As I write this, the European Commission has just launched an antitrust investigation into Meta. The charge: restricting third-party AI chatbots from WhatsApp while keeping its own Meta AI fully integrated. Small AI startups like Poke and Luzia had complained that Meta’s new policy — effective January 15, 2026— would block them from reaching customers through the platform. Meta’s defence? “The emergence of AI chatbots on our Business API puts a strain on our systems that they were not designed to support.”

If this sounds familiar, it should. Platform achieves dominance. Platform integrates its own service. Platform restricts competitors from accessing the same distribution channel. Apple did it with the App Store and payment systems. Google did it with Play Store. Now Meta is doing it with WhatsApp and AI chatbots.

The script doesn’t change. The actors do. Here’s the pattern I keep noticing. A new computing paradigm emerges. A platform achieves dominance. The platform starts extracting rent from everyone who depends on it. Developers and regulators fight back. The fight takes a decade. By the time anyone wins, the next platform is already building the next walled garden.

Microsoft bundled Internet Explorer with Windows. The antitrust case dragged on until 2001. By then, mobile was coming. Facebook invited developers onto its platform, let Zynga build a billion-dollar business on viral mechanics, then systematically killed those mechanics once it had captured the users. Google offered “don’t be evil” while constructing an advertising monopoly now facing antitrust action on three separate fronts.

A calculated assault

Tim Higgins’s iWar is the definitive chronicle of the previous iteration: Apple’s App Store versus everyone paying the 30 per cent tax to access a billion iPhone users. Excellent journalism — that fundamentally misunderstands the problem it’s documenting.

The narrative centres on Tim Sweeney’s calculated assault on Apple’s walled garden. In 2020, his “Project Liberty” bypassed Apple’s payment system, got Fortnite kicked off the App Store, and triggered a legal battle lasting five years and costing over a billion dollars.

Higgins reconstructs this with impressive access. Internal emails show Google panicking, offering Epic $208 million to prevent “contagion.” Court documents reveal Phil Schiller warning colleagues their workarounds wouldn’t fly with the judge— ignored. The book builds to the April 2025 contempt ruling: Apple had “willfully” defied orders, one executive “outright lied under oath,” Cook was called out by name. A federal judge referred Apple executives for possible criminal prosecution — the system saying “you lied to us.”

Cook’s response to the contempt finding? “We’re going to appeal.” That’s the whole strategy. Delay, appeal, delay, appeal. The system rewards it.

Since publication, the dominoes keep falling. The Ninth Circuit denied Apple’s stay. Google lost its appeal and settled. Australia ruled against both companies — the first contested application of their 2017 competition law reforms. And in October, OpenAI announced its own app store, with Altman confirming monetisation would be “part of the platform’s long-term model.” They’ve watched this whole fight and learned —probably how to extract without triggering the same backlash.

The cycle is already beginning again. Meta’s WhatsApp move is just the latest iteration.

Higgins frames all this as interpersonal drama — Schiller, the true believer, defending Jobs’s legacy, Cook, the cautious operator, Sweeney the iconoclast. Great characters. Compelling narrative. But swap out every executive in this story and you’d get the same outcome.

Collective irrationality

Thomas Schelling wrote about this in Micromotives and Macrobehavior back in 1978. His insight was that individual rationality often produces collective irrationality. Everyone making sensible choices for themselves, and yet the aggregate outcome is something nobody wanted.

His famous example was segregation — no individual needs to be especially racist for neighbourhoods to end up completely segregated; mild preferences for similar neighbours, compounded across thousands of decisions, produce extreme separation. The micro motives don’t match the macro behaviour.

Platform economics works the same way. Your average product manager at Google isn’t sitting there thinking “let’s build a monopoly.”

She’s thinking “I need this app on my platform to hit my growth target.” Apple’s finance team calculating the 27 per cent external commission wasn’t executing a conspiracy — they were minimising revenue impact while technically complying with a court order.

The lawyer drafting the anti-steering provisions was probably just trying to close a loophole someone identified in a review meeting. Each decision makes sense in isolation. Each person is just doing their job. Nobody is coordinating. Nobody needs to. The incentive structure does the coordinating for them.

Thousands of individual actors, each locally optimising, producing a global outcome that none of them specifically intended. Extraction doesn’t require a villain — just a system where every rational micro-decision nudges toward the same macro-equilibrium.

The 30 per cent commission didn’t emerge from some smoke-filled room; it emerged because that’s where the incentives naturally settle when you control the only door into a billion-user market.

Multi-sided marketplaces with asymmetric power do this reliably. The platform sits in the middle, both sides need it more than it needs any individual participant, and the screws tighten incrementally because each turn is locally rational. By the time anyone notices the pattern, it’s already the industry standard.

Google and Apple arrived at 30 per cent independently. So did Valve’s Steam. So did the console makers. Convergent evolution toward maximum extraction — not because of collusion, but because the game theory points there.

By the time you’ve proven Apple violated antitrust law, a thousand product managers at ten other platforms have already made the same locally-rational decisions producing the same globally-suboptimal outcomes. Punishing Tim Cook doesn’t change the feedback loop that created the behaviour.

We keep bringing 20th century tools to 21st century fights — tools designed to catch robber barons, not to rewire systems where the villain is the structure itself. There’s also something genuinely new here that the book documents without fully grasping.

Video game weaponised

When Apple booted Fortnite, Epic didn’t just file a lawsuit — it launched #FreeFortnite, an in-game event encouraging players to spam Apple’s social media, flood regulatory hearings, and treat the corporate dispute as a personal grievance. They created the “Tart Tycoon” skin — an apple-headed villain in a suit — so players could literally wear their contempt for Apple while playing. Millions of teenagers became an unpaid lobbying army, their parasocial relationship with a video game weaponised for litigation support.

And the book never mentions that Epic is also a platform. Unreal Engine powers half the games industry. Epic takes a cut. Sweeney fighting for the right to be the extractor rather than the extracted. The “rebel” framing obscures that this is really a fight between platforms at different layers of the stack.

The book ends with Sweeney hiking through an open forest after his legal victory, free of walled gardens. Poetic, but hollow. He’s celebrating while OpenAI is building its app store and Meta is restricting AI chatbots from WhatsApp.

iWar is essential reading for anyone involved in platform policy — not for its answers, but for its detailed documentation of what the current system produces. The journalism is excellent. The framework is missing. Higgins describes the symptoms exhaustively while misdiagnosing the disease.

Book Details

You can find the book here.

 (The reviewer is a product designer and researcher working in the tech sector)

Published on December 16, 2025