• Apple’s discussions about integrating AI models from external partners into Siri could signal a major strategic shift from in-house development by 2026
• Internal morale issues and talent retention challenges highlight Apple’s struggle to keep pace in the generative AI race

The tech world is buzzing with speculation after Bloomberg reported that Siri, Apple’s voice assistant, could be powered by AI technology from external partners, with the iPhone maker allegedly exploring partnerships with Anthropic and OpenAI to revolutionise its digital assistant.

According to Bloomberg’s sources, Apple has engaged in discussions with both Anthropic and OpenAI about incorporating their large language models into Siri. The conversations reportedly include requests for these companies to train specialised versions of their models that could operate on Apple’s cloud infrastructure for testing purposes.

This potential shift represents what Bloomberg characterises as a “monumental reversal” for Apple, which has historically relied on its proprietary Apple Foundation Models to power AI features. The company had been developing a new version of Siri scheduled for 2026 using entirely in-house technology.

Why would Apple consider this dramatic AI change on Siri?

The reported discussions suggest Apple may be acknowledging its challenges in the generative AI space. While Apple currently allows ChatGPT to handle certain web-based queries through Siri, the core assistant functionality remains powered by Apple’s technology.

Bloomberg’s sources indicate that adopting external AI models could help Apple achieve feature parity with AI assistants available on Android devices, potentially addressing criticism that the company has fallen behind in AI innovation.

In hindsight, the original Siri, launched in 2011, has faced increasing competition from more sophisticated AI chatbots. Apple’s efforts to modernise the assistant have encountered significant obstacles over the years.

A year ago, Apple demonstrated enhanced Siri capabilities that would allow the assistant to access personal data and analyse on-screen content more effectively. The company also showcased technology enabling more precise control of apps and features across Apple devices.

However, these improvements have faced repeated delays. Initially planned for early 2025, the enhanced features have been postponed indefinitely, with Bloomberg reporting they’re now targeted for next spring.

Internal uncertainty and talent challenges

Bloomberg’s sources describe Apple’s AI team as operating “with a high degree of uncertainty and a lack of clarity,” with executives still evaluating multiple strategic directions. Despite approving a multi-billion-dollar budget for 2026 to run proprietary models via cloud infrastructure, the company’s longer-term AI strategy remains unclear.

The potential shift to third-party technology has reportedly affected team morale. The Apple Siri AI development team, comprising roughly 100 engineers led by Pang Ruoming under senior director Daphne Luong, includes some of the industry’s most sought-after talent.

Bloomberg reports that team members have expressed concerns about the company considering external technology, viewing it as a reflection of their work. Some have indicated they might leave for more lucrative opportunities at competitors like Meta and OpenAI.

The talent retention challenge is significant. Meta has reportedly offered engineers annual packages ranging from $10 million to $40 million—or more—to join its Superintelligence Labs group. Bloomberg notes that Apple typically pays AI engineers significantly less than market rates, sometimes half or less than what competitors offer.

Recent departures underscore these challenges. Tom Gunter, described as one of Apple’s most senior large language model researchers, left last week after eight years with the company. Bloomberg’s sources consider him difficult to replace given his specialised skills and the competitive talent market.

Apple also nearly lost its MLX team—responsible for the key open-source system for developing machine learning models on Apple chips—before making counter-offers to retain them.

Industry context and precedent

The potential Apple Siri AI partnership wouldn’t be unprecedented in the industry. Samsung has taken a similar approach with its Galaxy AI features, which are largely based on Google’s Gemini technology despite Samsung’s branding. Amazon has also incorporated Anthropic’s technology to power its new Alexa+ service.

However, for Apple—a company that has built its reputation on controlling the entire user experience through proprietary technology—such a move would represent a significant philosophical shift.

The road ahead

Bloomberg emphasises that Apple’s exploration of third-party AI models remains in the early stages, with no final decision made. The company continues developing its internal “LLM Siri” project using proprietary models.

Apple executives reportedly believe that long-term ownership of AI models will become increasingly important as artificial intelligence becomes more central to product functionality. The company is working on various AI-dependent projects, including a tabletop robot and smart glasses.

Whether Apple ultimately chooses to partner with external AI providers or doubles down on internal development, the reported discussions highlight the intense pressure facing even the world’s most valuable technology company to remain competitive in the rapidly evolving AI landscape.

The stakes couldn’t be higher—Apple’s next move will either cement its AI comeback or confirm its struggle to remain relevant in the intelligence-first computing era.