Apple makes a huge bet on AI models becoming commodities
Lafargue Raphael/ABACA/Reuters
- Apple is integrating Google's Gemini AI into Siri, signaling a shift in AI strategy.
- Apple's approach allows Siri to swap AI providers for flexibility and cost savings.
- This move positions Apple to focus on user experience and distribution over AI model development.
Apple's decision to rebuild Siri around Google's Gemini AI models looks, at first glance, like an admission of failure. After years of promising breakthroughs, Apple is reportedly paying Google roughly $1 billion a year to keep its digital assistant relevant.
Look closer, though, and the move represents something more radical: a giant bet that AI models will become commodities.
Top tech reporter Mark Gurman wrote this week that Apple's revamped Siri, codenamed Campos, will launch later this year as a full-fledged chatbot embedded across iPhones, iPads, and Macs. The underlying intelligence will come from Google's Gemini.
The more important detail is architectural. Apple is designing Campos so that the underlying AI models can be swapped out over time, according to Gurman's report for Bloomberg.
"That means the company will have the flexibility to move away from Google-powered systems in the future if it so chooses," he wrote.
That design choice tells us a lot about Apple's AI strategy.
Rather than trying to win an AI arms race against Google, OpenAI, Anthropic, and Meta by spending tens of billions of dollars a year on AI infrastructure, Apple appears to be positioning itself as model-agnostic. Siri remains Apple's direct interface to more than 1 billion iPhone users. The "brains" underneath, however, can change.
If Apple is right, the leading AI models will converge in quality over time, becoming roughly interchangeable. At that point, differentiation shifts away from the model itself and toward distribution, integration, privacy controls, and user experience — areas where Apple already excels.
Apple could then choose whichever AI provider is best, cheapest, or most strategically useful every few years. So, Google's Gemini might power Siri today, but tomorrow it could be OpenAI's latest GPT offering, or Claude from Anthropic, or whatever Meta is cooking up next, or xAI's Grok, or Mistral's offerings, or even region-specific models such as DeepSeek or Alibaba's Qwen in China. Services like OpenRouter already demonstrate how models can be swapped in and out.
Apple is also playing a longer game. Its search deal with Google, worth an estimated $20 billion a year, shows how valuable iPhone distribution is. If AI models truly commoditize, Apple could eventually flip the Siri relationship the same way: AI model providers could end up paying Apple for access to its massive user base.
The strategy also saves Apple enormous capital expenditure. While rivals pour huge sums of money into AI data centers and chips, Apple can let others shoulder that cost. For instance, Google spent more than $60 billion in capex during the first three quarters of 2025, while Apple spent $12.7 billion in its latest fiscal year.
If this AI building boom turns about to be a bust, Apple may dodge a bullet. Of course, it might turn out that underinvesting in your own AI capabilities is a strategic blunder of epic proportions.
We'll see in the next few years. Right now, Apple just placed a huge bet that the real power in the AI era won't belong to model makers, but to whoever controls the interface.
Sign up for Business Insider's Tech Memo newsletter here. Reach out to me via email at abarr@businessinsider.com.