Palantir publishes sovereign AI and data manifesto as 54% of revenue comes from government contracts
Palantir has posted on X a condensed summary of the book "The Technological Republic" by CEO Alex Karp and Nicholas Zamiska. It urges customers not to hand their data to large language model vendors and to keep the models and the logs of their operation in-house. According to the company's 2025 annual report, 54% of its revenue came from government customers and 46% from commercial clients.
AI-processed from Habr AI; edited by Hamidun News
Palantir Technologies has published a post on the social network X titled "The Technological Republic, in brief," a condensed summary of the book "The Technological Republic" by the company's CEO Alex Karp and Nicholas Zamiska. The message: don't give your data to large language model vendors, and don't judge AI adoption by token consumption.
What Palantir proposes
Palantir offers customers a simple rule: keep the models, the traces of the system's work and the fine-tuning results with you, not with a third-party LLM vendor. In effect, the company is pitching a version of "sovereign AI": deploy models on your own side instead of passing data to external large language model providers. The manifesto is about control rather than savings: over data, access and the outputs of AI systems.
- The manifesto post is titled "The Technological Republic, in brief"
- It is a digest of the book "The Technological Republic" by Alex Karp and Nicholas Zamiska
- It was published on Palantir's official account on X
- The company describes its market as solutions "from the factory floors to the front lines"
Why it matters and who is speaking
Palantir is not an open-source startup. It sells tools for connecting data, managing access rights, domain-specific models and operator actions to the military, intelligence agencies and large businesses. According to Palantir Technologies' 2025 annual report filed with the US Securities and Exchange Commission (SEC), 54% of the company's revenue came from government customers and 46% from commercial clients.
"From the factory floors to the front lines" is how
Palantir describes the market for its solutions on the company's official website.
Government work is not a decorative "other nice use cases" section of the business. It is almost half of the company's revenue, which is why a call to keep data and models "in-house" sounds different from it than from an ordinary tool developer.
What this means
When a company with this revenue structure talks about "saving on tokens," it is not about thrift but about a new layer of control: where the data lives, who runs the model, who sees the requests, who gets the traces of the system's work, and who ends up as a mandatory part of the whole arrangement.
Frequently asked questions
What is "The Technological Republic"?
It is a book by Palantir CEO Alex Karp and Nicholas Zamiska. The company released a condensed summary of it as an X post titled "The Technological Republic, in brief."
How much of Palantir's revenue comes from governments?
According to the 2025 annual report filed with the SEC, 54% of Palantir's revenue came from government customers and 46% from commercial clients.
Key Questions Answered
What is "The Technological Republic"?
It is a book by Palantir CEO Alex Karp and Nicholas Zamiska. The company released a condensed summary of it as an X post titled "The Technological Republic, in brief."
How much of Palantir's revenue comes from governments?
According to the 2025 annual report filed with the SEC, 54% of Palantir's revenue came from government customers and 46% from commercial clients.
What Palantir proposes?
Palantir offers customers a simple rule: keep the models, the traces of the system's work and the fine-tuning results with you, not with a third-party LLM vendor. In effect, the company is pitching a version of "sovereign AI": deploy models on your own side instead of passing data to external large language model providers. The manifesto is about control rather than savings: over data, access and the outputs of AI systems. The manifesto post is titled "The Technological Republic, in brief" It is a digest of the book "The Technological Republic" by Alex Karp and Nicholas Zam
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.