Apple Bets on Privacy and On-Device AI as Rivals Race Ahead

[madmax]

As rivals funnel billions into massive AI data centers, Apple remains committed to its privacy-first strategy, running most AI tasks directly on users’ devices via Apple Silicon.
When cloud processing is needed, Apple uses Private Cloud Compute, designed to temporarily process and then delete data.

But this philosophy has consequences:
• Smaller on-device models = less capable AI
• Minimal data collection = fewer real-world training sets
• Heavy reliance on licensed and synthetic data

The result? Apple is now reportedly leaning on Google’s Gemini to power the next version of Siri — a remarkable twist in a rivalry that spans more than 15 years.

Subramanya now faces Apple’s biggest question in a decade:
Can Apple catch up in AI — or has its privacy-centric approach slowed it down beyond recovery?

[Nahid Hossen]

Apple doubling down on on-device AI proves privacy is still their biggest selling point.