Cognition Pushes Back on SpaceX Acquisition Report as AI Coding Race Intensifies

SpaceX did not reach a deal to buy AI coding startup Cognition, and Cognition CEO Scott Wu says the company is not for sale. The denial came after a Bloomberg report said Elon Musk’s aerospace company had explored acquiring the startup as it expands its artificial intelligence ambitions and tries to narrow the gap with OpenAI, Anthropic, and Google.

The episode underscores how quickly AI coding tools have become strategic assets in the competition for enterprise customers, talent, and computing power. It also shows that SpaceX’s AI push is moving beyond headlines: the company is already tying its future to software, model development, and the infrastructure needed to run them.

Wu publicly rejected the acquisition story on X shortly after it appeared, saying the account was inaccurate and that no takeover discussions were underway. SpaceX and Cognition did not respond to requests for comment.

What the reported SpaceX-Cognition talks say about Musk’s AI strategy

The reported interest in Cognition fits a broader pattern across Elon Musk’s companies, which have increasingly treated artificial intelligence as a core business priority rather than a side project. SpaceX has already acquired Musk’s AI company xAI, gone public in a blockbuster listing, and used the resulting market power to make the case that AI will become a major source of value for the company.

Musk has been telling employees that AI could account for the vast majority of SpaceX’s value within a few years. That is an ambitious claim, but it helps explain why the company has been looking for ways to deepen its AI capabilities as fast as possible.

The reported pursuit of Cognition also makes sense in practical terms. AI coding products have become one of the clearest business models in the sector, with companies willing to pay for tools that help engineers build software faster, automate repetitive tasks, and support enterprise workflows. For a company like SpaceX, which already sells advanced technology to governments, enterprises, and consumers, stronger software tools could become both an internal advantage and a product opportunity.

Why AI coding startups are suddenly so valuable

AI coding has emerged as one of the strongest commercial use cases in the broader AI boom. Unlike some consumer-facing chatbot products that are still searching for durable revenue, coding assistants can be sold directly to developers and enterprise teams with a clearer value proposition: faster output, less manual work, and lower development costs.

That dynamic has made coding startups highly attractive to strategic buyers. Anthropic’s rapid growth, powered in part by Claude Code, has shown how quickly a coding-focused product can turn into a major revenue engine. Cursor has also become a standout player in the category, and SpaceX recently completed a $60 billion acquisition of that startup, according to the report.

The Cursor deal was followed by a joint release this month of Grok 4.6, a new model the companies said scored well on coding and complex multi-step agentic tasks. That collaboration suggests SpaceX is not merely collecting AI assets; it is trying to build a stack that includes models, tooling, infrastructure, and enterprise reach.

How Cognition would have fit into that stack

Cognition would have added another layer to that strategy. The startup is best known for Devin, its AI coding agent, and it also brings an enterprise customer list that includes Mercedes-Benz, Citi, and Goldman Sachs. In strategic terms, that combination would have given SpaceX a product with proven enterprise appeal and another route to monetization beyond model development alone.

Owning Cognition could also have helped SpaceX position itself more directly against the fastest-growing AI software companies. In a market where large model developers are increasingly bundling their own coding tools, independent startups are becoming acquisition targets because they already have users, workflows, and paying accounts.

Below is a quick snapshot of the key developments surrounding the reported deal interest.

What did Scott Wu say about the acquisition report?

Wu said the report was wrong and made clear that Cognition has not been engaged in sale talks with SpaceX. His public response was brief but firm, aimed at shutting down speculation before it could harden into market lore.

Cognition is not for sale, and the two companies have not been in talks, Scott Wu said after the report was published.

That denial matters because it comes at a time when AI startups are often viewed less as standalone businesses and more as likely targets for larger platform companies. A direct rebuttal from a CEO can be an attempt to protect strategic optionality, reassure employees, and maintain bargaining power with investors and prospective partners.

The denial did not completely rule out collaboration. According to Bloomberg’s report, the companies may still be discussing ways to work together, including the possibility that Cognition could use SpaceX computing capacity. Wu did not publicly address that part of the report.

How computing capacity is becoming a bargaining chip

Access to compute has become one of the most important currencies in AI. Model training and inference require huge quantities of specialized hardware, and companies with excess capacity can use it to attract partners, generate revenue, or build strategic relationships.

SpaceX appears to be exploring that model. The company has reportedly been offering computing capacity to outside AI players, including Anthropic, until it needs more of that capacity for its own operations. If Cognition were to become a customer or partner, even without an acquisition, it would signal a closer integration between Musk’s infrastructure ambitions and the software companies that depend on that infrastructure.

That potential arrangement is important because it highlights a different route to influence in AI: not just owning the applications, but controlling the resources that make those applications possible. In a market where training large models can cost hundreds of millions of dollars and where inference demand continues to climb, compute access can be as valuable as equity.

What is SpaceX trying to build beyond rockets?

SpaceX is trying to turn itself into a broader technology platform, with AI as one of the pillars of that transformation. The company’s core business remains launch and aerospace infrastructure, but its strategic narrative increasingly includes software, autonomy, advanced models, and future-facing computing infrastructure such as space-based data centers.

That vision is unusual even by Silicon Valley standards. Building data centers in space would be technically difficult and economically uncertain, but it reflects the scale of ambition Musk is willing to attach to the company. It also suggests that SpaceX sees AI not merely as an internal productivity tool but as a long-term business layer that could sit atop its existing hardware dominance.

