Apple is finalizing a deal to leverage Google’s 1.2-trillion-parameter Gemini model to power the much-awaited Siri upgrade.
The thing about Apple’s famously self-reliant ecosystem? Sometimes it needs to call in the neighbors for help. And by “help,” I mean a cool $1 billion per year to rent Google’s brain.
For context, that’s roughly eight times more complex than Apple’s current 150-billion-parameter models. Eight times. Let that sink. And remember, every time Siri confidently misunderstood your simple request.
There’s an irony hidden here.
This is the same company that built its brand on doing things “the Apple way.” Privacy and control. The same company that’s been promising us a smarter Siri since, well, since Siri became a punchline. And now? They’re running Google’s models behind the scenes while marketing it all as Apple technology.
Honestly, Apple did test alternatives- OpenAI’s ChatGPT and Anthropic’s Claude before choosing Google. But here’s where it gets interesting: what reportedly tipped the scales wasn’t superior performance, but price. Nothing says “innovation leader” quite like shopping for the budget AI option.
Apple insists this is temporary.
The company claims it’s developing its own 1-trillion-parameter model that could be ready as soon as 2026. Sound familiar? It’s the same playbook they used with maps, weather data, and chips. Lean on someone else until you catch up.
Except this time, there’s no guarantee users will embrace the new Siri or that it can undo years of damage to the brand.
The planned spring 2026 launch gives Apple just enough time to slap its design language on Google’s tech and call it revolutionary. Gemini will handle the summarizing and planning functions- the parts that actually require intelligence.
Meanwhile, Apple already pays Google around $20 billion annually to be the default iPhone search engine. Now add another billion for AI assistance.
At this rate, Google isn’t just inside Apple’s walled garden. It’s paying the mortgage.
