Apple is reportedly exploring using artificial intelligence technology from Anthropic or OpenAI to power a new version of Siri, potentially abandoning its own in-house models in what would represent a significant strategic shift for the iPhone maker.

The company has held discussions with both AI firms about using their large language models for Siri, requesting custom versions that could run on Apple's cloud infrastructure for testing, according to people familiar with the matter, Bloomberg said.

After multiple rounds of testing overseen by Siri chief Mike Rockwell and software engineering head Craig Federighi, Apple executives concluded that Anthropic's technology shows the most promise for Siri's requirements, according to the report. This led to discussions between Apple's vice president of corporate development Adrian Perica and Anthropic about using Claude technology.

The move would mark a dramatic reversal for Apple, which currently powers most of its AI features with homegrown technology called Apple Foundation Models. The company had been planning a new version of its voice assistant running on that technology for 2026.

A switch to external AI models would acknowledge Apple's struggles to compete in generative artificial intelligence, the most important technological development in decades. The Siri assistant, originally released in 2011, has fallen behind popular AI chatbots despite Apple's attempts to upgrade the software.

Apple's investigation into third-party models remains at an early stage, with no final decision made. A competing internal project dubbed "LLM Siri" using in-house models continues active development.

The potential shift has created uncertainty within Apple's roughly 100-person foundation models team, led by distinguished engineer Ruoming Pang. Some team members have expressed unhappiness about considering third-party technology, with concerns this creates perception they are partially to blame for the company's AI shortcomings.

Apple faces significant talent retention challenges, with competitors offering substantially higher compensation packages. Meta Platforms has been offering some engineers annual pay packages between $10 million and $40 million to join its Superintelligence Labs group, whilst Apple typically pays its AI engineers half or less than market rates.

In financial discussions, Apple and Anthropic have disagreed over preliminary terms, with the AI startup seeking a multibillion-dollar annual fee that increases sharply each year. This has left Apple contemplating arrangements with OpenAI or other providers if it proceeds with external partnerships.

The deliberations follow executive shake-ups after AI chief John Giannandrea was sidelined following tepid responses to Apple Intelligence and Siri feature delays. Giannandrea has lost oversight of multiple AI projects, including Siri and the company's robotics unit.

---