Guess what? A company just one month old, without a product, with employees still getting hired, has already received a valuation of around $260 million, and raised a seed funding of $113 million. Mistral AI has received probably one of the biggest seed fundings in the history of AI ever, more than OpenAI as well. 

For instance, other AI startups like Anthropic and Cohere raised $124 million and $40 million in Series A, respectively, and Stability AI raised $101 million in seed round.

Mistral AI has not yet revealed the utilisation of its funding so far. However, the founders’ approach towards open source makes it clear that they would want to build models through publicly available data to avoid legal issues, as other companies have been facing a lot of backlash for training on users’ data. 

Currently the company is backed by Lightspeed Ventures, alongside Xavier Niel, JCDecaux Holding, Rodolphe Saadé, and Motier Ventures in France. La Famiglia and Headline from Germany; Exor Ventures from Italy; Sofina from Belgium; and First Minute Capital and LocalGlobe also participated in the seed round. 

Paul Murphy, another investor from Lightspeed Ventures Partners, said that one of the biggest reasons to invest in the company was the most impressive founding team. The company is also backed by former Google CEO Eric Schmidt.

“Then we thought, who is the most credible here? Then we met with the Mistral founders. We think there are only 70-100 people in the world right now with their expertise for language models and optimising them,” said the Lightspeed partners. 

The team from Europe

Based in Paris, Europe, Mistral AI has blown the AI startup race funds of bounds. Founded by Arthus Mensch from DeepMind, Guillaume Lample and Timothée Lacroix, from Meta AI. Lample is one of the core members of the team behind Meta AI’s LLaMa model, which has been leading the way in open-source.

When Lightspeed Ventures approached them, the founders were studying at École Polytechnique and École Normale Supérieure, and the trio have known each other since childhood.

Interestingly, this is also the first time a startup outside Silicon Valley is gaining all the spotlight. This means that apart from the giants like Meta, Apple, Google, and Amazon, the baton for AI is slowly being taken away from the big-tech and now is with developers. 

But, one question haunts, will this European startup survive the harsh regulatory hurdles like the GDPR and EU AI Act? Conversely, OpenAI’s success in the field links us to it being based out of the US, where laws are less stringent. These laws are also one of the reasons why OpenAI’s France (Europe) user base is the smallest – i.e. 3.98% of the users globally. In other words, for Mistral AI to make an impact and surpass OpenAI, it might have to consider moving out of Europe. 

Open Source Mantra 

According to TechCrunch, the founders are focused on building open source solutions for enterprises, and not building something consumer centric. The chief Mensch believes that making AI useful is one of the biggest challenges in the current generative AI field, and thus wants to release their products to the market in 2024. The research focused approach of the founders is visible even in their vision. 

The problem was that the company believes that OpenAI are not adhering to the word “open” in their name. Moreover, OpenAI’s proprietary approach has made large companies building models for themselves to be the norm. While Mensch and team believe that, “open source is a core part of our DNA.” 

Since the release of OpenAI models, the trio said that GPT felt like a shot in the arm for a lot of people in the AI world. “We could see the technology really start to accelerate last year,” said Mensch.

Hopefully the company sticks to its words. Initially, OpenAI also started out as an open source company and now we know where they are headed. Now, Meta believes it is leading the way for open source, but as Mark Zuckerberg said that it is only because their technology is not on par with GPT-4 or PaLM 2.

OpenAI Killer?

OpenAI’s Sam Altman has been raising concerns about the problem with open sourcing models as it can create harmful models by bad actors for the world. But the developer community has been loving what Meta and Stability AI have been contributing to the open source community with LLaMa and StableLM models. 

Now with Mistral AI in the picture, with a focus solely on building open source models and datasets, the resentment that people have been building towards OpenAI might get amplified.

It might be very hard for the startup to acquire customers that have already adopted OpenAI’s GPT technology into their enterprises and companies. For this, Mensch said that the company wants to make it easy for companies to build their own AI products, instead of just relying on theirs. We want to give them tools that are easy to use to create their own products,” he said. 

What about India? 

Antonie Moyroud, the lead investor from Lightspeed Ventures, said that, “We had a chance to see a few strong tech teams looking to raise funding to build LLM across India, the U.S. and Europe. But what was lacking was that while they were strong technically they hadn’t thought about the second order effects of LLM.” 

This might be the best time for India to recover from Sam Altman’s India visit trauma, alongside many VCs sentiments, and start their own AI company, as it is clear that everyone’s a winner in the AI race.