PLATFORM ECONOMICS

A platform business intermediates between two or more user groups — buyers and sellers, developers and users, advertisers and audiences — and creates value by reducing transaction costs and enabling interactions that wouldn't otherwise occur at scale.

Platform vs Pipeline

Traditional product businesses (pipelines) move value in one direction: inputs in, outputs out. Platforms move value in multiple directions simultaneously. A manufacturer builds a product and sells it; a marketplace connects buyers and sellers without owning the inventory. The critical structural difference is that platforms benefit from cross-side network effects — more sellers attract more buyers, which attracts more sellers — while pipelines generally don't. This is why platform businesses such as Apple's App Store, Amazon Marketplace, and Google Search have generated some of the highest returns on capital in business history: they can grow total value without proportionally growing costs, because each additional participant makes the network more valuable for everyone else.

The Two-Sided Challenge

Platforms must solve a chicken-and-egg problem at launch. No buyers without sellers, no sellers without buyers. The strategies that work — subsidizing one side, seeding supply, focusing on a narrow geography or category first — are well-documented. What's less discussed is that platforms face a governance problem at scale: the rules governing participant behavior become as important as the technology. How a platform handles fraud, quality, and bad actors determines whether participants trust it enough to transact. This governance layer is a genuine competitive advantage, not just operational overhead.

AI Platforms

The current wave of AI development is creating new platform dynamics. Foundation model providers — OpenAI, Anthropic, Google — function as platform intermediaries between AI capability and application developers. The strategic question is whether AI application developers will face the same dependency and margin compression that mobile app developers faced with iOS and Android. If models commoditize quickly, the platform layer shifts to whoever controls distribution, fine-tuning infrastructure, or proprietary data. Building a defensible position — a genuine economic moat — on top of commodity model infrastructure is the central strategic challenge for AI product companies today.

Platform Power and Regulatory Pressure

The largest platform businesses have attracted antitrust scrutiny precisely because their structural advantages are so durable. The EU's Digital Markets Act and US DOJ actions against Google and Apple reflect a view that self-preferencing, data advantages, and ecosystem lock-in constitute anticompetitive behavior when the platform also competes in adjacent markets. For technology strategists, this regulatory dynamic is material: it creates both risk — forced structural changes, interoperability requirements — and opportunity, as regulated access to data or distribution opens markets that incumbents would otherwise deny. Platforms designed with some degree of openness are more defensible politically than pure walled gardens, and those whose moats derive primarily from access control rather than genuine product superiority face the most structural exposure.

Open Questions

• Will AI foundation model providers consolidate into a small number of dominant platforms, or will open-source alternatives prevent that concentration?
• As regulators impose interoperability requirements, do platforms lose their network effect advantages or simply shift them to a different layer?
• Can a platform sustain trust with all sides of the market while also competing directly against participants — or does that conflict always resolve badly?

Part of the knowledge graph at The Best Blog Ever — reference definitions for ideas that matter.