Apple watched Amazon, Google, Microsoft, and Meta commit $650 billion to AI infrastructure and said "no thanks, we'll just buy yours."
Instead of building its own AI from scratch, Apple is paying Google $1 billion per year to license a custom 1.2 trillion parameter Gemini model to power a completely redesigned Siri, expected to ship with iOS 27 later this year.
For context, here's what the others are spending on AI data centers in 2026 alone: Amazon at $200 billion, Google at $175-185 billion, Microsoft at $80 billion, and Meta at $65 billion. That's roughly $650 billion combined, most of it going into hardware that depreciates fast.
Apple's bet is that it doesn't need to own the infrastructure. It has 2.2 billion active devices, 30%+ margins, and the most locked-in consumer ecosystem in tech. Plug in someone else's AI, distribute it through your hardware, and collect the margin without the capex.
The risk is dependency. Apple tried building its own AI models internally and it didn't work. That's the whole reason they're writing Google a check. If Google raises the price, changes the terms, or prioritizes its own products, Apple doesn't have a plan B.
The bigger question isn't really about Apple. It's whether the $650 billion AI infrastructure buildout will generate returns that justify the spend, or whether the smartest play was always to let everyone else burn the capital and license the winner.
Instead of building its own AI from scratch, Apple is paying Google $1 billion per year to license a custom 1.2 trillion parameter Gemini model to power a completely redesigned Siri, expected to ship with iOS 27 later this year.
For context, here's what the others are spending on AI data centers in 2026 alone: Amazon at $200 billion, Google at $175-185 billion, Microsoft at $80 billion, and Meta at $65 billion. That's roughly $650 billion combined, most of it going into hardware that depreciates fast.
Apple's bet is that it doesn't need to own the infrastructure. It has 2.2 billion active devices, 30%+ margins, and the most locked-in consumer ecosystem in tech. Plug in someone else's AI, distribute it through your hardware, and collect the margin without the capex.
The risk is dependency. Apple tried building its own AI models internally and it didn't work. That's the whole reason they're writing Google a check. If Google raises the price, changes the terms, or prioritizes its own products, Apple doesn't have a plan B.
The bigger question isn't really about Apple. It's whether the $650 billion AI infrastructure buildout will generate returns that justify the spend, or whether the smartest play was always to let everyone else burn the capital and license the winner.
21❤️2🤙2👍1
