breaking March 2, 2026

Core AI: Apple Replaces Core ML in Its WWDC 2026 Overhaul

Apple will replace Core ML with a new Core AI framework at WWDC 2026, marking a fundamental shift in how iOS 27 developers build AI-powered apps.

Core AI , Apple’s incoming replacement for Core ML , will debut at WWDC 2026, marking the most significant overhaul to the company’s on-device AI developer infrastructure in nearly a decade. Bloomberg’s Mark Gurman first reported the shift in his Power On newsletter on Sunday, revealing that the transition will ship with iOS 27. The move retires a framework Apple introduced in 2017 , built for a narrow machine learning world that no longer reflects how AI is developed and deployed in 2026.

The announcement arrives as Apple attempts to close a visible capability gap against rivals who have moved faster. Google has been shipping on-device AI through Gemini Nano. Microsoft has embedded Copilot into Windows at the system level. Meta has pushed open-weight models that any developer can run locally. For Apple, Core AI is the infrastructure answer to that competitive pressure , a statement that on-device AI will be a true first-class citizen in the Apple ecosystem going forward, not a late addition bolted onto an aging framework.

From “Machine Learning” to “AI”: Why the Naming Shift Matters

The move from Core ML to Core AI is more than a rebrand. Core ML, introduced in 2017, was built around a narrow definition of machine learning , classifying images, running regression models, deploying pre-trained weights on-device. It predates the large language model era and was never designed to handle the open-ended, generative, and agentic AI behaviors that developers now expect to ship.

Core AI, according to Gurman, is architected with modern AI in mind from the ground up. Its primary job is enabling developers to integrate both Apple’s own Foundation Models , the on-device models behind Apple Intelligence , and third-party AI systems into their apps, without requiring developers to build inference pipelines from scratch.

“Apple knows that ‘machine learning’ is a dated term that no longer resonates with developers or consumers,” Gurman wrote. The new name explicitly aligns Apple’s developer toolchain with the vocabulary and expectations of 2026.

What Core AI Means for Developers and Users

For app developers, Core AI is expected to dramatically reduce the friction of shipping AI features. Currently, integrating a third-party model , whether a locally-quantized open-source LLM or a cloud-based API , requires stitching together multiple frameworks, managing token contexts, and handling model input/output manually.

Core AI is expected to abstract much of that complexity. Notably, Gurman indicates that MCP , the Model Context Protocol, which has become a de-facto standard for connecting AI models to external tools and data sources , is a likely integration vector. If Apple ships native MCP support inside Core AI, it would be a significant legitimization of the protocol and an accelerant for the agentic app ecosystem on Apple platforms.

For end users, the practical impact will arrive through apps: smarter text summarization, more capable in-app assistants, context-aware suggestions, and richer automation , all running with Apple’s signature emphasis on on-device privacy.

The Broader Apple Intelligence Picture

Core AI arrives in the context of a turbulent stretch for Apple Intelligence. The company has faced criticism for slow rollout, limited Siri capabilities, and the delayed integration of third-party AI models. Apple is reportedly pushing a Gemini-powered Siri upgrade to iOS 26.5, having already delayed the feature past iOS 26.4, which is currently in developer beta.

iOS 27, the target platform for Core AI, is expected to carry a rebuilt Siri experience and a broader reimagining of how AI operates across the OS , not as a bolted-on assistant but as a system-level layer that apps can call directly. Core AI is the developer surface for that layer.

WWDC 2026 is shaping up to be one of the densest Apple developer conferences in recent memory, with Core AI joining previously rumored features including a new Siri chatbot interface and expanded Apple Intelligence APIs. The June event is now confirmed to be a major inflection point for the platform.

The move also positions Apple in direct competition with Google’s on-device AI initiatives and the expanding AI platform strategies being pursued by every major technology company right now. With Core AI, Apple is telling developers: build your AI stack on our platform, with our tooling, under our privacy guarantees. The stakes of that pitch have never been higher, and the wave of AI investment reshaping the industry is making it increasingly urgent that Apple closes the gap.

What Comes Next

WWDC 2026 begins June 9. Between now and then, Apple is expected to release iOS 26.4 to the public and begin seeding early iOS 26.5 betas. Developers curious about Core AI will likely get their first official technical preview through WWDC sessions and updated documentation, followed by a developer beta in the weeks after the keynote.

See also: Jack Dorsey Just Fired the Starting Gun on AI Layoffs | X01.

For related context, see OpenAI Closes 110B Round at 730 Billion Valuation.

Core AI is expected to abstract much of that complexity. Notably, Gurman indicates that MCP , the Model Context Protocol, which has become a de-facto standard for connecting AI models to external tools and data sources , is a likely integration vector. If Apple ships native MCP support inside Core AI, it would be a significant legitimization of the protocol and an accelerant for the agentic app ecosystem on Apple platforms.

For end users, the practical impact will arrive through apps: smarter text summarization, more capable in-app assistants, context-aware suggestions, and richer automation , all running with Apple’s signature emphasis on on-device privacy.

The Broader Apple Intelligence Picture

Core AI arrives in the context of a turbulent stretch for Apple Intelligence. The company has faced criticism for slow rollout, limited Siri capabilities, and the delayed integration of third-party AI models. Apple is reportedly pushing a Gemini-powered Siri upgrade to iOS 26.5, having already delayed the feature past iOS 26.4, which is currently in developer beta.

iOS 27, the target platform for Core AI, is expected to carry a rebuilt Siri experience and a broader reimagining of how AI operates across the OS , not as a bolted-on assistant but as a system-level layer that apps can call directly. Core AI is the developer surface for that layer.

WWDC 2026 is shaping up to be one of the densest Apple developer conferences in recent memory, with Core AI joining previously rumored features including a new Siri chatbot interface and expanded Apple Intelligence APIs. The June event is now confirmed to be a major inflection point for the platform.

The move also positions Apple in direct competition with Google’s on-device AI initiatives and the expanding AI platform strategies being pursued by every major technology company right now. With Core AI, Apple is telling developers: build your AI stack on our platform, with our tooling, under our privacy guarantees. The stakes of that pitch have never been higher, and the wave of AI investment reshaping the industry is making it increasingly urgent that Apple closes the gap.

What Comes Next

WWDC 2026 begins June 9. Between now and then, Apple is expected to release iOS 26.4 to the public and begin seeding early iOS 26.5 betas. Developers curious about Core AI will likely get their first official technical preview through WWDC sessions and updated documentation, followed by a developer beta in the weeks after the keynote.

The critical question will be how much of Core AI ships at launch versus being staged across the iOS 27 release cycle. Apple has a history of introducing frameworks that mature slowly , Core ML itself took multiple OS cycles before it supported the model types developers actually wanted. Whether Core AI launches with full third-party model support, native MCP integration, and a polished developer experience, or whether it arrives as a structural foundation to be incrementally expanded over the next year, will define whether it lands as a genuine platform shift or a promise deferred.

For now, the signal from Gurman is clear: Apple has decided that the machine learning era is over. The Core AI era begins this June.