← Back to writing
Blog · 01 / 18MAY 15, 2026BLOG6 min read

Siri 2.0 Is A Tax Disguised As An API

Apple wants every app to expose its features through App Intents so Siri can call them. The API is clean. The unwritten part of the contract is what should make every iOS developer pause.

All blogs →Get in touch

Siri 2.0 Is A Tax Disguised As An API

WWDC 2026 is three weeks out. The leaks are converging on the same story: a Gemini-powered Siri, a standalone Siri app, a Dynamic Island presence, an Extensions system for swapping in third-party models, and — the part that affects every iOS developer — a major expansion of App Intents so Siri can take action inside your app without the user ever opening it.

The technical pitch is straightforward. The economic pitch is where the trap is.

What App Intents Actually Are

If you haven't shipped an App Intent yet, the model is simple: you annotate parts of your app's functionality with declarative metadata — what the action does, what parameters it takes, what entities it operates on — and Siri (or Spotlight, or Shortcuts) can invoke them directly. Order a coffee. Start a workout. Send a message. Tag a photo. The user says it, Siri parses it, your app executes it, the result shows up in the assistant surface.

In iOS 18 and 19 this was a nice-to-have. By iOS 27, it becomes the primary distribution surface for app functionality. The new Siri is being marketed as the interface, not a feature. The apps that expose strong App Intents become visible inside that interface. The ones that don't become invisible.

That's the carrot.

The Stick Nobody's Talking About

Per multiple reports from prominent developers Apple has been briefing, the company has explicitly declined to rule out introducing a commission on Siri integrations in the future. The phrasing is careful. Nothing is being announced. Nothing is being committed to. The door is just being left open.

If you're a developer who lived through the 30% App Store cut, the In-App Purchase mandate, the Reader app fight, the Spotify/Epic litigation, the EU's Digital Markets Act compliance theater, the "Core Technology Fee" that landed in 2024 — you know exactly how this script ends.

The pattern is the same every time:

  1. Apple ships a new platform feature that developers must adopt for distribution
  2. Apple does not charge for it initially
  3. Apple positions the feature as essential for visibility
  4. Two to four years in, a fee structure appears
  5. The fee is framed as reasonable because of the "platform investment"
  6. The fee becomes mandatory because the alternative is invisibility

The map of "things that started as free APIs and became revenue lines" includes: App Search, In-App Purchases, the Apple Sign-In requirement, push notifications via APNs (still free, but the precedent for monetizing infrastructure is well-established), and now — strongly signaled — Siri integrations.

Why Big Apps Are Hesitating

The 9to5Mac piece from May 13 ran the explicit quote: prominent app companies are slow-walking their Siri integrations specifically because of fee uncertainty. Not because the API is hard. The API is fine. The economic model is unknown, and that's enough to freeze investment.

The math from a CFO's chair:

  • Building robust App Intents for a non-trivial app is roughly 2-6 engineering weeks
  • The QA matrix expands because every intent now needs to work in voice, text, contextual, and headless modes
  • The exposure is permanent — once the user has an expectation that "Siri can do X in your app," removing it is a degradation
  • If Apple announces a 5-15% take on revenue attributed to Siri-driven actions in 2027, every line item above becomes a sunk cost in service of a new tax

For a Fortune 500 app, the calculation is "wait for Apple to clarify the economic model." For a solo developer or small team, the calculation should be similar — but the messaging from Apple's developer relations team will be intense, and the FOMO about being absent from Siri's surface will be real.

The Substance Density Test

I wrote about substance density in a previous post: how much of a vendor's marketing claim survives 30 minutes of hands-on testing. Siri 2.0 mostly passes that test on the technical side. The App Intents API is clean. The framework is well-documented. The integration is not the problem.

The substance gap is in the economic surface:

  • Who owns the user? When a user says "order me coffee" and Siri picks the merchant, is the merchant Siri picked your app or your competitor's? On what basis? Discoverable? Auditable? Appealable?
  • Who owns the data? App Intents pass parameters through Apple's runtime. What gets logged? What gets used for model improvement? What's the privacy posture for sensitive verticals — health, finance, identity?
  • Who owns the revenue? This is the unanswered question. The answer "we haven't decided" should be read as "we have decided, but we're not telling you yet."

A clean API surface with an unclear economic surface is always more dangerous than a worse API with clearer economics. Apple's developer ecosystem has been here a half-dozen times.

What I Recommend

Three positions, depending on what you're building:

1. If you're shipping a free or ad-supported app, build App Intents. The risk surface is low (no revenue for Apple to take a percentage of), and the discoverability benefit is real. Treat Siri integration as SEO for the next decade of iOS.

2. If you're shipping a paid app or one with significant IAP revenue, build a minimal App Intents surface — enough to be present, not enough to become economically dependent. The "what would it cost to retreat" calculation matters. Don't build flows that only exist inside Siri.

3. If you're shipping in a regulated vertical (health, finance, kids, identity), wait until WWDC, then wait again for the privacy and data-handling specs to settle, then wait again for the first wave of compliance precedent. The cost of moving slowly here is lower than the cost of being the test case.

The cross-cutting advice for everyone: keep your app's core flows workable without Siri. The day Apple announces the commission, you don't want to be the team whose product is unrecoverable from the new economic reality.

The Bigger Pattern

Apple's most successful platform moves have always been about taking a category of work developers do for free and progressively monetizing it. Storefront. Payments. Search. Push at scale. The App Store rules. The Core Technology Fee.

The "natural language is the new UI" pitch is the largest such opportunity in Apple's history. Every action in every app becomes a billable surface area if the assistant layer is the front door. The Gemini partnership and the Extensions system tell you Apple knows this — they're already structuring the market so that they're the platform on which other AI vendors compete, with Apple collecting rent from both sides.

Siri 2.0 is genuinely impressive technology. It's also a beachhead. The API is the welcome mat. What gets billed once you're inside is the open question, and "we haven't ruled it out" is not a reassurance.

Build the integration. Don't build a business that depends on it being free.

Sources:

Enjoyed this read?

Get the next one in your inbox →