Free as an Economic Model

In the AI Story

Each of Anderson's four free models is already visible in the AI market. Freemium dominates: ChatGPT and Claude offer limited free tiers and paid frontier access, with conversion rates calibrated in real time. Cross-subsidy drives the major platforms: Google offers Gemini to preserve ecosystem lock-in, Microsoft offers Copilot to sell Azure, Apple integrates AI to sell hardware. Advertising has not yet colonized AI tools but the economics predict it will. Gift economy animates the open-source community — Meta's Llama, the open-weight alternatives that compete with proprietary offerings through reputation and ecosystem effects.

The trajectory is unambiguous. The marginal cost of AI inference is declining with each hardware generation and each efficiency improvement. Competitive pressure increases with each new entrant. The price of frontier capability, now one hundred dollars per month, will fall toward the marginal cost of inference, which is approaching zero. Segal predicts this explicitly in The Orange Pill when he writes that frontier capability will soon become 'dirt cheap.'

The incumbent pain is severe and consistent with every previous abundance transition. The Software Death Cross is the music industry's Napster moment, compressed into months rather than years. The trillion dollars of market value that vanished from SaaS companies in early 2026 is the digital-distribution-era version of recorded music revenue collapse.

Value migrates to the adjacent scarce layers: data (which accumulates with use and cannot be instantly reproduced), trust (which requires institutional time to build), and integration (which depends on network effects no individual creator can replicate). The product becomes the loss leader; the adjacent layer becomes the business.

Origin

Anderson published Free: The Future of a Radical Price in 2009, extending the economic logic of The Long Tail into pricing theory. The book was controversial — Malcolm Gladwell's New Yorker review was particularly skeptical — but the predictions have been substantially borne out. Music, journalism, software, and now AI capability have followed the trajectory Anderson described.

The theoretical foundation is classical microeconomics: in competitive markets, price converges on marginal cost. What Anderson contributed was the recognition that digital goods have marginal costs so close to zero that the convergence produces free products as an equilibrium rather than an anomaly.

Key Ideas

The price trajectory is structural. Competitive dynamics in digital markets produce free products regardless of what producers want.

Four monetization models. Freemium, advertising, cross-subsidy, and gift economy are the only sustainable paths when the core product is free.

Adjacent layers capture value. When the product is abundant, value migrates to data, trust, integration, and curation — the scarce complements that remain.

Incumbents suffer brutally. The transition from paid to free destroys business models calibrated to scarcity; the pain is the price of the democratization it enables.

The floor matters. The speed at which AI capability reaches universal affordability determines when the developer in Lagos can fully participate in the long tail of creation.