Anthropic's IPO Reveals Heavy Dependence on Amazon and Google
Anthropic's confidential IPO filing reveals that the AI company routed nearly half of its revenue through Amazon and Google's cloud marketplaces in 2025, creating a complex relationship where the tech giants serve as investors, computing suppliers, distribution partners, and direct competitors. The company's revenue surged more than tenfold to $4.6 billion last year while operating losses doubled to $8 billion, with consumption-based fees expected to dominate future revenue streams. The concentration of sales through just two major platforms presents a significant business risk as Anthropic seeks a valuation around $2 trillion.
Anthropic's financial structure reveals a rapidly intensifying partnership dynamic with two dominant technology platforms. The company's revenue trajectory accelerated dramatically during 2025, yet simultaneous operating losses expanded proportionally, reflecting the capital-intensive nature of developing and deploying large language models. The arrangement whereby Amazon and Google simultaneously function as investors, infrastructure providers, sales channels, and competitive rivals creates a web of interdependencies that warrants investor scrutiny.
The magnitude of Anthropic's computing commitments underscores the extraordinary infrastructure requirements underpinning modern AI development. Total long-term spending obligations nearly eight times exceeded annual revenue by mid-2026, locking the company into sustained reliance on its primary partners' capacity and pricing decisions for years ahead.
This concentration of distribution and computing dependencies could reshape competitive dynamics in enterprise AI services, potentially limiting customer choice if Amazon and Google leverage their gatekeeper positions. Investors may face uncertainty regarding whether Anthropic can maintain operational independence or negotiate favorable terms as its reliance deepens. The circular financial relationship—where cloud providers simultaneously profit from distribution fees, computing contracts, and equity stakes—raises questions about long-term pricing sustainability and whether smaller AI competitors might face structural disadvantages in accessing equivalent infrastructure and market channels.