Apple Plans to Replace Core ML with Core AI Framework at WWDC 2026, Signaling a Developer-Facing AI Platform Shift
Apple is reportedly planning to retire Core ML at WWDC 2026 and introduce Core AI, a modernized framework intended to help developers integrate outside AI models into iOS apps, according to Bloomberg's Mark Gurman.
Overview
Apple is reportedly preparing to retire Core ML, the machine learning framework it introduced in 2017, and replace it with a new system called Core AI at its Worldwide Developers Conference in June 2026, according to Bloomberg’s Mark Gurman, as reported by 9to5Mac. The planned transition, expected to accompany iOS 27, would mark a structural change in how Apple exposes artificial intelligence capabilities to third-party developers.
From Machine Learning to Generative AI
Core ML, launched in 2017, was built around traditional machine learning tasks — image classification, natural language processing, sound recognition — and allowed developers to bundle trained models directly into their apps. The framework served its purpose for a generation of on-device ML features, but as the industry pivoted toward large language models and generative pipelines, the tooling began to show its age.
According to Gurman’s Power On newsletter as cited by 9to5Mac, Apple is “planning a few other software-based AI upgrades, including a new framework called Core AI,” with “the idea to replace the long-existing Core ML with something a bit more modern.” Gurman also noted a deliberate nomenclature shift: “The switch from ‘ML’ to ‘AI’ is significant. Apple knows that ‘machine learning’ is a dated term that no longer resonates with developers or consumers. The general purpose of Core AI, though, remains the same: helping developers integrate outside AI models into their apps.”
Developer Implications
The framing of Core AI as a tool for helping developers “integrate outside AI models” suggests Apple intends the framework to serve not just its own on-device models, but to provide a system-level integration point for third-party AI capabilities. iLounge similarly notes the transition addresses the expanded scope of what developers are building, from task-specific model bundles to generative AI workflows. Specific technical details about supported model types, input modalities, or integration protocols have not been reported by primary sources.
Apple’s standard approach of maintaining migration pathways suggests both frameworks may coexist during a transition period, though Core AI is positioned as the primary target for new development going forward.
Broader AI Strategy at WWDC 2026
Core AI is expected to be one of several AI-focused announcements at WWDC 2026. Apple has separately been reported to be working on a significant Siri overhaul that would incorporate Google Gemini models, according to MacRumors, with a redesigned Siri expected for devices running iOS 26.4 and a more advanced version arriving in iOS 27.
Taken together, the reported Core AI framework and the Siri overhaul represent two layers of Apple’s response to the generative AI era: one facing developers, the other facing consumers. Core AI would provide the integration plumbing; a revamped Siri would provide the user-facing surface. Both efforts, if they proceed as described, converge on a broader ambition to embed AI as a platform layer throughout iOS.
Context and Caveats
All current reporting on Core AI traces to a single source: Bloomberg’s Gurman, via his Power On newsletter. Gurman has a strong track record on Apple pre-announcements, but Apple has not officially confirmed any of these plans. WWDC 2026 is expected in June, and the company typically maintains strict embargoes on developer framework details until the keynote.
The reported shift from “ML” to “AI” in the framework name reflects how fundamentally the target use case has changed: from bundling small, task-specific models into apps to providing a system-wide runtime for integrating large, general-purpose AI models. That change, if Apple executes it as described, could give iOS a meaningful edge in the on-device AI application space.