Loading article…
Palantir CEO Alex Karp calls OpenAI's deployment unit a “complete farce” and says AI firms are “unlikeable,” highlighting tensions as token‑based pricing
Palantir’s Alex Karp told CNBC that OpenAI’s new deployment company is “a complete farce” and that most AI firms “don’t understand how unlikeable they are,” underscoring growing friction between Palantir’s forward‑deployed model and the token‑driven approaches of OpenAI and Anthropic【2】.
| At a glance | |
|---|---|
| Person | Alex Karp, Palantir CEO |
| Critique | OpenAI deployment unit “complete farce” |
| Claim | AI firms “unlikeable” and “very expensive” |
| Context | Token‑based pricing now central to AI business models【1】 |
Karp argued that AI companies are “chillaxing over their latte” while ignoring the practical realities of token‑based billing, which charges customers per million tokens of input ($5) and output ($30) for OpenAI’s GPT‑5.5 model【1】. He said the pricing structure makes products “very expensive” and that the companies “don’t actually work the way customers expect.” By framing token consumption as a “farce,” Karp positions Palantir’s forward‑deployed, problem‑solving approach as a more customer‑centric alternative.
Wall Street analysts note that tokens are now the primary metric for AI usage, with Google reporting 16 billion tokens per minute—up from 10 billion the prior quarter【1】. OpenAI and Anthropic have filed confidential IPO prospectuses, and their token pricing will be a key focus for investors unfamiliar with the unit. The token model ties revenue directly to compute usage, a stark contrast to Palantir’s subscription‑based, outcome‑driven contracts.
OpenAI’s token pricing and its upcoming deployment unit aim to replicate Palantir’s forward‑deployed model, but Karp’s comments suggest he sees the effort as misguided. Meanwhile, hardware players like Cerebras tout token generation speed as a competitive edge, claiming their chips can produce tokens faster than Nvidia’s processors【1】. If token consumption continues to grow exponentially, AI firms will need ever‑larger compute capacity, potentially reinforcing the advantage of vertically integrated players like SpaceX, which claim a “token cost advantage” from owning data centers【1】.
Karp’s blunt assessment highlights a cultural clash: Palantir’s focus on solving concrete problems versus the token‑centric, usage‑based economics that dominate the AI frontier. Whether this tension reshapes how AI services are packaged and priced remains to be seen.
Coverage is mostly measured — 279 of 300 reports stay neutral.
Every Monday — the token unlocks, Fed dates & catalysts set to move crypto and markets this week. So you’re never blindsided.
Free · 3-min read · one-click unsubscribe
AI-assisted synthesis by the TrendWatcher Editorial Desk · sourced from 2 outlets · Jul 1, 2026 · How we report
OpenAI warns that AI technology has democratized access to hacking tools, enabling large-scale, automated attacks that could threaten hospitals, water plants, and internet infrastructure.
OpenAI stated it cannot be confident that SpaceX will comply with its terms of service, citing previous contract violations by other companies owned by Elon Musk.
OpenAI announced that it plans to shut off Cursor's access to its models on November 12.