Apple now depends on Anthropic for much of its internal AI work, even after signing a public partnership with Google. Inside the company, teams use Anthropic models to build products, test ideas, and run internal tools. This setup shows that Apple’s real AI engine sits behind the scenes, not in its consumer features.
Bloomberg’s Mark Gurman shared this in a recent interview on TBPN. He explained how deep Anthropic’s role runs inside Apple and why the company first tried to base Siri and other tools on Claude before turning to Google.
Apple’s internal tools run on Claude
Gurman described Apple’s current setup in direct terms. He said, “Apple runs on Anthropic at this point. Anthropic is powering a lot of the stuff Apple is doing internally in terms of product development, a lot of their internal tools.” He also added, “They have custom versions of Claude running on their own servers internally.”
This means Apple does not just use off-the-shelf AI. It runs private versions of Claude inside its own infrastructure. Teams use these models to write code, test features, and speed up product work. So while Apple shows Google Gemini to users through Siri, the company itself relies on Anthropic to get real work done.
Why Apple did not make Anthropic its main partner
Gurman also explained why Apple did not end up launching Siri on top of Claude. He said Apple planned to rebuild Siri around Anthropic before the Google deal happened. He described the talks in clear terms: “Apple actually was going to rebuild Siri around Claude. But Anthropic was holding them over a barrel.”
The problem was money. According to Gurman, Anthropic asked for several billion dollars per year, and those fees would double each year for the next three years. Apple saw that cost as too high. As a result, Apple shifted toward Google, whose Gemini deal costs about one billion dollars a year.
At the same time, Apple also had to think about its long running Safari search deal with Google. That uncertainty pushed Apple to look at Anthropic and OpenAI first. Even so, price and timing ended up favoring Google for the public Siri rollout.
Apple’s AI strategy
Even after choosing Google for Siri, Apple did not walk away from Anthropic. The company still uses Claude for internal work, which shows where Apple places its trust. Product teams already depend on it, and their workflows now revolve around it.
Gurman summed it up when he said, “This Google deal just came together a few months ago. They were not going to use Google.” That line makes one thing clear. Apple’s real AI backbone already existed before Gemini entered the picture.
In practice, this leaves Apple with two AI tracks. One runs through Google for public features like Siri. The other runs through Anthropic for product development and internal tools. When you look at how Apple builds and tests its software, one fact stands out. Apple runs on Anthropic at this point.
