Apple Quietly Outsourced Siri to Gemini. The Honest Read for Indie Apple-Ecosystem Devs.
Apple Quietly Outsourced Siri to Gemini. The Honest Read for Indie Apple-Ecosystem Devs.
Apple's Q2 print landed yesterday with a beat on revenue and a $100B buyback authorization, but the line that buried itself in the analyst commentary was the confirmation that Google's Gemini will power the next Siri. The plan had been telegraphed since January, but yesterday was the financial-disclosure version where Apple effectively committed to it.
Most coverage today will frame this as Apple losing the AI race. That's lazy. The contrarian read — and the one nobody is writing — is that this is unambiguously good news for indie Apple-ecosystem developers. Here's why, and the specific App Intents homework you should do this week to capture the upside.
What Apple actually committed to
The setup is that Apple is licensing Gemini under terms similar to its long-running search-default deal with Google. Gemini does the heavy reasoning. Apple keeps the on-device "personal context" layer (your messages, calendar, photos), the privacy framing, and the App Intents surface that connects Siri to third-party apps.
That last layer is the one indie devs should care about. App Intents is the framework that lets a Siri request like "send Anna the receipt for last night's dinner" route to your accounting app's "send-receipt" intent rather than to a generic web search. In iOS 18 it was a "nice to have" surface. In iOS 26, with a Gemini-powered orchestrator behind it, it becomes a real distribution channel.
The reason to take this seriously: Gemini is good enough at multi-step reasoning that a user request can now reliably get decomposed into the right App Intent calls in the right order. The previous Siri implementation could not do this. With the better orchestrator, the App Intent surface goes from "almost works" to "works." That's a step change for any indie iOS developer who exposes intents.
Why this is good news for indie devs specifically
The conventional read is that Apple losing the foundation-model race means Apple's developers lose. The contrarian read is that Apple losing the foundation-model race means Apple's developers win, because Apple is forced to compete on the layer indie devs actually depend on — the integration surface — rather than the layer they don't.
If Apple had won the foundation-model race, Siri would have been an Apple-only orchestrator with limited extensibility, and the App Intents surface would have been a lower priority. Apple losing forces them to make App Intents excellent because that's where their differentiation now lives. Better integration surface, more direct access to user requests, lower distribution friction for indie apps.
This is the same dynamic that played out with Apple Pay. Apple didn't win the payments-network race. They became the integration layer. That worked out fine for indie merchants who plugged into the integration. App Intents is the same shape of bet, ten years later.
The privacy framing is doing real work
There's a second-order effect worth naming. Apple's privacy framing — on-device personal context, server-side queries scrubbed of identifiers, no training on user data — is now load-bearing for the entire arrangement with Google. If the privacy story breaks, the deal collapses publicly and Apple eats a brand crisis.
That means Apple has a strong incentive to make the privacy boundaries crisp and the developer documentation around what data flows where actually accurate. For indie devs, this is the first time the privacy posture is forced to be unambiguous. You'll get clearer guidance about what Siri can and can't see when it queries your app, and clearer rules about how your App Intent responses are handled.
Practically: read the iOS 26 App Intents documentation when it ships at WWDC, and pay particular attention to the "personal context" boundary section. That section will tell you exactly what data your app can rely on Siri having about the user when it routes a request, and what your app needs to provide explicitly. Most indie devs will skim that section and miss the actionable bits. Don't.
The App Intents homework for this week
Three concrete actions worth doing in the next seven days, before the WWDC noise hits and everyone is competing for the same attention.
First, audit which App Intents your app already exposes. Most indie iOS apps have a handful of intents from earlier years that have rotted slightly — they reference UI flows that have changed, or return error states that are no longer reachable. The next-generation Siri will route real user requests through these. If they're rotten, your app gets routed to and then fails awkwardly. Fix that this week.
Second, write down the five most common things your users do in your app. For each one, ask: is there an App Intent for this? If not, write one. The new orchestrator will make exposed intents discoverable in a way they previously weren't, and the marginal cost of adding an intent is low. Apps that ship five well-designed intents in the iOS 26 timeframe will get disproportionate distribution from Siri routing.
Third, add a small piece of telemetry that logs when an intent is invoked from Siri vs. from another surface. The data you collect over the next six months will tell you which intents are actually pulling traffic and which were a waste of effort. You want this telemetry in place before the iOS 26 rollout, not after, so you have a clean baseline.
Total work: a focused weekend, maybe two evenings if your app is small. The upside is that when iOS 26 ships and Siri actually starts routing requests reliably, your app is ready for the first wave of routing traffic. Most indie apps will not be.
The honest counter-take
The deal is not signed in stone. Apple has historically renegotiated these arrangements when leverage shifts. If OpenAI or Anthropic offers better terms in twelve months, Apple will entertain the switch, and indie devs will end up living through one more orchestrator rewrite. The App Intents surface is durable; the model behind it is not.
The right response is to optimize for the surface, not the model. The intents you expose this week will work the same way regardless of which company's foundation model is making the routing decision. That's the hedge. It's also the reason the App Intents homework is the right answer — it's the part of the stack that doesn't depend on which model wins.
The other counter-take I keep coming back to: Apple-ecosystem indie development has been declining as a focus for solo operators since roughly 2022. The App Store economics, the review-team friction, and the iOS 26 SDK deadline mess have all pushed indie energy toward web-first stacks. This deal doesn't reverse that trend. It does, however, make the case for staying engaged with App Intents specifically — that surface is now the highest-leverage place to spend Apple-ecosystem development time, and it's worth the focused weekend even if the rest of your iOS investment is on autopilot.