Summary
OpenAI announced in a late Friday post that it will wind down its contract supplying models to Cursor, the popular AI coding platform recently acquired by SpaceX AI. Matthew Berman breaks down the announcement, the decade-plus of corporate history behind it, and what it signals for the competitive landscape among frontier AI labs and the developers who depend on them.
The conflict traces back to the longstanding Elon Musk–Sam Altman rift: Musk’s founding role at OpenAI, his 2017 departure after failing to take over as CEO, the 2023 launch of xAI as a direct competitor, and subsequent litigation over OpenAI’s nonprofit-to-for-profit transition. OpenAI’s stated reason — that it cannot trust SpaceX to comply with its terms of service given a history of contract violations by Musk’s companies — is unusually pointed by corporate standards. SpaceX AI’s Michael Truel responded that OpenAI models represent only about 5% of Cursor’s traffic, a claim Berman scrutinizes carefully, noting that token volume and revenue are very different metrics and that frontier model tokens from OpenAI and Anthropic likely represent a disproportionate share of actual spend.
Berman also covers the strategic dimension: Cursor’s trove of developer coding session data is central to the SpaceX acquisition rationale, Anthropic and Google remain available as model providers, and Cursor has been building its own second-tier model for roughly a year. The episode raises pointed questions about the risks of deep integration with any single AI vendor’s models inside developer tooling.
📺 Source: Matthew Berman · Published September 01, 2026
🏷️ Format: News Analysis
