Mistral AI will offer Mistral AI’s premium models through Models as a Service (MaaS) in Azure AI Studio and Azure Machine Learning model catalogue. This broadens the catalogue beyond OpenAI and will now feature myriad open-source and commercial options.

Purchasing Mistral AI’s models using Microsoft Azure Consumption Commitment (MACC) is also now possible. Azure’s AI-optimised infrastructure allows Mistral AI to better promote, sell, and distribute its models to Microsoft’s global customers.

The last key area Boyd mentions is AI research and development. Microsoft and Mistral AI plan to collaborate on developing purpose-specific models for specific customers, including European public sector applications.

Microsoft and OpenAI Under The Microscope

Several antitrust bodies are investigating Microsoft and OpenAI’s relationship, including the US’s Federal Trade Commission (FTC) and the UK’s Competition and Markets Authority (CMA).

These probes arrive two months after November’s dramatic OpenAI saga, in which its CEO Sam Altman was ousted by its board before being reinstated just four days later following intense pressure from both employees and investors.

Over 700 of OpenAI’s 770 workers said they would resign unless Altman was allowed to return, while Microsoft, as OpenAI’s largest investor, exerted pressure on OpenAI’s board.

Microsoft released a statement two days after Altman’s sacking saying that the tech giant had hired both Altman and his OpenAI Cofounder Greg Brockman to manage a new AI research team. In an open letter to their board, the 700-strong group of OpenAI employees said that Microsoft had promised them jobs if they followed through on their threat of resigning.

After Altman’s return, Microsoft CEO Satya Nadella spoke about improving OpenAI’s governance. Microsoft also since secured a nonvoting position on OpenAI’s board.

Following November’s melodrama, the CMA delivered an Invitation to Comment to both tech companies, which comprises the first part of the CMA’s information-gathering initial review and is posted before initiating an official phase one investigation. The regulatory body asked Microsoft, OpenAI, and any other relevant third party whether “recent developments” had seen the partnership expand into a “relevant merger situation”.

Meanwhile, the FTC has announced that Microsoft and OpenAI’s relationship is among several tech giants and AI startup dynamics are being probed, including relationships led by Amazon and Alphabet.



from UC Today https://ift.tt/edPVA3p