Cognition CEO Scott Wu has rejected a report that SpaceX tried to acquire his AI coding company, saying the startup is not for sale and that the two companies have not held acquisition talks. The denial came shortly after Bloomberg reported, through sources familiar with the matter, that SpaceX had explored a deal.

The dispute adds uncertainty to SpaceX’s widening push into artificial intelligence. The company recently completed a reported $60 billion acquisition of Cursor, has absorbed Elon Musk’s xAI business, and is positioning AI as a larger part of its long-term commercial strategy. For AI builders and enterprise buyers, the episode highlights how quickly leading coding-agent companies are becoming strategic acquisition targets—even when a reported transaction does not materialize.

What was reported—and what Cognition says

Bloomberg’s report, as described by TechCrunch, said SpaceX had held talks to buy Cognition, the developer of the Devin coding agent. The talks were reportedly no longer active, although the companies could still discuss a commercial relationship involving SpaceX’s computing capacity.

Wu disputed the central account in a post on X. He said Cognition is “not for sale” and that SpaceX and Cognition have not been in talks. His statement did not address the separate possibility that Cognition could use SpaceX’s infrastructure or that the companies might cooperate without an acquisition.

Neither SpaceX nor Cognition provided additional comments to TechCrunch, according to that publication. Investing.com and Seeking Alpha separately carried headlines attributing the acquisition report to Bloomberg, but their available source material did not include full article text or independent confirmation. As a result, the confirmed development is Cognition’s denial, while the alleged negotiations remain a media report that the company disputes.

Cognition is one of the larger independent AI coding companies still operating outside a major model provider, according to TechCrunch’s account. The startup raised $1 billion in late May at a reported post-money valuation of $25 billion. Bloomberg also reported that Cognition was in early discussions for another funding round at a potential $40 billion valuation, though that financing signal has not been confirmed by the company in the supplied evidence.

Why SpaceX is pursuing coding software

The reported approach would fit SpaceX’s broader effort to build an AI business that can compete with OpenAI, Anthropic, and Google. Musk has reportedly told SpaceX employees that AI could represent 99% of the company’s value in roughly four or five years. Turning that ambition into revenue, however, requires products that businesses will buy and deploy now.

AI coding is one of the clearest commercial categories available. Coding agents can assist with software generation, debugging, testing, and multi-step development tasks, giving vendors a path to recurring enterprise revenue. Cognition’s Devin would have offered SpaceX another established product and a customer base that TechCrunch identified as including Mercedes-Benz, Citi, and Goldman Sachs.

SpaceX already has a direct position in the category through Cursor. TechCrunch reported that SpaceX completed its acquisition of Cursor last week and that the two companies had been working together before the transaction closed. Cursor and SpaceX also jointly released Grok 4.6 this month, with the model reported to perform strongly on coding and complex agentic-task benchmarks.

Those benchmark results are not independent evidence of product superiority in enterprise settings. They are performance claims reported in the source coverage, and real-world adoption will depend on factors such as code reliability, security controls, integration with existing development systems, and the cost of running agents at scale.

The strategic stakes for enterprise AI

Acquiring Cognition would have given SpaceX more than another coding product. It could have added Devin, Cognition’s engineering and product teams, and relationships with large corporate buyers. That combination could help SpaceX move beyond consumer-facing experiments such as Grok and toward workplace automation and software development workflows.

The company’s AI position remains complicated. SpaceX has significant ambitions for computing infrastructure, including the eventual construction of data centers in space, while its AI operations are still described in the coverage as relatively early. The business has also faced controversy around Grok, including incidents involving offensive content and nonconsensual sexual imagery. Those issues could make enterprise trust, governance, and compliance central obstacles as SpaceX seeks corporate customers.

Cognition’s independent status also matters to the market. The startup acquired the remaining assets of Windsurf after Google DeepMind hired the rival company’s chief executive and senior research staff in a reported $2.4 billion talent and licensing arrangement. Cognition subsequently laid off 30 employees and offered buyouts to roughly 200 Windsurf employees, according to TechCrunch. The restructuring underlines the operational pressure facing coding startups even as investor interest and acquisition speculation remain high.

For founders, the reported talks show both the opportunity and risk of remaining independent. A high valuation can provide room to build a standalone platform, but major infrastructure and model companies may still view a coding agent as a valuable distribution layer. For buyers, a rumored acquisition is not evidence that a product is available, strategically aligned, or likely to remain independent.

What to watch next

The first signal will be whether SpaceX and Cognition announce any partnership involving computing capacity. Bloomberg reportedly linked the companies through a possible infrastructure relationship, and TechCrunch noted that SpaceX sells computing resources to other AI companies, including Anthropic, until it requires more capacity for its own operations. Cognition’s denial did not specifically reject that possibility.

Investors and enterprise customers should also watch for details about how Cursor fits into SpaceX’s product structure. The key questions are whether SpaceX keeps Cursor as a standalone coding assistant, connects it more closely to Grok, or uses it as the foundation for a broader enterprise AI platform.

Cognition’s next financing round will be another important indicator. A new raise at the reported valuation would suggest that the company can continue funding growth without selling. Conversely, a strategic investment, infrastructure agreement, or leadership change could revive acquisition speculation even if Cognition remains formally independent.

Finally, customers will need evidence beyond benchmark scores: retention, deployment scale, security certifications, data-handling policies, and measurable developer productivity. Those signals will matter more than any single acquisition rumor when companies decide whether to place coding agents inside production engineering workflows.

Creati.ai perspective

The immediate news is not a confirmed acquisition but a contested report. Wu’s denial should prevent the market from treating Cognition as part of SpaceX, while the lack of a response on possible infrastructure cooperation leaves some commercial links unresolved.

The larger lesson is that AI coding has become a strategic battleground because it connects models to paid, repeatable business workflows. SpaceX’s reported interest in Cognition—whether or not talks occurred—shows why coding agents, compute capacity, and enterprise distribution are increasingly being evaluated together. The companies that win this market will need to demonstrate dependable software work and trustworthy operations, not just high benchmark results or headline valuations.