TheChek News Flow

Tech giant unveils partnership with French AI start-up, as regulators probe $13bn alliance with ChatGPT maker

Microsoft has struck a deal with French artificial intelligence start-up Mistral as it seeks to broaden its involvement in the fast-growing industry beyond OpenAI.

The agreement, announced by Mistral on Monday, will involve the US tech giant helping to bring its AI models to market.

Microsoft will also take a minor stake in the 10-month-old Paris-based company, although the financial details have not been disclosed.

The partnership makes Mistral the second company to provide commercial language models available on Microsoft’s Azure cloud computing platform.

“Microsoft’s trust in our model is a step forward in our journey to put frontier AI in everyone’s hands,” said Arthur Mensch, cofounder and CEO of Mistral AI.

Microsoft has already invested about $13bn in San Francisco-based OpenAI, an alliance that is being reviewed by competition watchdogs in the US, EU and UK.

Other Big Tech rivals, such as Google and Amazon, are also investing heavily in building generative AI — software that can produce text, images and code in seconds — which analysts believe has the capacity to shake up industries across the world.

Mistral, which builds large language models, the underlying technology that powers generative AI products, secured a €2bn valuation in December in a funding round worth roughly €400mn.

Its models are “open source”, meaning technical details will be released publicly.

This contrasts with the approach of competitors such as ChatGPT maker OpenAI, whose latest model GPT-4 is a so-called black box in which the data and code used to build the model are not available to third parties.

OpenAI was estimated to be worth $86bn during its recent secondary share sale last year.

Paris-based AI startup Mistral AI is gradually building an alternative to OpenAI and Anthropic as its latest announcement shows. The company is launching a new flagship large language model called Mistral Large. When it comes to reasoning capabilities, it is designed to rival other top-tier models, such as GPT-4 and Claude 2.

In addition to Mistral Large, the startup is also launching its own alternative to ChatGPT with a new service called Le Chat. This chat assistant is currently available in beta.

If you’re not familiar with Mistral AI, the company is better known for its capitalization table as it raised an obscene amount of money in very little time to develop foundational AI models. The company was officially incorporated in May 2023. Just a few weeks after that, Mistral AI raised a $112 million seed round. In December, the company closed a $415 million funding round with Andreessen Horowitz (a16z) leading the round.

Founded by alums from Google’s DeepMind and Meta, Mistral AI originally positioned itself as an AI company with an open-source focus. While Mistral AI’s first model was released under an open-source license with access to model weights, that’s not the case for its larger models.

Mistral AI’s business model looks more and more like OpenAI’s business model as the company offers Mistral Large through a paid API with usage-based pricing. It currently costs $8 per million of input tokens and $24 per million of output tokens to query Mistral Large. In artificial language jargon, tokens represent small chunks of words — e.g., the word “TechCrunch” will be split in two tokens “Tech” and “Crunch” when processed by an AI model.

By default, Mistral AI supports context windows of 32k tokens (generally more than 20,000 words in English). Mistral Large supports English, French, Spanish, German and Italian.

As a comparison, GPT-4 with a 32k-token context window currently costs $60 per million of input tokens and $120 per million of output tokens. So Mistral Large is currently 5 to 7.5 times cheaper than GPT-4-32k. Things are changing at a rapid pace and AI companies update their pricing regularly.

But how does Mistral Large stack up against GPT-4 and Claude 2? As always, it’s very hard to tell. Mistral AI claims that it ranks second after GPT-4 based on several benchmarks. But there could be some benchmark cherry-picking and disparities in real-life usage. We’ll have to dig more to see how it performs in our tests.

An alternative to ChatGPT

Mistral AI is also launching a chat assistant today called Le Chat. Anyone can sign up and try it out on chat.mistral.ai. The company says that it is a beta release for now and that there could be “quirks”.

