Apple is partnering with Google and Nvidia to power the next-generation, AI-enhanced version of Siri, according to a report from The Information. The move marks a significant strategic shift for Apple, which typically prefers to develop its technology in-house or diversify across multiple suppliers.
The report indicates that some Siri queries will be processed on Google Cloud using a licensed version of Gemini. Google’s data centers will use Nvidia’s Blackwell B200 chips with a built-in confidential compute feature that encrypts user data during processing—a critical requirement for Apple’s privacy commitments.
A new direction for Apple
This partnership is unusual for Apple, which has historically kept its core services running on its own infrastructure. However, the computational demands of modern AI-powered assistants have forced the company to seek external help. Apple’s own data center capabilities and AI models are not yet at the level needed to compete with ChatGPT, Gemini, and other advanced AI assistants.
The new Siri is expected to launch in September alongside iOS 20 and the iPhone 18 series. It will reportedly be capable of handling more complex queries, maintaining context across conversations, and integrating more deeply with third-party apps. The partnership with Google and Nvidia is specifically for the “cloud-based” queries—simpler requests that can be processed on-device will still use Apple’s own neural engine.
Privacy implications
Apple has emphasized that the confidential computing features built into the Nvidia Blackwell chips will ensure that Apple cannot see user data sent to Google Cloud, and Google cannot access it either. The system is designed so that encrypted tokens are processed and discarded, with no data stored on either company’s servers beyond what’s necessary to complete the request.
This approach could serve as a model for how tech companies can partner on AI without compromising user privacy—a concern that has become central to the AI assistant wars.
