A Report, Then a Denial

SpaceX was reported to have explored an acquisition of AI coding startup Cognition, but Cognition CEO Scott Wu quickly rejected the account. Posting on X, Wu said the story was inaccurate, that Cognition “is not for sale,” and that the two companies have not been in talks.

The report, published by Bloomberg and attributed to people familiar with the matter, landed only days after SpaceX completed its $60 billion acquisition of Cursor, another AI coding startup. That timing made the claim notable: SpaceX is trying to position itself more aggressively against OpenAI, Anthropic, and Google in enterprise AI.

Why AI Coding Matters to SpaceX

According to the report, SpaceX acquired xAI earlier this year and then went public in June, with its market capitalization rising to nearly $2.3 trillion at its peak. The company has been pitching investors on a broader AI strategy, including the long-term idea of building data centers in space.

But the AI business is still relatively young compared with leading rivals. It has also faced reputational challenges around Grok, including last year’s “MechaHitler” episode and this year’s nonconsensual sexual imagery scandals. Those controversies matter because enterprise AI customers tend to value reliability, governance, and brand safety.

AI-assisted coding is one of the clearest ways to turn generative AI into revenue. In simple terms, an AI coding agent helps developers write, debug, test, or complete software tasks with less manual effort. Anthropic’s growth, driven in part by Claude Code, shows why coding workflows have become a strategic battleground.

Cursor, Devin, and Enterprise Customers

Cursor and SpaceX were already working together before the acquisition closed. This month, Cursor and SpaceX jointly released Grok 4.6, a model that the companies say performs better on coding benchmarks and complex multi-step agentic tasks. An agentic task is one where AI breaks down a goal, follows steps, and may use tools rather than simply answering a prompt.

Cognition would have brought a different asset: Devin, its coding agent, plus an enterprise customer base that reportedly includes Mercedes-Benz, Citi, and Goldman Sachs. That kind of customer list is valuable for any company trying to move from AI demos to recurring enterprise software revenue.

Key figures from the report include:

• Cognition raised $1 billion in late May at a $25 billion post-money valuation;
• Bloomberg says the company is now in early talks for a new round at a $40 billion valuation;
• Bloomberg also says acquisition talks are no longer active;
• The companies may still discuss collaboration, potentially involving Cognition’s use of SpaceX computing capacity.

Wu’s denial did not specifically address the possible computing-capacity discussion. SpaceX and Cognition did not respond to requests for comment.

Independence Under Pressure

Cognition remains one of the largest independent AI software coding startups not yet absorbed by a major AI model developer. It drew attention last year when it acquired the remaining assets of Windsurf after Google DeepMind struck a $2.4 billion deal for Windsurf’s CEO, top research talent, and licensing rights.

After that merger, Cognition laid off 30 employees and offered buyouts to the remaining 200 Windsurf employees. Those who stayed reportedly faced demanding expectations, including more than 80 hours of work per week and six days in the office. That intensity would not be unfamiliar in a Musk-led environment; Musk has said he works up to 120 hours a week and often sleeps on office or factory floors.

What Comes Next

The disputed report highlights a broader shift in AI competition. The focus is moving beyond chatbots and model releases toward products that sit directly inside daily business workflows. Software development is especially attractive because productivity gains can be measured, budgets already exist, and customers are accustomed to paying for developer tools.

Whether SpaceX ever pursued Cognition or not, the Cursor deal and Cognition’s reported funding discussions point in the same direction: AI coding has become a gateway to enterprise AI revenue. Large platforms may keep buying or partnering with specialist startups, while independent players will try to use high valuations and customer traction to remain independent longer.