Apple is preparing to showcase a new AI-powered battery management system at its Worldwide Developers Conference (WWDC) on June 9, 2025, marking one of the company's most practical applications of artificial intelligence to date.
The new feature will analyze how users interact with their devices and make power-saving adjustments on the fly. If implemented effectively, this could provide substantial improvements to battery life across Apple's device lineup.
The battery management system is expected to be particularly important for the rumored "iPhone 17 Air" - an ultra-slim handset that will join Apple's lineup later this year. Without this AI optimization, the phone's smaller battery might struggle to last through a typical day of use. According to Bloomberg's Mark Gurman, the iPhone 17 Air would last several hours less than other iPhones without this feature, but with the AI battery management activated, the svelte device could become much more practical.
The AI-powered system will analyze individual usage patterns to adjust power consumption dynamically. It can limit the power draw of specific apps and system features based on how the device is typically used. Building upon Apple's existing machine learning capabilities seen in features like Optimized Battery Charging (available since iOS 13), this new system takes a more comprehensive approach to power management while avoiding the mixed reception that some generative AI features have received.
iOS 26 will also introduce a lock screen indicator showing estimated time until charge completion, refined from earlier versions. This will be particularly helpful for iPhone 17 Air users, whose smaller battery may present challenges for some users. Importantly, the AI battery management feature will be optional, allowing users to disable it if they prefer manual control over their device's power settings.
In addition to the battery management system, Apple is expected to announce plans to open up its on-device foundation models to third-party developers at WWDC. These are the same ~3B parameter models Apple currently uses for features like text summarization and autocorrect. This represents a meaningful milestone for Apple's AI platform, giving developers powerful tools to natively integrate into their apps and potentially unlock genuinely useful features.