Apple is reportedly poised to pay rival Google approximately $1 billion annually for a temporary integration of Google’s advanced Gemini artificial intelligence to power a major upgrade of its Siri voice assistant.
The significant expenditure underscores Apple’s strategic move to quickly enhance its AI capabilities. It also highlights a reliance on a direct competitor for a critical component of its user experience.
According to a Bloomberg report, Apple tested AI models from other developers, including OpenAI’s ChatGPT and Anthropic’s Claude, before settling on Google’s Gemini.
This collaboration is intended as an interim solution. Apple’s long-term goal remains to develop its own robust AI technology.
The planned upgrade marks a substantial leap in Siri’s intelligence. The Gemini model Apple intends to use boasts 1.2 trillion parameters, significantly larger than the 150 billion parameters in Apple’s current cloud-based AI, known as Apple Intelligence.
Gemini is expected to assist Siri with more complex tasks. These include summarizing intricate information and planning according to difficult commands. Some existing Siri features will continue to utilize Apple’s in-house AI.
Apple stated that the AI operations will run on its Private Cloud Compute servers. This measure is intended to ensure user data remains segregated and does not intermingle with Google’s systems.
Despite the agreement with Google, Apple has made it clear that it does not want Gemini as a permanent solution. The company’s development teams are actively working on their own AI.
This internal project aims to create a 1 trillion-parameter AI model. Apple hopes this proprietary system will be advanced enough to replace Gemini within the coming year.
However, Apple acknowledges that “catching up” is challenging. Google continuously improves Gemini, setting a moving target for its competitors. The new Siri, powered by this enhanced AI, is expected to launch next year.