AI sovereignty but make it French

Arthur Mensch stands in New Delhi (2026) warning that three or four companies have too much power over AI. Beautiful. The CEO of Mistral, valued at twelve billion euros through the exact same venture capital circus that bankrolls OpenAI and Anthropic, is here to save you from American tech giants. The new Robin Hood.

Here’s what’s actually happening. India is paying two hundred and fifty billion dollars to switch dealers. Instead of buying your AI crack from OpenAI, you’re now buying it from Mistral. Congratulations on your independence. Oh wait, the servers those French models run on? Those belong to Amazon, Google and Microsoft. Sovereignty for idiots, with extra steps.

Mensch babbles about open source like it’s a moral position. Open source means everyone can see how the hallucinating oracle works instead of just the company that built it. Brilliant, transparent unreliability. India is now buying access to technology that lies just as hard as the American version, but with a French accent and a story about democratization.

The funny irony? Mistral is selling this as resistance to power concentration while raising capital from General Catalyst and ASML to join the exact same accumulation dance. The money India is spending isn’t going to public infrastructure. It’s flowing to Reliance and Adani, two consortiums that will soon be running the same surveillance business models as Silicon Valley, just in Mumbai.

Countries are now building their own AI capacity because nobody wants to be the colonel who has to explain afterwards why they were dependent on American companies for critical infrastructure. Rational? No. Politically inevitable? Absolutely. India is buying back its freedom from American tech giants by applying for a job with the new Napoleon.