Apple announced a fundamental redesign of its Apple Intelligence platform, moving away from a purely proprietary approach to embrace Google's Gemini AI models in a deep collaboration that promises to deliver state-of-the-art reasoning, image understanding, and multimodal capabilities across iPhones, iPads, and Macs. The shift marks a significant pivot for the company as it races to catch up with competitors in the AI race while maintaining its signature privacy-first positioning.

What Is Apple's New AI Architecture and How Does It Work?

At the heart of Apple's revamped strategy sits a new system orchestrator, a central coordinator that manages Apple Intelligence features securely across all Apple platforms. This orchestrator tailors responses based on which app you're using and what task you're performing, enabling what Apple describes as "truly system-wide intelligence". The architecture processes requests both on your device and through Apple's Private Cloud Compute infrastructure, a hybrid approach the company says keeps user data inaccessible to Apple itself or third parties.

The upgraded models, developed jointly with Google, support capabilities that were previously unavailable on Apple devices. These include realistic image creation, advanced photo editing with spatial reframing that lets you adjust how a photo is framed after it was taken, and visual question answering that can analyze what's on your screen. Certain higher-powered device versions will also gain speech generation, improved dictation accuracy, and stronger natural language understanding, though Apple has not yet specified which devices qualify.

How Does the Redesigned Siri Compare to Previous Versions?

The new Siri AI represents perhaps the most dramatic overhaul of Apple's voice assistant since its 2011 debut. Apple software chief Craig Federighi emphasized that the company believes "truly helpful AI should be centered around you and your needs," contrasting Apple's approach with competitors he characterized as "pursuing AI for the sake of AI" without clear regard for users. The redesigned Siri can now handle complex, multi-step tasks that require understanding context from your device and personal information.

Craig Federighi

For example, the new Siri can create a menu by gathering recipes from the web and your own text messages, then invite friends from a group chat for a viewing party. It can also analyze images on your screen, such as determining whether a backpack will fit as a carry-on for a flight or whether hiking boots will fit inside luggage. A standalone Siri AI app will launch later this year, though it will not initially be available in Europe due to regulatory issues, and Apple is still working out compliance requirements for China.

How Are Developers Getting Access to These New AI Capabilities?

Apple is providing developers with new tools to integrate AI into their apps more easily. Xcode 27, Apple's development environment, now includes expanded agentic coding capabilities that bring models and agents from Anthropic, Google, and OpenAI directly into a developer's workflow. The updated Foundation Models framework serves as a single native Swift API that supports on-device models with image input, server models, and the ability to build custom skills.

Developers can tap into the next generation of Apple Foundation Models, which were custom-built in collaboration with Google and its Gemini models. For small developers, Apple is offering a significant incentive: those enrolled in the App Store Small Business Program with fewer than 2 million total first-time App Store downloads can access the next generation of Apple Foundation Models running on Private Cloud Compute at no cloud API cost. Developers also have the flexibility to leverage models from other providers like Claude and Gemini, or any provider that implements the new language model protocol.

Ways Developers Can Build AI Features Into Their Apps

• Foundation Models Framework: A native Swift API that supports powerful on-device models with image input, server models, and custom skills, allowing developers to integrate Apple's next-generation models or choose from providers like Claude and Gemini.
• Core AI Framework: A brand-new framework designed specifically for running custom models on device, optimized for the unified memory and Neural Engine of Apple silicon, enabling developers to deploy full-scale large language models locally.
• App Intents Framework Updates: Enhanced capabilities that connect apps to Siri AI features including personal context understanding, app actions, and onscreen awareness, helping users discover and access app content more easily.
• Dynamic Profiles: A new capability that enables developers to update how AI models interact with their apps on the fly, making it easier to create adaptive AI experiences.

Susan Prescott, Apple's vice president of Worldwide Developer Relations, stated that "developers are at the heart of the Apple ecosystem, and our goal is to provide them with the best possible tools and technologies to build the future". She emphasized that Xcode 27's new intelligence frameworks and agentic coding capabilities give developers the tools they need to focus on bringing their ideas to life.

Susan Prescott, Apple's vice president of Worldwide Developer Relations

Beyond AI, Xcode 27 itself has been significantly improved. The application is now Apple silicon only, 30 percent smaller, and delivers faster performance with a simpler setup process. Xcode Cloud, Apple's cloud-based build service, is now up to 2 times faster and includes new support for apps using Metal graphics and visionOS builds.

Why Is Apple Partnering With Google Instead of Building Everything Alone?

The collaboration with Google represents a pragmatic acknowledgment that Apple cannot match the scale and resources Google has invested in large language models. By integrating Gemini models into its architecture, Apple gains access to state-of-the-art AI capabilities without the massive research and development costs of building competing models from scratch. This approach allows Apple to focus on what it does best: integrating AI seamlessly into user-facing products while maintaining strict privacy controls.

Apple has framed its approach as fundamentally different from competitors. The company reiterated that Apple Intelligence relies on on-device processing and Private Cloud Compute, with a promise that user data is only used to execute the immediate request and is not accessible to Apple or third parties. Apple added that outside experts can verify those privacy guarantees "at any time".

This announcement comes at a pivotal moment for Apple. The company has faced criticism for stumbling in its efforts to deliver new AI features that were promised nearly two years ago. CEO Tim Cook, who is stepping down in September after 15 years leading the company, received an extended standing ovation at the World Wide Developers Conference and told the audience he is "deeply grateful to have been on this journey with you". The transition to new CEO John Ternus arrives as artificial intelligence has unleashed the most upheaval within the tech industry since the original iPhone was unveiled in 2007.

The scale of Apple's developer ecosystem makes this announcement significant. The conference drew developers from approximately 65 countries to Apple's Silicon Valley headquarters, all focused on software development for Apple platforms. With these new tools and capabilities, Apple is signaling that it intends to make AI development as accessible and integrated into the Apple experience as possible, whether developers choose to use Apple's own models or bring their own.