The challenge is execution. SpaceX’s xAI efforts are still considered early relative to leaders such as OpenAI, Anthropic, and Google. Meanwhile, the company must manage the reputational baggage that has come with Grok, which has faced criticism for controversial outputs and safety lapses. Those issues matter because enterprise customers generally demand reliability, brand safety, and strong trust signals before integrating AI into critical workflows.

Why Grok’s controversies matter for enterprise adoption

The controversy around Grok is more than a public relations problem. It is a reminder that enterprise AI buyers often judge vendors on predictability, governance, and safety, not just model quality. Incidents involving harmful or embarrassing outputs can slow sales cycles and make customers cautious.

For SpaceX, that means a strong AI strategy needs more than flashy model releases. It needs credibility, controls, and a product portfolio that can stand up to scrutiny from large companies and regulated industries. That is one reason a startup like Cognition, with established enterprise relationships, would have looked attractive if the acquisition report had been true.

What happened to Cognition before this report?

Cognition has moved quickly from startup status to one of the most prominent independent coding companies in AI. In late May, it raised $1 billion at a $25 billion post-money valuation, placing it among the most richly valued private AI software firms.

Bloomberg also reported that the company is now in early conversations about a new funding round at a $40 billion valuation. If accurate, that would indicate investor appetite remains strong despite a crowded market and rising expectations around product performance and growth.

The company’s rise has been shaped in part by the turbulence around rival Windsurf. Last year, Cognition acquired the remaining assets of Windsurf after Google DeepMind hired away the startup’s CEO and top researchers in a $2.4 billion talent-and-licensing deal.

That transaction gave Cognition a way to consolidate talent and technology in a fast-moving category. But it also exposed the company to the hard operational realities of post-merger integration, including layoffs and cultural consolidation.

How did the Windsurf deal affect Cognition?

The Windsurf acquisition gave Cognition more scale, but it also came with immediate restructuring. The company laid off 30 employees and offered buyouts to the remaining 200 Windsurf staff members. Those who stayed were reportedly expected to accept extremely demanding working conditions, including long hours and a six-day in-office schedule.

Those rules reflected a deliberate culture choice: a startup atmosphere built around intensity, speed, and total commitment. The approach is not unusual in Musk-adjacent companies, where long workweeks and extreme dedication are often framed as necessary to win in highly competitive industries.

In that sense, Cognition’s internal expectations would not have been out of place inside SpaceX. Musk has repeatedly described his own work habits in similarly extreme terms, saying he works very long hours and sometimes sleeps at work. The overlap in philosophy helps explain why the acquisition rumor felt plausible to many observers even before Wu denied it.

Why does the rumor matter even if the deal never happened?

The report matters because it reveals how aggressively major companies are positioning themselves for the next phase of the AI market. Even without a completed acquisition, the fact that SpaceX was linked to Cognition suggests the company is evaluating where control over software, customers, and compute will matter most.

It also highlights a broader market reality: elite AI startups are no longer judged only by their products, but by how they fit into the ambitions of larger tech and industrial players. When a startup has strong enterprise traction, a well-known product, and access to valuable talent, it becomes a strategic object as much as a standalone business.

That helps explain why the reported deal resonated so widely. SpaceX is not known primarily as a software company, but Musk’s ecosystem increasingly blurs the line between aerospace, AI, and infrastructure. If one of the world’s largest private companies is willing to chase AI coding startups, that is a sign the category has become central to the future of enterprise software.

What happens next for SpaceX and Cognition?

The most immediate outcome is that the acquisition talk appears to be over, at least for now. Bloomberg said the deal discussions are no longer active, though cooperation may still be on the table. That leaves open the possibility of some form of commercial arrangement centered on compute or product collaboration rather than ownership.

For Cognition, the reported interest may reinforce its valuation and strategic importance even as it remains independent. For SpaceX, the episode shows the company is still searching for the right combination of assets to accelerate its AI ambitions. The company’s next moves will likely be watched closely by investors, enterprise customers, and competitors alike.

The larger takeaway is that AI coding has become one of the most contested corners of the technology industry. Startups such as Cognition, Cursor, and Anthropic are not just building tools; they are building leverage. Whoever controls the best coding products may shape how software gets made across the enterprise economy.

For now, Cognition insists it is not on the market. SpaceX, meanwhile, continues to behave like a company that wants more than rockets — and is willing to buy, partner, or build its way into the center of the AI race.

Key developments at a glance

• Bloomberg reported that SpaceX attempted to acquire AI coding startup Cognition.
• Cognition CEO Scott Wu publicly denied the report and said the company is not for sale.
• SpaceX has already acquired xAI and reportedly bought Cursor for $60 billion.
• Cognition raised $1 billion in late May at a $25 billion valuation.
• Bloomberg says Cognition may now be seeking fresh funding at a $40 billion valuation.
• The two companies could still explore cooperation around compute resources.

Why investors are watching this closely

This story matters to investors because it shows how quickly valuations in AI software are being shaped by strategic demand. A startup with enterprise traction and a hot product can move from a funding story to an acquisition rumor to a potential platform deal in a matter of days.

It also matters because the AI market is entering a consolidation phase. Large buyers want differentiated products, proprietary user bases, and access to scarce infrastructure. Cognition has all three in some measure, which is why it drew attention even if a deal never materialized.

In the end, the report is less about a single abandoned takeover and more about the direction of the market. The companies with the most capital, compute, and ambition are now competing not only to build the best models, but to own the tools that make those models indispensable.