At the 2026 Worldwide Developers Conference, Apple laid out a vision for how to integrate AI into its products that was strikingly calm, responsible, and relevant.
In contrast to the job-killing, security-destroying, human-replacing hype spread by the likes of Anthropic and OpenAI, company executives toned down their usual superlative-filled spiel to convey how AI tools can actually help software developers and those using Apple products.
Features like Safari’s Notify Me (website change notification) and a low-code extension creation service for browsers called Describe an Extension look like solid uses for machine learning technology.
Some of Cupertino’s more subdued marketing may be due to the company eating crow as a result of poor AI performance. But it also coincides with a lack of enthusiasm in three areas the company is focused on: improving its platform, strengthening child safety, and Apple Intelligence.
Platform improvements like 30% faster app launches, 70% faster photo loading, and a more efficient CPU scheduler aren’t features that marketing departments know what to do with, even if they result in noticeable improvements to the user experience. And while Child Safety is welcomed by some and is politically expedient at the moment, it’s fundamentally about restricting the use of Apple products rather than expanding their use.
As a result, Apple Intelligence has not been fully successful since its introduction in 2024.
“Apple is rebuilding from the ground up to make AI native, useful, and invisible across the devices people already use every day,” Francisco Geronimo, VP of client devices at IDC, said in an email. register.
“This is important because a winning AI experience for consumers won’t be the loudest or most technically complex; it will be the one that understands context, respects privacy, works reliably across apps, and reduces friction without forcing users to change their behavior.”
Much of the developer keynote focused on improvements to Siri (now rebranded as Siri AI), which will be generally available when v27 of Apple’s various platforms is released this fall. Apple developers now have access to better versions of these releases.
But beyond claims that Siri is fit for purpose, the Apple executives in attendance managed to highlight the company’s substantial advantages in terms of privacy, integration, and cost. And they made a good pitch about developing AI applications on the Apple platform and using the Swift programming language to do so.
“Many AI providers talk about privacy today, but by default, most AI providers keep your interactions private and you are responsible for protecting your privacy,” explained Craig Federighi, Apple’s senior vice president of software engineering. “The same goes for using temporary chats, deleting conversations, or even turning off features entirely. At Apple, we believe privacy in AI is non-negotiable.”
Apple has over-promised privacy in the past, describing it as a human right and treating it as a government-granted perk, but its AI privacy story centered around private cloud computing was compelling enough to prompt Google to emulate it.
Anyone developing applications using AI tools must consider data security and data privacy. Cloud-based AI models can easily capture sensitive data. Apple is giving developers the ability to use Google’s Gemini model family and the new multimodal-based Foundation Models framework on-device or in private cloud computing, with the ability to integrate with cloud-based model providers and custom models if needed.
Moreover, it’s done in a way that respects the realities of software development. Not all developers can risk connecting their apps to expensive AI APIs (e.g. Claude or Codex) that can result in AI charges that exceed the app’s revenue.
That’s why Apple is making the Foundation Model framework available on Private Cloud Compute, with no cloud API cost, for developers who haven’t yet made it.
“Developers with fewer than 2 million initial App Store downloads will now be able to use Apple Foundation Models running on Private Cloud Compute with no cloud API cost,” said Joshua Shafer, Apple’s senior director of software, during Apple’s Platforms State of the Union presentation. “This is access to frontier-level intelligence with unparalleled privacy protections, because infrastructure costs shouldn’t prevent ideas from being considered.”
or by infrastructure barriers. One of Apple’s advantages is that it has control over both hardware and software. And the company is leveraging its technology stack to solve contextual problems. AI models perform better when they have access to contextual information. Developers may not be able to provide AI services with enough useful information because that information is typically siled by application boundaries, permissions, and other types of controls.
Apple announced both enhancements to existing technologies and new technologies to make contextual information more accessible to AI models and improve AI-oriented development.
For example, Spotlight, Apple’s on-device search indexing service, has been rewritten to be less problematic. It has a long history of unstable service, requiring users to remove and re-add their storage devices to trigger a re-index. In addition, Spotlight is integrated with Siri in hopes of making the service more effective at searching for files and displaying relevant data within apps to inform AI queries.
Apple’s Xcode 27 includes a variety of improvements, but the most notable changes were introduced in February, with Xcode 26.3 adding support for Anthropic’s Claude Agent and OpenAI’s Codex. This list has now expanded to include Google’s Gemini and agent customizations. Xcode can be quite difficult and complex for those who aren’t seasoned Apple platform developers, so the integration of the IDE with an AI coding agent is a meaningful improvement. Being able to ask an AI agent to identify small faults in your configuration is a welcome change.
The App Intents framework has been expanded to help developers better leverage Siri AI capabilities through understanding personal context, accessing app actions, and on-screen activity.
There’s also a new Core AI framework, which, as Apple says, “enables you to load, specialize, and run AI models completely on-device with a modern, memory-safe Swift API, preserving user data privacy and app responsiveness, with zero server dependencies or token costs.”
If frontier model leaders like Anthropic and OpenAI continue to raise prices, Apple’s local model story could become even more compelling. ®
