Palantir's official Twitter account just dropped a 9-point statement explaining 'AI Sovereignty,' and right after, their CEO went on CNBC to call out OpenAI and Anthropic. Key points from the statement: ✅ AI sovereignty determines the future of organizations Giving up control over your data and models is essentially handing over your future choices to someone else, who will most likely profit from your data while you bear the losses. ✅ Data is the core asset of a business Handing it over to third parties for model training means giving away your accumulated advantages, along with your ability to build the next competitive edge. ✅ Criticism of the 'pay-per-token' model This approach encourages companies to write disposable scripts instead of solid, sustainable software, creating a false sense of rapid progress. In an interview the next day, Karp doubled down on this logic, bluntly stating, 'These models are being irresponsibly oversold,' and emphasized that clients need to figure out one key thing: who exactly controls the model weights. Palantir, long known for providing data analytics and AI tools to the U.S. military and intelligence agencies, raises a thought-provoking question with this statement and Karp's remarks: 'Should companies entrust their lifeblood to a few closed-source large model companies?'