Alex Karp is angry. The Palantir chief executive told CNBC on Wednesday that companies are burning cash on AI services that hand over their most valuable data in exchange for almost nothing. “Every single enterprise in this country, these people are livid.
They are paying for tokens that create no value. These people are stealing the weights and alpha of my business,” Karp said.
This is not the first time a technology shift has forced a reckoning between premium vendors and cheaper alternatives. The pattern repeats. In the 1990s, enterprises paid top dollar for proprietary Unix systems from Sun Microsystems and HP.
Then Linux arrived. Free, open-source, and good enough.
It ate the data center. In the 2000s, Oracle databases dominated — until PostgreSQL and MySQL proved they could handle most workloads for zero licensing cost. Cloud computing followed the same arc: AWS launched in 2006 as a low-cost option to buying servers; today it is the market leader.
Now the same cycle is hitting artificial intelligence. OpenAI and Anthropic sell frontier models at premium token-based prices. But a new generation of cheaper Chinese AI models is gaining traction.
The calculus is shifting. Enterprises are asking whether the premium is worth it.
The answer, for many, is no. Soaring AI bills are pushing companies to prioritize affordability over marginal improvements in model performance, according to a Reuters report. Tiezhen Wang, former head of the Asia-Pacific at Hugging Face, told Rest of World that many firms build products on proprietary models first, then switch to open-source alternatives as they scale. Switching later, he said, could allow firms to reduce spending on AI tokens substantially.
History suggests this is not a temporary cost-cutting measure. It is how markets mature.
The first movers — the Suns, the Oracles, the OpenAIs — charge a premium for being first. They own the customer relationship. Then the second wave arrives.
Lower-cost options that are 80% as good for 20% of the price. Enterprises grumble about quality. Then they switch anyway.
The premium vendors cut prices. The cycle repeats.
What makes this moment different is the geopolitical dimension. Chinese AI models are not just cheaper — they are Chinese. That introduces questions about data sovereignty, security, and regulatory compliance that did not exist when the choice was between Red Hat Linux and Solaris.
Karp’s warning about companies handing over “the weights and alpha of my business” takes on added weight when those weights might end up on servers in Beijing. But for CFOs watching AI budgets balloon, and CTOs weighing performance against cost, the appeal is straightforward.
The open-source approach, supporters argue, could accelerate adoption precisely because of its cost advantages. The same argument was made for Linux. For PostgreSQL.
For Kubernetes. Each time, the incumbents predicted disaster. Each time, the market proved them wrong.
The pricing battle is already here. Cheaper alternatives are competing with premium frontier models.
The question is not whether enterprises will switch — they already are. The question is how fast, and what the incumbents do about it. If history is any guide, they will cut prices, then complain about unfair competition, then cut prices again.
And the cycle will continue until the next technology shift comes along.
{“@context”: “https://schema.org”, “@type”: “NewsArticle”, “headline”: “Palantir CEO Alex Karp criticizes AI models for draining company data value.”, “description”: “Alex Karp is angry. The Palantir chief executive told CNBC on Wednesday that companies are burning cash on AI services that hand over their most…”, “datePublished”: “2026-07-02T07:15:41+00:00T09:30:00+00:00”, “author”: {“@type”: “Person”, “name”: “Li Wei”}, “publisher”: {“@type”: “Organization”, “name”: “InfoPulse Today”, “url”: “https://newsfeed.infopulsetoday.com/”}}
