Apple considers outsourcing Siri's brainpower to OpenAI's ChatGPT or Anthropic's Claude

Apple is reportedly in advanced discussions with leading AI firms OpenAI and Anthropic as it considers a major overhaul of Siri, potentially sidelining its own artificial intelligence efforts. The company is exploring the use of external large language models (LLMs) to power future iterations of its voice assistant, in a move that could mark a significant shift in Apple’s approach to generative AI.

According to a Bloomberg report by Mark Gurman, Apple has asked both OpenAI and Anthropic to customise and train their AI models for use on Apple’s cloud infrastructure. This would allow the tech giant to maintain control over user privacy while leveraging more advanced third-party AI systems.

If implemented, the move would represent a notable reversal from Apple’s existing strategy. The company has long prioritised developing its own in-house AI tools, branded as Apple Foundation Models, and has already allocated a multibillion-dollar budget to operate these models via the cloud by 2026. However, insiders told Bloomberg that the performance of Apple’s internal models is currently viewed as inferior to offerings from external providers.

Executives including Craig Federighi and John Giannandrea are said to be increasingly open to partnering with outside firms to accelerate progress. One project under consideration is a 2026 update to Siri, coinciding with the launch of iOS 26.4, that would bring deeper contextual awareness and the ability to execute more complex, multi-step tasks based on on-screen content and user data.

Despite these talks, no final decision has been made. A separate internal initiative called “LLM Siri” remains in active development, and could result in a redesigned assistant powered entirely by Apple’s own AI models by 2025.

The shifting strategy has reportedly led to internal tensions. Bloomberg's Mark Gurman reports that one senior Apple AI engineer, Tom Gunter, left the company last week. The team behind Apple’s MLX open-source AI framework has also allegedly threatened to quit, citing frustration with the company’s lack of direction in generative AI.

Apple’s challenges come amid increasing competition for AI talent. Companies such as Meta and OpenAI are offering salaries that can exceed Apple’s compensation by more than double. Meta CEO Mark Zuckerberg recently announced a new unit called Meta Superintelligence Labs, focused on building artificial general intelligence (AGI), and has been hiring aggressively from Apple’s competitors.

Apple's willingness to consider third-party AI providers mirrors a strategy already employed by rivals. Samsung, for instance, uses Google’s Gemini AI under its Galaxy AI branding. Anthropic’s Claude is also used by Amazon to enhance its new Alexa+ service.

Apple has also explored potential acquisitions to boost its AI capabilities, including talks with Perplexity and Thinking Machines Lab, a startup founded by former OpenAI Chief Technology Officer Mira Murati.

While Apple’s long-term goal remains to eventually own and operate its own AI models, its current focus appears to be on quickly narrowing the gap with competitors and ensuring that Siri no longer lags behind its Android counterparts.