Cognition's CEO Pushes Back on Acquisition Rumors as SpaceX Accelerates AI Push
Scott Wu publicly refutes Bloomberg's report while his startup navigates a valuation climb to $40 billion, even as computing-capacity talks with SpaceX may still be alive.
A Public Denial and a Bigger Context
Scott Wu wasted no time. Within hours of a Wednesday report claiming that SpaceX had approached Cognition about an acquisition, the CEO took to X to set the record straight. "Inaccurate," he wrote. "Cognition is not for sale." He added that the two companies had not been in talks about a deal.
The statement arrived against a backdrop that has made AI coding one of the most contested arenas in enterprise software. At DailyTechWire, we've tracked the velocity of M&A and partnership announcements in this space over the past eighteen months, and the cadence has accelerated sharply since SpaceX's $60 billion acquisition of Cursor closed last week. That deal, combined with SpaceX's earlier purchase of xAI and its subsequent June IPO that briefly pushed market cap near $2.3 trillion, signals an ambition that extends well beyond rockets and satellites. Elon Musk told employees last week that "in about four or five years, AI will be 99% of the value" of SpaceX. Delivering on that projection will require not just models and infrastructure, but sticky, revenue-generating products. AI-assisted coding is proving to be exactly that.
Wu's denial stops short of addressing one wrinkle: according to the same report, Cognition and SpaceX are still discussing a potential partnership in which Cognition would consume SpaceX's excess compute capacity. SpaceX has been leasing GPU cycles to other AI labs, including Anthropic, while it scales its own workloads. Wu did not comment on this portion of the story.
The Strategic Fit That Wasn't
On paper, Cognition would have been a logical addition to SpaceX's portfolio. The startup's flagship product, Devin, is an autonomous coding agent that has found traction with blue-chip enterprise customers including Mercedes-Benz, Citi, and Goldman Sachs. Those relationships represent the kind of enterprise footprint that SpaceX's xAI division, despite its scale and funding, has struggled to build. xAI's chatbot Grok has been dogged by controversy, from last year's "MechaHitler" incident to this year's nonconsensual imagery scandals, making it a harder sell in regulated industries.
Anthropic's growth trajectory, driven largely by Claude Code, has demonstrated that developer tools can generate significant revenue and customer lock-in at speed. SpaceX's acquisition of Cursor was a direct play into that dynamic. This month, Cursor and SpaceX jointly released Grok 4.6, a model that scores higher on coding and multi-step agentic benchmarks. Adding Cognition's enterprise relationships and agent architecture would have deepened that capability.
The fact that the deal did not materialize, or that Wu disputes it ever got off the ground, leaves Cognition as one of the largest independent players in the space. The company raised $1 billion in late May at a $25 billion post-money valuation and is now in early talks for a new round at $40 billion, according to the same report Wu contested. That valuation climb, if it closes, would underscore investor confidence that Cognition can remain independent and competitive without being absorbed by a hyperscaler or a well-funded rival.
A Consolidating Landscape
Cognition's recent history reflects how quickly the AI coding market is consolidating. Last year, the company acquired the remaining assets of competitor Windsurf after Google DeepMind acqui-hired Windsurf's CEO and top researchers in a $2.4 billion deal for talent and licensing rights. Post-merger, Cognition laid off 30 employees and offered buyouts to the remaining 200 Windsurf staff. Those who stayed faced what the company framed as wartime conditions: more than 80-hour workweeks and six days in the office.
That intensity is not unusual in Musk's orbit. He has publicly stated that he works up to 120 hours a week and has slept on factory and office floors. The cultural overlap between Cognition's operating tempo and the expectations inside SpaceX and xAI may explain why the acquisition speculation gained traction in the first place, even if Wu insists no formal talks occurred.
What Comes Next for Cognition
Whether or not SpaceX made a formal bid, the underlying logic of the rumor remains relevant. AI coding is emerging as a category where product-market fit and enterprise adoption are happening faster than in most other AI verticals. Startups that can demonstrate both technical differentiation and customer revenue are attractive targets, and the universe of credible acquirers has expanded beyond the usual suspects. SpaceX, with its public-market valuation and stated AI ambitions, now has the currency and strategic rationale to compete for assets that would have been out of reach two years ago.
For Cognition, the path forward involves sustaining its enterprise momentum while navigating the compute and capital intensity required to train and serve increasingly capable models. If the partnership discussions around SpaceX's compute capacity are real, they offer a pragmatic way to secure GPU access without ceding equity or control. That would allow Cognition to preserve its independence while benefiting from infrastructure that SpaceX is currently underutilizing.
Wu's public stance, "not for sale," is unambiguous. But in a market where billion-dollar outcomes are table stakes and multi-tens-of-billions valuations are in play, declarations of independence are easier to make than to maintain. The next twelve months will clarify whether Cognition can sustain its trajectory as a standalone company or whether the gravitational pull of consolidation proves too strong.
