How long is Anthropic’s lease with SpaceX? Opinions vary.
Original reporting by TechCrunch
Earlier this month, the AI sector observed a major compute deal unfold: Elon Musk's xAI reportedly pledged billions of dollars monthly to Anthropic for exclusive access to its advanced Colossus cluster. Hailed as a strategic coup for both companies, the agreement was presented as providing xAI with much-needed revenue and helping Anthropic advance in the perpetual race for compute power. This high-profile partnership was widely seen as a significant move, poised to reshape the competitive landscape.
Contradictory claims
However, the definitive nature of this colossal agreement has now been called into question. In a recent post on X, Elon Musk significantly downplayed the scale and duration of SpaceX’s commitment, suggesting a far more transient arrangement. Musk asserted that SpaceX had not committed to a multi-year lease, instead describing it as a 180-day agreement with a 90-day mutual cancellation clause. He even hinted at potentially reclaiming the compute capacity if resources became tight. This statement directly contradicts SpaceX’s recent S-1 filing, which explicitly details Anthropic agreeing to monthly fees "through May 2029"—a clear three-year term. The filing repeatedly confirms this duration, raising serious questions about the discrepancy and transparency surrounding such a significant financial and strategic move, particularly during a quiet period for the company.
The apparent discrepancy between Elon Musk's public comments on X and SpaceX's recent S-1 filing casts a significant shadow over what was initially hailed as a mutually beneficial compute deal between xAI and Anthropic. While Musk characterized the Colossus lease as a short-term, 180-day arrangement initiated by xAI, even hinting at a potential early recall, the regulatory filing clearly details a three-year commitment from Anthropic to pay substantial monthly fees through May 2029. This direct contradiction, especially regarding a foundational strategic partnership and made during a quiet period surrounding a securities offering, raises serious questions beyond the immediate transaction. It complicates the perception of a deal previously framed as a "coup" for both parties, introducing an element of doubt regarding its true nature and stability.
Broader Implications of Transparency
At its core, this situation underscores a persistent challenge in the rapidly evolving AI landscape: the reliability of information from key industry figures versus official corporate disclosures. Such a material difference in describing a foundational partnership, particularly one concerning crucial compute resources, can significantly erode investor confidence. This erosion isn't limited to just the specific companies involved; it also impacts the perceived transparency and stability of the AI sector as a whole. For Anthropic, whose financial stability and competitive positioning were seemingly bolstered by this long-term revenue stream, the ambiguity introduces unwelcome uncertainty regarding future cash flow and strategic planning. Looking ahead, this incident could prompt increased scrutiny from regulators and investors alike, demanding greater clarity in how significant compute deals—the lifeblood of modern AI development—are communicated. It further highlights the fragility of strategic alliances built on a foundation of potentially shifting public narratives, potentially making future collaborations in the intensely competitive AI compute race more cautious and meticulously documented, with a greater emphasis on ironclad terms rather than public statements.