Platform Taxes and Rent-Seeking

Platform taxes are the fees, commissions, and revenue shares that dominant platforms extract from economic activity conducted on their networks. Rent-seeking is the broader economic concept: extracting value from a position of control without creating proportional new value. In the technology economy, the two are deeply intertwined — platforms that achieve monopoly or near-monopoly positions in distribution can impose tolls that shape entire industries, redirecting entrepreneurial energy and determining which businesses are viable.

The 30% Standard and Its Discontents

The canonical platform tax is Apple's 30% commission on App Store purchases and in-app transactions — a rate that has become a de facto standard across mobile platforms. Google's Play Store charges the same. Valve's Steam takes 30% of PC game revenue (with tiered reductions at high volumes). These fees are not set by competitive markets — they're set by gatekeepers who control access to users.

The economic impact is profound. A 30% tax on revenue means that businesses operating on these platforms must either achieve margins above 30% (difficult for most), pass costs to consumers, or accept dramatically lower profitability than they would in an open market. For games and interactive experiences, where content creation costs are high and margins are already thin, the platform tax can be the difference between a sustainable studio and one that slowly bleeds out.

Epic Games' legal battle with Apple over Fortnite's removal from the App Store was fundamentally a fight over rent-seeking. Epic's argument: Apple provides distribution infrastructure (valuable, but not 30%-of-all-revenue valuable) and uses its hardware monopoly to prevent sideloading and alternative app stores. Apple's counterargument: curation, security, and the trust environment justify the fee. The courts partially sided with Apple, but the EU's Digital Markets Act and similar legislation worldwide have begun forcing open alternative distribution channels — a regulatory response to what economists recognize as monopoly rent extraction.

Rent-Seeking vs. Value Creation

Classical rent-seeking theory, developed by economists Gordon Tullock and Anne Krueger, describes how actors invest resources to capture existing wealth rather than create new wealth. The key insight: rent-seeking is not just the extraction itself but the distortion it creates. When platforms extract outsized fees, developers build for platforms with lower taxes (the agentic web over mobile apps), design around payment restrictions (linking to external payment flows), or avoid building entire categories of products that can't survive the tax.

Unity's 2023 per-install runtime fee debacle illustrated how quickly rent-seeking can backfire. By attempting to insert a new tax layer into an ecosystem that had already optimized around existing pricing, Unity triggered a developer revolt that forced a full reversal and ultimately contributed to leadership change. The lesson: platform taxes that exceed what participants perceive as proportional to value delivered generate intense resistance — and in an economy with alternatives, migration.

Not all platform fees are rent-seeking. Valve's Steam takes 30%, but developers voluntarily choose Steam because its discovery, community, update infrastructure, and install base genuinely drive sales. The test is whether the fee is commensurate with the value the platform creates — or whether it's extractable only because the platform controls a chokepoint.

Platform Taxes in the Agentic Economy

The agentic economy is restructuring where platform taxes can be imposed. As AI agents increasingly mediate transactions — discovering, evaluating, and purchasing on behalf of users — the traditional app store chokepoint becomes less relevant, but new chokepoints emerge. Who controls the agent's default tools? Which inference providers get priority? What commission does a platform charge when an AI agent books a service on a user's behalf?

Cloud providers impose their own platform taxes: markup on compute, egress fees that make leaving expensive, and proprietary service integrations that create switching costs. NVIDIA's CUDA ecosystem is arguably a platform tax on the entire AI industry — the software moat means competitors' chips must be dramatically better, not just comparable, to win adoption.

The open-source and open-weight movements in AI can be understood partly as resistance to platform taxes. Meta's decision to release Llama as open-weight isn't altruism — it's a strategic move to prevent any single API provider from becoming a rent-collecting gatekeeper over AI inference. NVIDIA's $26 billion commitment to open-weight models follows the same logic: widely adopted open models create demand for NVIDIA hardware without letting a closed-model provider capture the toll position between NVIDIA's silicon and end users.

The agentic economy will inevitably produce new platform taxes. The question is whether the architecture — open protocols, interoperable agents, competitive inference markets — creates enough competition at each layer to keep rents closer to the value actually delivered, or whether new monopoly chokepoints emerge that extract Gilded Age margins from the AI revolution.