Palantir’s Shout of “Enough is Enough”: Other Companies’ Fear of the “Winner-Takes-All” Effect in Large Models

```

A fierce statement from Palantir CEO Alex Karp has brought long-standing tensions in the tech industry to the surface—large AI labs are accumulating momentum using clients' data and decisions, while traditional enterprises are increasingly worried about becoming mere "value contributors" in the AI wave.

In the past two weeks, Karp first delivered nearly 20 minutes of intense criticism on CNBC, directly accusing AI labs of exaggerating their capabilities, overpricing tokens, and claiming that every major enterprise client he interacts with is "burning with anger" over these issues.

Shortly after, Palantir released a white paper titled "Institutional Sovereignty in the AI Era," listing fifteen recommendations for enterprises and governments to guard against core data erosion by AI giants like OpenAI and Anthropic. These two actions quickly sparked widespread discussion in the tech community.

The core of this debate is a question that is being asked louder and louder: Who captures the value in the AI era—is it the enterprises deploying AI, or the labs developing the underlying models?

This question is not only about business dynamics; it has spread into policy games and geopolitical competition, and directly threatens the valuation of traditional software companies.

Not Just Karp Speaking Out

Karp himself admits his position is not neutral.

Palantir's core products are built atop foundational models, serving as the intermediary layer connecting AI and enterprise customers—giving it direct commercial interests in the tussle between enterprises and AI labs.

To outside accusations that he is venting emotions, Karp responds:

No, this is the voice of American business, conveyed through me.

Notably, Karp is not the only tech executive warning about this imbalance.

Microsoft CEO Satya Nadella has recently published articles and repeatedly voiced similar concerns, focused on the question: Can enterprises truly retain the "learning outcomes" accrued from using AI models?

Nadella said this month at an event at Stanford University:

If you’re just a consumer of foundational models, I’m not sure how you can retain enterprise value, let alone create it.

AI Labs’ “Encroachment” Logic

Karp's criticisms touch a deeper anxiety in the tech industry.

Former White House AI lead David Sacks immediately echoed this viewpoint on social media, directly targeting Anthropic. Sacks wrote:

Anthropic has rolled out Claude Science, Claude Security, Claude Legal, and Claude Code—each product directly enters fields previously served by companies building applications based on its models.

Sacks further commented:

This pattern is consistent: observe where value is created, then dive in directly. Dominate the model layer first, then use this position to seize the most lucrative vertical markets.

This "observe-copy-expand" path has left many enterprises building commercial applications on large model APIs uneasy. For these companies, contributing data and usage scenarios to AI labs may be arming competitors for market entry.

Neither OpenAI nor Anthropic has publicly responded to Karp’s criticism. Both companies’ current policies state that enterprise client data will not be used to train their models.

An insider at an AI lab dismissed the issue, saying:

Responding to Karp-style theatrics is foolish; he is simply advocating for his own interests.

The Winners Are Not Yet Decided

The deep background to this debate is the entire industry’s uncertainty about the ownership of AI value.

WallstreetCN mentions that, this Thursday, media reported Starbucks is using AI to replace software previously sourced from Microsoft and IBM, and both companies’ stock prices subsequently came under pressure.

This case is seen as a microcosm of AI rapidly reshaping the enterprise software landscape. Analysts point out that, in the time it takes to brew a coffee, the winners and losers of the AI era may swap places, once again confirming a harsh reality: today’s tech giants may not secure tomorrow’s leadership.

Meanwhile, Meta last weekend announced the launch of a new AI model and introduced a paid tier. According to Bloomberg, Meta CEO Mark Zuckerberg said in an interview that he sees an opportunity to compete on price:

Some labs’ pricing is very extreme, with excessively high profit margins. We believe it’s entirely possible to provide cutting-edge or high-level intelligence services at a more affordable price.

This statement further intensifies competition in the foundational model market and indirectly confirms that Karp’s criticism of AI lab overpricing is not baseless.

Risk Warning and DisclaimerThe market has risks; investment should be cautious. This article does not constitute personal investment advice, nor does it take into account the specific investment objectives, financial circumstances, or needs of individual users. Users should consider whether any opinions, viewpoints, or conclusions in this article are suitable for their own situation. Any investments made accordingly are at your own risk. ```