Amazon’s $18B Double Bet: Why They’re Backing Both Anthropic and OpenAI

Amazon has invested $8 billion in Anthropic. Now they’re in talks to invest $10 billion+ in OpenAI at a $500 billion valuation. The same company is backing both sides of the frontier model — as explored in the intelligence factory race between AI labs — race. This isn’t hedging – it’s a masterclass in platform economics.

The Capital Structure of AI

The deal crystallizes AI’s new reality: foundation model companies need so much money that competitors become investors. Amazon funds OpenAI while backing Anthropic. Microsoft backs both too. The AI race costs too much for clean competitive lines.

This is unprecedented in technology history. Imagine if during the browser wars, Microsoft had invested in both Netscape and Internet Explorer. Or if during mobile, Apple had funded Android. The capital requirements of frontier AI have created strange bedfellows.

The Anthropic Investment Logic

Amazon’s $8B Anthropic investment is a masterclass in platform economics:

Model Leadership: Claude’s capabilities often match or exceed GPT-4 in enterprise use cases

Safety Alignment: Anthropic’s Constitutional AI appeals to risk-averse enterprises

Acquisition Option: Amazon’s investment provides right of first refusal

No Governance Headaches: Minority investor with no board seat, unlike Microsoft’s complex OpenAI arrangement

Why OpenAI Too?

If you’re Amazon, why invest in both? Because AWS is the distribution layer. Every model that runs on AWS generates infrastructure — as explored in the economics of AI compute infrastructure — revenue. Whether Anthropic wins or OpenAI wins, Amazon wins – as long as they run on Trainium and serve through AWS.

This is the “commoditize the complements” strategy applied to AI itself. Let others compete on models while you capture the infrastructure layer beneath all of them.

Key Takeaway

As the AI Value Chain shows, the platform layer captures value regardless of which application wins. Amazon isn’t betting on a model – they’re betting on being the platform all models need.

Source: Amazon’s AI Superstructure on The Business Engineer