NEW YORK — Alex Karp did not mince words. The Palantir chief executive told CNBC on Wednesday that too many American companies are burning cash on AI services that deliver nothing useful. Worse, he said, those same services are siphoning off the proprietary data that makes a business valuable.
“These people are stealing the weights and alpha of my business,” Karp said. He was blunt about the resentment building in boardrooms.
“Every single enterprise in this country, these people are livid. They are paying for tokens that create no value.” Karp’s outburst lands in a market already shifting underfoot. A Reuters report shows that soaring AI bills are forcing companies to rethink their priorities.
The calculus is straightforward: pay a premium for a frontier model from OpenAI or Anthropic, or take a cheaper alternative that may be only slightly worse. For CFOs watching their budgets swell and CTOs trying to justify the spend, the answer increasingly tilts toward thrift.
Enter the Chinese models. A new generation of lower-cost AI alternatives from China is gaining traction. These systems compete on price, not on raw performance benchmarks.
They are not trying to beat GPT-4 at every test. They are trying to be good enough for a fraction of the cost. And that pitch is landing.
Tiezhen Wang, former head of the Asia-Pacific region at Hugging Face, told Rest of World that many companies start with proprietary models. They build their first products on the big names.
But as they scale, the bill grows. Wang said that switching to open-source alternatives later can cut spending on AI tokens substantially. The open-source route, he argued, could speed up adoption precisely because it costs less.
The result is a pricing battle across the industry. Cheaper alternatives are competing directly with premium frontier models.
The gap between what you pay and what you get is narrowing. For a company running hundreds of thousands of queries a day, the difference between a 99% accurate model and a 97% accurate model may not matter. The difference between a $10,000 monthly bill and a $2,000 monthly bill does.
Karp’s warning is not just about cost. It is about control. He said companies are handing over their most valuable business information in exchange for tokens that generate no value.
That trade-off looks worse when the bills come due. The CFOs are livid, as Karp put it.
The CTOs are weighing performance against cost. And the Chinese AI firms are waiting. The shift is not theoretical.
It is happening now. Enterprises are reassessing what they get for their AI spending.
The premium models still lead in raw capability. But capability does not matter if nobody can afford to use it at scale. The open-source alternatives, many of them Chinese, are filling that gap.
They are not stealing the market. They are inheriting it.
{“@context”: “https://schema.org”, “@type”: “NewsArticle”, “headline”: “Palantir CEO warns US firms wasting cash on AI services that steal data”, “description”: “NEW YORK — Alex Karp did not mince words. The Palantir chief executive told CNBC on Wednesday that too many American companies are burning cash…”, “datePublished”: “2026-07-02T07:15:41+00:00T09:30:00+00:00”, “author”: {“@type”: “Person”, “name”: “Henry Thomas”}, “publisher”: {“@type”: “Organization”, “name”: “InfoPulse Today”, “url”: “https://newsfeed.infopulsetoday.com/”}}
