Apple has restructured its artificial intelligence strategy under software chief Craig Federighi, accelerating plans to overhaul Siri by relying on external AI models after years of internal delays and organizational friction.
According to a detailed report from The Information, Apple's approach to artificial intelligence has undergone a significant shift over the past year. Apple software chief Craig Federighi is said to be at the center of that shift, having assumed direct oversight of the company's AI organization and is now driving decisions that will shape the future of Siri and other Apple Intelligence features across the product lineup.
Last fall, Federighi apparently addressed a joint meeting of Apple's software and AI teams, expressing enthusiasm for closer collaboration while also signaling dissatisfaction with the company's pace of progress in artificial intelligence. Some members of Apple's foundation models team interpreted the remarks as criticism of their work.
In December, Apple moved to consolidate its AI leadership under Federighi, completing a transition that had begun earlier in the year when responsibility for Siri was removed from the AI group and brought under Federighi's software division. In January, Apple announced plans to use Google's Gemini AI models to power future AI upgrades, including an improved version of Siri. In Federighi's view, integrating a third-party model would allow Apple to finally ship a revamped Siri later this year after controversially postponing the update in 2025.
However, the report also outlines internal concerns about the implications of placing AI under Federighi's control. People who have worked closely with him described him as highly cost-conscious and skeptical of investments with uncertain returns. This approach stands in notable contrast to rivals such as OpenAI, Meta Platforms, and Google, who invest tens of billions of dollars in data centers, chips, and AI researchers.
Apple has attempted to limit infrastructure spending by emphasizing on-device processing and its Private Cloud Compute system, which uses Apple silicon. The company was said to be waiting for the cost of AI computation and talent to decline, betting that most consumer use cases will eventually be handled locally on devices.
Federighi apparently viewed AI as unpredictable and difficult to control, preferring deterministic software behavior that could be clearly specified during design reviews. He rejected proposals to use AI to dynamically reorganize the iPhone home screen, arguing that such changes would confuse users.
Tensions over AI strategy have surfaced internally before. Around 2019, Mike Rockwell, who was leading development of the Vision Pro headset, reportedly proposed an AI-driven interface. He criticized Federighi's software approach as overly conservative, prompting a rebuke. Rockwell was later placed in charge of Siri in early 2025 and now reports directly to Federighi.
Despite his earlier skepticism, Federighi's stance shifted following the release of ChatGPT in late 2022. People close to him said he became convinced of the potential of large language models after experimenting with the technology and instructed his teams to explore ways to integrate similar capabilities into Apple products. Federighi reportedly concluded that Apple's internal models did not perform adequately on devices, while members of the foundation models team believed they were being blamed for challenges related to model optimization, which fell under the software organization's responsibilities.
Some team members complained they were not given sufficient guidance on how their models would ultimately be used, limiting their ability to compete with external alternatives. Around the time Apple removed Siri oversight from Giannandrea and assigned it to Rockwell, with Federighi directing the broader effort, Federighi instructed teams to evaluate deep integration of third-party models.
Despite the partnership with Google, Apple plans to continue developing its own AI models, particularly those designed to run on devices. Apple reportedly intends to shrink and adapt models derived from external partners so they can run more fully on Apple hardware, reducing long-term dependence. To support that goal, Apple is said to be considering acquisitions of smaller AI firms specializing in model compression and optimization.
See The Information's full report for more.