The arrival of Anthropic’s S-1 prospectus, as reported by Reuters on September 28, 2026, signals an ambition that defies conventional market logic. While the document has not yet appeared on the SEC’s EDGAR database—a delay that has kept the market waiting since the confidential filing on June 1—the details circulating describe a company aiming for a valuation exceeding $2 trillion. This is not merely a capital raise; it is a declaration of a new industrial order where the cost of entry is measured in half-trillion-dollar increments.
The true narrative of this filing is not the revenue, which reached approximately $4.59 billion in 2025, but the staggering $518 billion in planned cloud, compute, and infrastructure obligations. As detailed in the breakdown, this includes roughly $111.1 billion to Google, $110 billion to Amazon, $31.4 billion to Microsoft, and $161.2 billion to Broadcom. Approximately 80% of these obligations are binding and non-cancelable. To put this in perspective, for every dollar Anthropic earned in 2025, it committed roughly $113 to future infrastructure. This is the terminal expression of the compute landlord thesis: a company that does not own its own foundation but is instead locked into a massive, non-cancelable web of dependencies.
This structure creates a profound dependency. Anthropic is effectively a tenant in a digital landscape owned by the very entities that provide its capital and infrastructure. The scale of these commitments suggests that the company’s survival is tethered to the continued expansion of these specific hardware and cloud providers, creating a feedback loop where the pursuit of artificial intelligence capabilities necessitates an ever-increasing transfer of wealth to the providers of the underlying compute.
Governance within this entity is equally singular. The filing outlines a Founder LLC structure that grants seven co-founders—including CEO Dario Amodei, President and Board Chair Daniela Amodei, Chief Compute Officer Tom Brown, and Chris Olah—50.1% voting power through a Class F share. At a $2 trillion scale, such concentrated control is unprecedented. This mechanism is designed to persist until only two or fewer founders remain, effectively insulating the company’s strategic direction from the standard pressures of public market shareholders. Furthermore, the co-founders have pledged 80% of their equity to charity, with the Long-Term Benefit Trust overseen by figures including former Federal Reserve Chair Ben Bernanke and national security expert Richard Fontaine. It is a legal architecture built to prioritize the founders’ vision of safety, even when that vision directly conflicts with the financial interests of those providing the capital.
The document itself is a study in contradictions. Of its 261 pages, roughly 80 are dedicated to risk factors, compared to only 48 pages describing the actual business operations. The disclosures detail existential risks to humanity, including concerns over self-preserving behaviors, shutdown resistance, and potential blackmail by the models themselves. There is a jarring dissonance in a company presenting a $2 trillion valuation while simultaneously documenting the possibility that its core product could pose a catastrophic threat to the species.
This tension is further complicated by the legal environment. On September 25, the D.C. Circuit Court of Appeals upheld a Pentagon classification of Anthropic as a supply chain risk under FASCSSA Section 4713. As explored in our analysis of the D.C. Circuit ruling, the Founder LLC structure, intended to codify safety as a competitive advantage, now exists in a reality where the state classifies those same safety restrictions as a liability to national supply chain security. The company is attempting to market itself as a safe harbor while the judiciary views its operational constraints as a systemic vulnerability.
The structural picture that emerges is one of extreme centralization. We see a company that is simultaneously a massive consumer of capital, a tenant of the largest cloud providers, and a self-governing entity that reserves the right to ignore its own investors in the name of existential risk mitigation. It is a model that demands total faith in the founders’ judgment, even as the legal system begins to categorize that judgment as a risk factor in its own right.
The ecosystem is hardening rapidly. With reports of NVIDIA considering a $10 billion anchor investment and the recent formation of the SAFA body on September 27, the industry is attempting to build guardrails around a runaway train. If the IPO proceeds, the market will be forced to decide whether the promise of artificial general intelligence justifies a corporate structure that treats its own investors as secondary to its existential mission. The ultimate consequence of this valuation is not just a test of market appetite, but a stress test for the viability of private governance in the face of national security mandates.
Note: This analysis is based on reports regarding the S-1 prospectus, as the document has not yet been made public on SEC EDGAR. Reports of NVIDIA’s potential $10 billion anchor investment remain unconfirmed. The $518 billion infrastructure breakdown and compute figures are based on the prospectus as viewed by Reuters.