business

Palantir CEO Karp Slams OpenAI and Anthropic Token Pricing Model

Alex Karp argues soaring token costs are pushing enterprises away from closed AI models toward open-weight alternatives built for efficiency.

Palantir CEO Alex Karp launched a sharp public attack on OpenAI and Anthropic Monday, declaring that runaway token costs in their pricing models signal a fundamental breakdown in how leading AI companies are structured. "Something has gone completely wrong," Karp said, framing the issue as a systemic failure rather than a temporary market adjustment.

At the heart of Karp's critique is the practice he calls "tokenmaxxing" — a strategy in which AI providers maximize revenue by charging enterprises for every token processed, regardless of whether that processing delivers proportionate value. As those costs spiral upward, Karp argues, corporate customers are being forced to make an uncomfortable choice between budget discipline and AI capability.

Read more Apple Negotiates With Blacklisted Chinese Chipmakers Over Memory Supply →

The consequence, according to Karp, is a growing migration toward open-weight models — publicly available AI architectures that companies can run and optimize on their own infrastructure without paying per-token fees to a centralized provider. This shift, if it accelerates, could represent a meaningful competitive threat to the closed-model business strategies that both OpenAI and Anthropic have built their valuations around.

Karp's comments carry weight beyond a typical executive rivalry. Palantir sits at the intersection of enterprise software and AI deployment, giving the company a front-row view of how Fortune 500 firms and government agencies actually budget for and consume AI services. His framing — that efficiency, not raw model power, is becoming the dominant corporate priority — reflects a broader tension developing across the AI industry as the post-hype phase of enterprise adoption sets in.

The remarks add to mounting scrutiny of whether the token-based revenue model is sustainable as competition intensifies and open-source alternatives mature. Continue reading at US Top News and Analysis.

Continue reading at US Top News and Analysis →

Frequently Asked Questions

Q.What is tokenmaxxing and why is Karp criticizing it?

Tokenmaxxing refers to AI providers maximizing revenue by charging enterprises for every token processed. Karp argues this practice inflates costs without delivering proportionate value to customers.

Q.What are open-weight models and why are companies moving toward them?

Open-weight models are publicly available AI architectures that companies can run on their own infrastructure, avoiding per-token fees charged by providers like OpenAI and Anthropic. Karp says skyrocketing token costs are accelerating this shift.

Q.Why does Alex Karp's criticism of OpenAI and Anthropic matter?

As CEO of Palantir, a company that deploys AI for major enterprises and government agencies, Karp has direct insight into how large organizations budget for AI. His comments reflect growing corporate pressure to prioritize efficiency over raw model capability.

More in business →