SpaceX recently approached artificial intelligence coding startup Cognition AI regarding a potential acquisition, according to a report from Bloomberg citing people familiar with the matter. The exploratory outreach would have marked a major consolidation in the autonomous software engineering sector.
Cognition co-founder and Chief Executive Officer Scott Wu responded publicly that the company is not for sale and that formal acquisition negotiations are not underway. However, sources indicated that the two companies continue to explore commercial and infrastructure collaborations, including the potential for Cognition to utilize SpaceX compute resources to train and operate its models.
Valuation Trajectory and Autonomous Engineering Scale
Cognition AI, the creator of the Devin autonomous software engineer, has expanded rapidly over the past year. In May 2026, the startup secured more than $1 billion in fresh capital at a $26 billion valuation in a funding round co-led by Lux Capital, General Catalyst, and 8VC, with participation from Peter Thiel's Founders Fund, Ribbit Capital, and Atreides Management.
The company's valuation has surged in tandem with its aggressive product expansion. Following its acquisition of coding assistant startup Windsurf in late 2025 for roughly $400 million, Cognition integrated multi-file editing, AST repo mapping, and autonomous agent loops into a unified developer platform.
The startup positions itself as model-agnostic software infrastructure for enterprise engineering teams, avoiding exclusive alignment with any single hyperscaler or foundation model provider.
Aerospace Compute and Vertical Integration
The interest from SpaceX highlights growing intersections between aerospace engineering, advanced industrial automation, and agentic AI systems. Autonomous coding agents are increasingly deployed to maintain mission-critical flight software, automate regression testing suites, and accelerate firmware development cycles.
While an outright acquisition has been rebuffed, discussions around sharing dedicated high-performance computing clusters reflect the severe compute constraints faced by software intelligence startups scaling reinforcement learning and test-time search loops.
