Gallery

Contacts

405 W. Greenlawn Ave Lansing, Michigan 48910

contact@techjacksolutions.com

+1-616-320-4064

Skip to content
Technology Daily Brief Vendor Claim

Apple Licenses Google Gemini for Siri at ~$1B/Year: What the WWDC 2026 Deal Means for Developers

3 min read Bloomberg Partial Moderate A
Apple has agreed to license a customized 1.2-trillion-parameter Google Gemini model to power cloud-based Siri features at approximately $1 billion per year, according to Bloomberg, making this the first time Apple has contracted a competitor's frontier model as the primary cloud reasoning engine for a core OS feature. Queries run through Apple's Private Cloud Compute infrastructure, which Apple says prevents Google from receiving user-identifying data.
Annual Gemini licensing, ~$1B

Key Takeaways

  • Apple is licensing a custom 1.2T-parameter Google Gemini model to power cloud Siri features at ~$1B/year, per Bloomberg, the first time Apple has used a competitor's frontier model for a core OS feature
  • Queries route through Apple's Private Cloud Compute, which Apple says prevents Google from accessing user-identifying data, this is Apple's stated design intent, not independently audited for this deployment
  • Apple is reportedly introducing an iOS 27 Extensions system allowing third-party AI integrations (ChatGPT, Claude, Gemini); Gemini holds the default slot, source: Bloomberg
  • The 1.2T parameter figure is Bloomberg-reported for a custom, unreleased model with no public benchmarks or independent evaluation available

Apple opened WWDC 2026 in Cupertino on June 8 with an announcement that redraws the boundary between building and buying in frontier AI. Apple’s developer platform now sits on top of a licensed Google model, a 1.2-trillion-parameter customized Gemini build, for cloud Siri reasoning, with licensing fees reported by Bloomberg at approximately $1 billion per year.

That number deserves a moment. A billion dollars annually to use a competitor’s model is not a stopgap. It’s a strategic posture.

The architecture matters for anyone evaluating this through a compliance or product lens. Cloud queries to Gemini route through Apple’s Private Cloud Compute, a documented Apple infrastructure designed for private AI processing. According to Apple, the design means Google doesn’t receive user-identifying data during those queries. That’s Apple’s stated intent for PCC, it hasn’t been independently audited for this specific Gemini deployment. Don’t treat the privacy guarantee as more than a vendor design claim until third-party evaluation exists.

The catch is the Extensions system. Apple is reportedly introducing an Extensions setting within Apple Intelligence on iOS 27 that would let users configure third-party AI integrations – ChatGPT, Claude, and Gemini among them, according to Bloomberg. That’s a significant developer surface. It means every major AI lab now has a formal path into Apple’s install base. Gemini holds the default slot. OpenAI and Anthropic are in the Extensions marketplace. The difference between default and optional is not trivial at billion-device scale.

For developers, WWDC 2026 is a session week, not just a keynote. Apple’s developer releases confirm iOS 27 betas are going out to developers. What the technical sessions clarify about Extensions API access, capability tiers, and data routing will matter more than the headline number.

What to watch

The Extensions API documentation. If Apple gates capability levels, giving Gemini features that Claude and ChatGPT don’t get, that’s a competitive moat inside Apple’s own platform. If the Extensions system is genuinely parity, it’s the most consequential developer distribution opportunity since the App Store. Watch what Apple publishes at the developer session level, not just the keynote.

For enterprise IT teams, the data governance question is already live. Gemini is now inside the Siri query path on Apple devices. The PCC routing is real and documented. Whether that satisfies your organization’s data residency or AI usage policies depends on your specific requirements, not Apple’s press release. Compliance teams running Apple device fleets should be mapping this against their AI governance frameworks now, before iOS 27 ships to production.

The part nobody mentions: the 1.2 trillion parameter figure is Bloomberg’s reporting on a custom model. It’s not a publicly released model with public benchmarks. There’s no Epoch AI evaluation. No SWE-Bench score. No independent throughput data. When you see “1.2T parameters” in coverage this week, that’s a Bloomberg-reported figure on a proprietary, unreleased build. Treat capability claims accordingly.

Apple conceded it couldn’t compete at the frontier model layer. Google got $1B/year in licensing revenue and the default AI slot on a billion devices. Both parties got what they needed. The question for every other AI lab is whether a spot in the Extensions marketplace is worth the terms Apple sets.

View Source
More Technology intelligence
View all Technology

Related Coverage

Stay ahead on Technology

Get verified AI intelligence delivered daily. No hype, no speculation, just what matters.

Explore the AI News Hub