Apple has confirmed it will pay Google roughly $1 billion per year to use Gemini as the backbone for some of Siri’s long-delayed AI upgrades. It’s a striking move for a company that prides itself on doing things in-house – and a tacit admission that Apple is badly behind the pack on AI.
Under the new multi-year agreement, Apple will base its next generation of Apple Foundation Models on Google’s Gemini technology, helping power a more personalized Siri later this year. Apple says everything will still run on device or via its Private Cloud Compute system, preserving its “industry-leading” privacy standards.
The deal inevitably draws comparisons to Apple’s long-running search agreement with Google, which reportedly earns Apple around $20 billion a year (and has recently been under scrutiny in the courts). Once again, Apple appears willing to lean on a rival’s technology when the alternative is falling further behind.
That urgency is easy to understand. Over the past year, Siri has been comically hopeless at times – a far cry from the confident promises Apple made when Apple Intelligence was unveiled back in 2024.
Apple frames Gemini as the best compromise between capability and control. The company says it evaluated multiple options and chose Google because it could deliver cutting-edge models while still allowing Apple to enforce strict privacy protections through custom, largely on-device deployments. That focus on privacy has long been one of Apple’s biggest roadblocks in AI – and one of the reasons it’s lagging today.
OpenAI was reportedly considered, and Siri already hands off to ChatGPT when it runs out of answers. But according to reports, OpenAI either declined the offer, or Apple simply preferred Google’s track record. We may never know which.
For now, Apple insists this is an interim solution while it works on much larger in-house models, potentially ready as soon as next year. We’ll believe that when we see it.