Why Apple Relies on Claude Internally While Choosing Google for Public AI Services
While Apple appears, on the surface, to be aligning closely with Google by selecting Gemini as the next-generation intelligence behind Siri, new information suggests a very different reality inside Apple Park. According to Mark Gurman of Bloomberg, Apple’s internal operations are powered largely by Anthropic’s Claude, rather than its public-facing AI partners.
Gurman reports that Apple has deployed customized versions of Claude on its own internal servers, making the model a core tool for engineers and developers. These systems are reportedly used for writing code, testing new features, and handling internal documentation, indicating a high level of trust in Anthropic’s technology within Apple’s development pipeline.
Claude as the Backbone of Apple’s Internal Development
The decision to rely on Claude internally suggests that Apple views Anthropic’s models as highly accurate and reliable compared to other large language models currently available. Within engineering teams, Claude has become a key productivity tool, effectively supporting the development of future Apple hardware and software products.
This reliance highlights an important reality in Apple’s AI strategy. Despite its public image as a self-sufficient technology giant, Apple has not yet reached a point where its internally developed large language models can fully replace third-party solutions in critical workflows.
Why Claude Was Not Chosen for Siri
Given Apple’s apparent confidence in Anthropic, many have questioned why Claude was not selected as the primary AI engine for Siri. The answer appears to be largely financial. Reports indicate that Anthropic demanded licensing fees totaling several billion USD per year, with terms that would double costs annually over the first three years of the agreement.
Even for a company with Apple’s financial resources, these conditions were considered excessively expensive and unsustainable for a consumer-facing service deployed at global scale.
Google’s Offer Made More Business Sense
By contrast, Google presented a significantly more cost-effective proposal. In addition to already paying Apple approximately 20 billion USD annually to remain the default search engine on Safari, Google reportedly offered Gemini AI services for roughly 1 billion USD per year.
This made Gemini a far more viable option for powering Siri for the general public, allowing Apple to deliver competitive AI capabilities without incurring extreme long-term costs. As a result, Apple opted to use Google’s technology for public-facing services while reserving Anthropic’s Claude for internal use.
A Multi-Partner AI Strategy
This situation places Apple in a unique position. While iPhone users worldwide will interact with Siri powered by Gemini, the devices and software themselves are being designed and engineered with the assistance of Claude. This dual approach suggests that Apple is hedging its bets rather than committing fully to a single AI provider.
In addition to Google and Anthropic, Apple has also integrated OpenAI’s ChatGPT into certain optional features, further reinforcing its strategy of diversification. By spreading AI dependencies across multiple partners, Apple reduces risk and maintains leverage as the AI landscape continues to evolve.
Looking Toward Apple’s Own AI Future
Ultimately, Apple’s long-term goal appears to be the development of its own proprietary large language model, often referred to in reports as LLM Siri. Until that technology matures sufficiently, Apple is likely to continue relying on external specialists to remain competitive in the AI race.
For now, Apple’s approach reflects a pragmatic balance between cost efficiency, performance reliability, and strategic flexibility, even if it means trusting external AI models more than its own.
Source: 9to5Mac
