Palantir CEO Alex Karp says enterprises are fed up with frontier AI labs like OpenAI and Anthropic

Palantir CEO Launches Scathing Attack on Frontier AI Labs

Palantir CEO Alex Karp delivered a blistering critique of frontier AI labs during a CNBC interview on Wednesday, claiming that every single enterprise customer his company works with is frustrated with companies like OpenAI and Anthropic1

1

. According to Alex Karp, these AI labs operate on a "hyper religion of hyper optimism" that fails to reflect the actual experiences of their corporate customers1

1

. The Palantir chief's comments highlight a growing disconnect between the promises of cutting-edge AI developers and the practical value enterprises are actually extracting from their technologies.

Enterprise Customers' Needs Ignored by AI Labs

Karp's central complaint revolves around what he perceives as a fundamental misunderstanding of enterprise customers' needs by frontier AI labs. "They believe all problems present, past, and future, including the ones they create but don't acknowledge, are going to be solved by them," Karp said of companies like OpenAI and Anthropic1

1

. He characterized their approach as largely religious, with AI labs telling businesses that their current problems will simply disappear tomorrow. The Palantir CEO suggested that frontier labs don't engage with enterprises or understand the technical challenges their customers face when deploying AI models1

1

. Karp even revealed that Palantir leadership has debated paying potential customers to visit frontier labs before signing contracts, claiming people emerge "screaming" that these solutions could never work for their businesses1

1

.

Tokenmaxxing and the ROI Problem

The criticism of frontier AI labs extends to what Karp termed "tokenmaxxing"—the practice of viewing token consumption as a measure of productivity and usefulness rather than actual business outcomes1

1

2

2

. This approach is becoming increasingly expensive for companies, with OpenAI reportedly considering reducing its per-token charge to attract more customers in its growing competition with Anthropic, which Karp called the "leading frontier firm"1

1

. The ROI of AI projects has become a critical issue, with recent Gartner estimates showing that only 28 percent of AI use cases fully meet ROI expectations, and most fail to ever get out of the pilot stage1

1

. Despite these dismal returns, business leaders continue investing heavily in AI deployment, searching for value that Karp insists isn't there without proper infrastructure.

Palantir's Foundry as the Alternative Solution

Karp positioned Palantir's approach as the antidote to frontier lab failures, with enterprise frustration driving businesses toward the company's Foundry systems1

1

. These data integration platforms act as AI-agnostic systems for unifying disparate data sources and pairing them with whatever LLMs a customer chooses to deploy. "It is not that LLMs aren't crucial for the world, it's just that the implementation is where the value is, certainly in the next 7 years," Karp explained1

1

. He even claimed that most of what Anthropic publicly celebrates as successful is "running on Palantir"1

1

. When OpenAI announced its acquisition of UK-based AI consulting firm Tomoro to form the OpenAI Deployment Company aimed at helping customers generate returns, Karp dismissed it as "a complete farce," arguing that AI labs don't understand "how unlikeable they are" because "the product doesn't actually work and it's very expensive"1

1

.