Access to the service is free (for now) and users can choose between three different models — Mistral Small, Mistral Large and a prototype model that has been designed to be brief and concise called Mistral Next. It’s also worth noting that Le Chat can’t access the web when you use it.

The company also plans to launch a paid version of Le Chat for enterprise clients. In addition to central billing, enterprise clients will be able to define moderation mechanisms.

A partnership with Microsoft

Finally, Mistral AI is also using today’s news drop to announce a partnership with Microsoft. In addition to Mistral’s own API platform, Microsoft is going to provide Mistral models to its Azure customers.

It’s another model in Azure’s model catalog, which doesn’t seem that big a deal. And yet, it also means that Mistral AI and Microsoft are now holding talks for collaboration opportunities and potentially more. The first benefit of that partnership is that Mistral AI will likely attract more customers with this new distribution channel.

As for Microsoft, the company is the main investor of OpenAI’s capped profit subsidiary. But it has also welcomed other AI models on its cloud computing platform. For instance, Microsoft and Meta partner to offer Llama large language models on Azure.

This open partnership strategy is a nice way to keep its Azure customers in its product ecosystem. It might also help when it comes to anticompetitive scrutiny.

Chinese auto group’s European head says success is due to BYD’s unique technology and management efficiency

BYD has rejected EU accusations that the success of Chinese car companies stemmed from state aid as Brussels forges ahead with a probe into whether local subsidies were helping electric cars made in China undercut European-made models.

“Our success is not because of the subsidy, it’s because we have unique technology . . . and our management efficiency is high,” said Michael Shu, European president of BYD, one of China’s leading auto groups.

“It’s because we invested in this technology much earlier, and much more, than competitors. It’s not because of the subsidy,” he added in the company’s first public comment on the probe launched last year.

The European Commission’s investigation into Chinese competitiveness in the electric vehicle market is widely expected to lead to higher tariffs. China’s ambassador to Europe previously called the move “unfair”.

The manufacturing subsidies the carmaker receives from China are “very limited”, Shu told the Financial Times in an interview at the Geneva Motor Show.

Carmakers such as Renault and Stellantis have warned that a wave of cheaper Chinese models will undercut those produced by European companies. But Shu said of the established industry: “They are afraid they cannot offer a competitive price.”

BYD, which is backed by Warren Buffett, has launched three models across Europe and is expanding its line-up and sales countries. Shu previously said the company aimed to be one of the top EV sellers in the region by 2030.

He said the company’s cost was “getting lower” as it produced more cars. The business has also launched its own ship to transport vehicles from China to Europe, which further reduces costs, he added.

Although any increase in tariffs may affect the speed of its rollout, “we are making some preparations . . . we will try our best to stabilise the sales price of the products in the market”, he said. Keeping prices stable “is quite important for long term relationship with customers, and brand image”.

Higher tariffs may in fact lead the carmaker to increase the planned capacity of the new EV factory it is building in Hungary, to reduce the number of imported cars, he added.

BYD expects to begin producing cars from the plant before the end of next year, despite having only announced the project in December, an incredibly rapid pace for an industry where factories can take four or more years to complete.

Shu said the Hungary plant planned to produce at least 150,000 cars a year, a minimum level for standard car factories. “We are still discussing how fast we can go and how many cars we can sell”, but it has to be at least 150,000 “otherwise the cost will be too high”.

BYD has not disclosed investment figures yet because it is still in negotiations with the EU over financial support for the project, he said.

The group already runs a battery factory in the country, and chose the site of the factory in Szeged because it is close to the university that the company sees as a potential recruiting ground.

The company has also launched its own car-carrying vessel to ship vehicles from China, an industry first. The move reduces its emissions by using greener fuel and cuts the company’s transportation costs compared with outsourcing its shipments to other companies, Shu said.

“Shipping capacity is an important bottleneck,” he said, adding the new vessel could sail at “BYD’s planning” rather than leaving cars in port waiting for a ship, which was expensive.