Amazon, Google, and Microsoft are each building complete AI stacks, but they’re placing fundamentally different bets on where value will accrue. Each company leads in different layers, reflecting distinct theories about how the AI economy will unfold.
These bets are not just strategic preferences — they are mutually exclusive theories of value. If models commoditize, Amazon wins. If modelquality remains decisive, Google wins. If distribution captures value, Microsoft wins. The next 24 months will reveal which theory is correct.
Amazon: “Models will commoditize. Value accrues to infrastructure.” Amazon’s thesis is that foundation models will become interchangeable commodities — like databases or compute instances today. If true, the winners are those who control the orchestration layer, the governance frameworks, and the underlying silicon.
Google: “Model quality is decisive. Own the intelligence, own everything.”Google believes that AI is fundamentally different from previous technology waves. The model itself — its reasoning capability, its knowledge, its reliability — is the product. If models don’t commoditize, whoever builds the best brain wins.
Microsoft: “Distribution is everything. Own the workflow, capture the value.”Microsoft’s thesis is that AI value will accrue at the application layer — specifically, where work actually happens. Models are commoditizing (they’ll just buy the best one). Infrastructure is commoditizing (Azure is good enough). But 400M+ commercial seats in M365 is the moat.
The $10 Trillion Question
Does AI value accrue to infrastructure (Amazon), modelquality (Google), or application integration (Microsoft)?
Each company has built a presence across all six layers of the AI stack, but none dominates everywhere:
Amazon: Full stack, 69% average completeness. Leadership in 1 of 6 layers (Agent Infrastructure). The bet is that this single layer — the orchestration and governance layer — is the one that matters most as agents proliferate.
Google: Most vertically integrated, with leadership in 3 of 6 layers. But agent infrastructure remains light on enterprise governance — a potential gap as regulated industries adopt AI agents.
Microsoft: Strongest at the application layer, leading in 2 of 6 layers. But dependency on the OpenAI partnership and weak silicon position creates strategic vulnerability.
Gennaro is the creator of FourWeekMBA, which reached about four million business people, comprising C-level executives, investors, analysts, product managers, and aspiring digital entrepreneurs in 2022 alone | He is also Director of Sales for a high-tech scaleup in the AI Industry | In 2012, Gennaro earned an International MBA with emphasis on Corporate Finance and Business Strategy.
