European technological sovereignty in the field of artificial intelligence cannot rely solely on the French company "Mistral", warned French Economy Minister Roland Lescure, reported Agence France-Presse, BTA reported.
“If European AI is reduced to "Mistral", then we are lost“, Lescure told journalists in San Francisco.
“We cannot put all our eggs in one basket. We need to develop an entire ecosystem“, the minister said during his visit to Silicon Valley, where he met with investors and companies from the AI sector such as „OpenAI“, as well as the financial technology company „Stripe“.
Asked by France Press whether „Mistral“ could abandon the development of the most advanced artificial intelligence models and focus mainly on providing computing infrastructure, Lescure refused to express a preference.
„Will they focus more on the development of fundamental models, or will they become a very efficient and profitable services company - we'll see“, he said. Asked if he had recommended the first option to the company's management, the minister replied: „Companies do what they can and what they want“. At the same time, he defined „Mistral“ as a company „of which we can be proud“.
According to Lescure The most advanced American models are still in the lead, but the Chinese are not far behind, and there are developments in Europe that could lead to new progress.
„Mistral“ is often described as the only European company that creates top-class artificial intelligence models, although in terms of indicators they are still inferior to the leading American and Chinese competitors, notes Agence France-Presse.
In June, the European Commission entrusted an Italian-German consortium, in which „Mistral“ is not involved, with the development of an advanced European model on the EU's public supercomputers.
„Mistral“, valued at nearly 12 billion euros and in which the French state has an indirect stake through the state-owned investment bank „Bpifrance“, has increased its investment in infrastructure this year. The company opened its first data center in France and signed trade agreement with „Microsoft“ (Microsoft), while continuing to develop its own models.
Asked if there was a risk that the administration of US President Donald Trump could use the opportunity to limit Europeans' access to the most powerful artificial intelligence models as a means of pressure, Lescure answered with one word: „unpredictability“.
In June, Washington forced the American company „Anthropic“ (Anthropic), creator of the „Claude“ (Claude), to suspend access to all foreign citizens to its most powerful models for two weeks. The measure has again heightened concerns in Europe about the continent's technological dependence, AFP notes.