Reid Hoffman criticizes SpaceX's AI strategy as 'buying your way into relevance' and calls xAI a 'complete train wreck,' while raising alarm over the U.S. government's unilateral suspension of Anthropic's Fable and Mythos models.

What happened

LinkedIn co-founder and investor in both Anthropic and OpenAI Reid Hoffman said on the Pioneers of AI podcast that SpaceX is not an AI company but rather a 'premium-priced CoreWeave' leasing infrastructure, and that xAI—which has lost all 11 of its original co-founders and is on its 'third restart'—is a 'complete train wreck' for building foundational models. He also called SpaceX's acquisition of Cursor coding tool evidence of AI absence, not capability, comparing the strategy to a roll-up conglomerate.

Why it matters

Hoffman's assessment matters because he speaks from the vantage point of a founder, board member at Microsoft, and investor in multiple AI companies—making his skepticism a substantive critique rather than outside commentary. His alarm about the U.S. government forcing Anthropic to pull its Fable and Mythos models on June 11th as an export control order suggests unpredictable regulatory intervention poses a concrete investor risk, especially as Anthropic prepares for a major IPO. The asymmetry—Anthropic penalized while OpenAI was not—undermines confidence in rule-of-law governance of AI policy.

What to watch

Hoffman argued there is room for both OpenAI and Anthropic to win, positioning them in distinct competitive lanes: Anthropic strong in code and expanding into design and legal, while OpenAI functions more like a consumer search front-end. He also flagged whether Cursor, now owned by SpaceX, has 'already peaked,' noting it 'seems to be fading over the horizon' as Claude Code and Codex have gained ground since early 2026.