Apple Explores Outside AI Partners as Siri Overhaul Faces Delays
Bloomberg reports that Apple is in early-stage discussions with Anthropic and OpenAI about deploying their AI models on Apple's infrastructure, marking a potential shift away from the company's homegrown approach to powering Siri.
To strengthen Siri's artificial intelligence capabilities, Apple is exploring partnerships with external AI providers rather than relying solely on internal development. A Bloomberg report indicates that the Cupertino firm has initiated preliminary conversations with both Anthropic and OpenAI regarding the possibility of running their models on Apple's Private Cloud Compute infrastructure.
Such a move would represent a significant departure for Apple, which currently powers its Apple Intelligence features using proprietary Apple Foundation Models. The company has reportedly discussed with both firms the prospect of adapting their models for deployment across Apple's cloud systems. However, Apple has not yet committed to this direction; the organization continues developing an alternative version of Siri that would rely on its own models. An AI-enhanced iteration of Siri is slated to arrive in 2026, following postponements that pushed back the originally planned release alongside iOS 18, iPadOS 18, and macOS 15.
Apple already integrates OpenAI technology into certain Siri functions, specifically for handling search queries and producing generative content. These capabilities operate as optional add-ons that users must explicitly activate, and Apple retains control over the assistant itself. Adopting a model from Anthropic or OpenAI could position Apple more competitively in the AI landscape and align with strategies employed by other device manufacturers. Samsung, for instance, offers Galaxy AI features while also incorporating Google's Gemini as a core component of its interface. Reports have also surfaced suggesting Apple has explored the possibility of acquiring Perplexity outright.
Apple introduced the iPhone 16 last year with the marketing emphasis of being "built for Apple Intelligence," yet the company has since struggled to deliver on its AI ambitions. Leadership changes have contributed to this challenge: responsibility for Siri shifted from AI executive John Giannandrea to Craig Federighi, who leads software engineering, and Mike Rockwell, Siri's leader who previously oversaw the Vision Pro launch. According to Bloomberg, the prospect of transitioning to an external AI model is affecting morale within Apple's AI division at a moment when competing technology firms are actively recruiting talent and offering premium compensation packages.
The implications extend beyond the iPhone. Apple Intelligence and Siri functionality span the iPad and Mac product lines as well. Giannandrea's reassignment also meant losing oversight of Apple's robotics efforts. The delays affecting Siri have additionally postponed development of a new device category—a hybrid combining HomePod and iPad capabilities designed to function as a home assistant comparable to an Amazon Echo Show.
Apple's internally developed models are not disappearing from the company's roadmap. During WWDC, Apple announced plans to make its on-device models—the foundation for features like Genmoji and Writing Tools—available to developers this year, enabling third-party creators to build AI functionality using Apple's framework. These local models offer privacy advantages but operate at reduced speeds and have delivered less impressive results compared to cloud-based alternatives from competitors. Apple does maintain servers for certain cloud-dependent requests. Deploying a third-party model across Apple's infrastructure could potentially deliver the advantages of both approaches: maintaining user privacy while simultaneously providing the performance gains necessary to compete in the broader AI marketplace.
Source: Tom's Hardware