Will AI-native companies (like Anthropic) disrupt Palantir?

$PLTR's premium is based on its AI platform (AIP) being irreplaceable. That companies’ disparate datasets need Palantir’s operating system to enable AI. It seems AI agents can now handle at least some of that!

• Palantir's bull case rests on the premise that integrating LLMs with messy enterprise data requires its proprietary stack and an army of Palantir’s forward-deployed engineers.

  Anthropic is now building a new enterprise model. ‘Computer Use’ and ‘Claude for Enterprise’ are products that may not need Palantir’s integration layer in a few years. Anthropic’s agents are commoditizing exactly what PLTR charges a premium for. The forward-deployed model ages poorly as agentic AI handles orchestration, stripping out the pricing power of expensive human deployments. CIOs under cost pressure will increasingly choose lightweight, API-first solutions over a rigid proprietary OS.

• At ~75x revenue, PLTR is priced for permanent irreplaceability. Even a partial erosion of the integration moat justifies significant multiple compression. I don’t see how Palantir can compete with AI-native companies like Anthropic and OpenAI. It does not have the resources to create its own LLM-based agents. I am sure it’s easier for Anthropic to partner with a consulting firm and deploy its agents than for Palantir to build its own models.

Palantir is valued at ~$330B for $5B in revenue. It will need to grow its revenue to at least $30B to justify this valuation. To do so while its moat is weakening will be extremely tough. Am I missing something?