Report: Apple Shifts To Partner-first AI, Eyes Gemini For Siri

Apple is exploring a deal to use Google’s Gemini artificial-intelligence system to power a long-overdue overhaul of Siri after internal efforts fell short.

• Apple may use Google’s Gemini to revamp Siri after in-house AI fell short.
• Gemini would be adapted for Apple’s private servers; launch target is 2026.
• Stocks rose on the report; Apple also weighed Anthropic and OpenAI, with deal terms still unclear.
  
  

Commenting on the report that was first picked up by Bloomberg, CNBC’s Steve Kovach said Apple is considering Google’s Gemini AI system because internal efforts to develop its own AI technology has failed. Alphabet shares rose about 3% intraday to a record high on the report, while Apple gained roughly 1.5% as investors appear to be happy that Apple is doing something to advance its otherwise disappointing approach to AI.

For the tie-up to work, Google would “have to change its AI system a little bit to work on Apple’s private servers,” Kovach reported, noting that Apple has also evaluated large language models from Anthropic and OpenAI. The potential agreement will most likely be seen as a shift in CEO Tim Cook’s AI strategy away from building in-house and relying on external partners instead.

This shouldn’t come as much as a surprise. Back in October 2024 a Bloomberg study found that Siri’s accuracy and response speed was 25% and 30% worse compared to ChatGPT. Apple’s own internal research found its AI ambitions are at least two and a half years behind its rivals.

On Apple’s recent earnings call, Cook was asked a question about whether models are becoming commoditized. As Kovach paraphrased it, Cook “didn’t want to talk about that because it would give away his strategy,” a response that analysts read as consistent with a partnership-first posture.

The Siri revamp is “not expected to launch until 2026,” which Kovach described as “the new deadline Apple set for itself.” That extended timetable suggests Apple aims to include the technology into iPhone software and services, likely with on-device and private-cloud components, rather than rush out a stopgap.

Key uncertainties remain, especially on pricing and revenue sharing.

“We don’t know what the economics of this are like,” Kovach said. “Is it going to be Google paying Apple? Is it going to be Apple paying Google? How those economics works are going to be really interesting.”

He also highlighted a separate report that Alphabet secured a six-year, $10 billion cloud contract with Meta, adding to momentum around Google’s enterprise AI push.