Apple is considering daily usage limits for its Siri AI features due to on-device compute constraints, according to sources familiar with the discussions [1]. The potential caps would restrict the number of AI-powered interactions per device per day — a significant departure from Apple's marketing of unlimited on-device intelligence.

The constraint stems from Apple's own privacy architecture. By processing AI requests on-device rather than in the cloud, Apple cannot distribute compute load across its data centers [2]. Each device must handle its own inference, and the processing ceiling is fixed by the hardware in the user's hand.

Bloomberg framed the potential limits as a phased rollout — a standard approach for new technology. X users read the constraint as evidence that Apple's on-device model fundamentally cannot scale. "The privacy-first model has a compute ceiling. Apple's marketing promised unlimited AI. The hardware can't deliver it," one widely shared post argued [3].

The usage caps would apply to Apple Intelligence features introduced at WWDC, including the system-wide interaction layer and advanced Siri capabilities. Basic Siri functions — timers, weather, simple queries — would remain unaffected.

Apple declined to comment on the potential limits. The company has not disclosed what the daily cap would be or when any limits might take effect.

The constraint explains Apple's reported interest in the Google Gemini partnership — cloud compute capacity that on-device inference cannot provide.

-- KENJI NAKAMURA, Tokyo