ISSUE #252
Apple Just Paid $1B to Break Its Own Promise
The privacy company wired Siri into Google. The real lesson isn't build-vs-buy — it's the price tag nobody reads.
For two years, Apple promised a smarter Siri and showed us roadmap slides. Yesterday at WWDC it finally shipped the rebuild — "Siri AI," conversational, context-aware, living in its own app. The catch: the brain isn't Apple's. It's Google's. Apple is reportedly paying ~$1 billion a year to run Apple Intelligence on Google's Gemini models.
What happens when you throw out the GTM playbook
That investor was wrong. Gamma is now worth $2B, with 50M users and more than half their growth driven by word of mouth.
Each one took a GTM risk most founders would never greenlight. Each one paid off.
TLDR: Apple — the "what happens on your iPhone stays on your iPhone" company — just rented its AI brain from Google. The takeaway for you isn't build vs. buy (that debate's over). It's that every AI tool carries a second price tag — data, dependency, lock-in — and almost nobody reads it before signing. Below: a prompt that reads it for you.
Apple bought the smartest brain in the room — and a new landlord
Here's what actually changed. The old Siri, when you asked it something hard, would shrug and offer to "ask ChatGPT" — a visible handoff to an outside company. The new Siri doesn't do that anymore. It doesn't need to, because the intelligence is now baked in through Apple Foundation Models that Apple built in collaboration with Google's Gemini. The top tier, "AFM Cloud Pro," even runs on Nvidia chips inside Google's cloud.
On the same stage, Apple's Craig Federighi insisted "privacy in AI is non-negotiable." Both things are true at once — and that's the lesson. Apple didn't sell you out. It made a calculated trade: it gave up a piece of its independence (a dependency on a direct competitor, a billion-dollar yearly bill, a brain it doesn't fully control) to get an assistant that finally works.
You make the same trade every time you adopt an AI tool — you just don't see the invoice. The free transcription tool that trains on your calls. The writing assistant that owns your prompt history. The agent platform that makes leaving a six-month migration project. The sticker price is rarely the real price. If the most disciplined company on earth takes on a dependency this big to get good AI, the move isn't to avoid the trade — it's to know exactly what you're trading before you sign.
So here's a prompt that runs the math on any tool you're about to adopt.
The Prompt (Copy This)
It interviews you about the AI tool you're considering — what it touches, what it stores, how hard it'd be to leave — then hands you a one-page verdict: the real cost beyond the monthly fee, your single biggest exposure, and whether to sign, negotiate, or walk. Apple-grade due diligence, in about three minutes.
Prompt Proof Table
| Who's running it | Tool they're weighing | Hidden price the audit surfaces | Verdict |
| Agency owner | Free AI notetaker on client calls | Vendor trains on recorded client calls — your NDAs may not cover it | Walk |
| SaaS founder | Agent platform for support | Workflows lock to their format — ~5-month rebuild to switch later | Negotiate exit terms |
| Solo consultant | Paid writing assistant | Low exposure — no client data, easy export, cancel anytime | Sign |
| Clinic manager | AI intake chatbot | Patient data crosses into a model with unclear HIPAA posture | Walk |
| Same prompt. YOUR tool, YOUR data. Read the second price tag before you sign. Try it → |
Apple spent fifteen years and untold ad dollars convincing you it would never let anyone else near your data. Then it found a brain worth a billion a year — and made the trade in front of forty million viewers. The question isn't whether you'll make the same kind of trade. You already are. The only question is whether you read the fine print first.