Apple is preparing its most consequential Siri reset in years, with plans under discussion to use Google’s Gemini models and Nvidia’s Blackwell chips to power more capable artificial intelligence features in iOS 27.
The move would mark a notable shift for a company that has long preferred tight control over its hardware, software and cloud infrastructure. Siri, once a defining feature of the iPhone, has fallen behind rival assistants built around generative AI. Apple now faces pressure to show that its delayed assistant overhaul can handle complex requests, understand personal context and act across apps without undermining the company’s privacy commitments.
The upgraded Siri is expected to be tied to iOS 27, with a preview possible around Apple’s annual developers conference. The system is being designed to deliver more conversational responses, interpret what is visible on a user’s screen and carry out deeper in-app actions. Those were among the capabilities first promised under Apple Intelligence, but the company pushed back delivery after internal testing showed the experience was not reliable enough for broad release.
The reported architecture would place Google and Nvidia at the centre of Apple’s AI catch-up effort. Google’s Gemini models could provide the reasoning layer for more advanced Siri requests, while Nvidia’s Blackwell B200 chips inside Google Cloud data centres would handle demanding inference workloads. The arrangement would allow Apple to tap large-scale AI computing capacity without waiting for its own server infrastructure to match the performance required for a widely deployed assistant.
Apple’s challenge is not merely technical. The company has marketed Apple Intelligence as a privacy-first system built around on-device processing and Private Cloud Compute, where more complex tasks are handled on Apple silicon in secure servers. Using Google-hosted infrastructure and Nvidia hardware would require Apple to reassure users that sensitive personal data remains protected. Nvidia’s confidential computing tools, designed to encrypt data during processing, are expected to be part of that answer.
The stakes are high because Siri has become a visible weakness in Apple’s broader AI strategy. OpenAI, Google, Microsoft, Anthropic and Amazon have all pushed voice and chat interfaces towards more capable digital agents. Google has embedded Gemini across Android and Workspace, Microsoft has built Copilot into Windows and Office, and Amazon has been reworking Alexa around generative AI. Apple, by contrast, has moved more cautiously, favouring controlled features such as writing tools, image generation, notification summaries and selected ChatGPT integration.
That caution has carried costs. The first wave of Apple Intelligence features did not fully answer investor concerns that the iPhone maker was lagging in generative AI. Delays to the personalised Siri upgrade added to doubts over execution. A stronger assistant could help Apple defend the iPhone’s premium position, deepen user dependence on its ecosystem and create a new route for services revenue if advanced AI functions are eventually bundled into paid offerings.
Hardware requirements could also become part of the commercial equation. More advanced AI features are likely to demand newer chips and larger memory capacity on devices. That may encourage iPhone upgrades, especially among users with older models that cannot run the full Apple Intelligence suite. At the same time, limiting the best functions to newer devices risks frustrating customers who expected major software improvements on existing iPhones.
A partnership with Google would not be unprecedented. Google already pays Apple heavily to remain the default search engine on Safari, a relationship that has drawn regulatory scrutiny in the United States. A deeper AI arrangement could attract further examination, particularly if Apple positions Siri as a gateway to third-party services, search, shopping and app actions. Regulators are already assessing the influence of dominant technology platforms over digital markets, and AI assistants may become the next area of concern.
For Google, the reported role would strengthen Gemini’s position as a core supplier of AI infrastructure beyond its own products. For Nvidia, another large-scale deployment on Blackwell chips would reinforce its grip on the AI computing market, where demand for high-performance processors continues to outstrip supply. For Apple, the calculation is more delicate: relying on two major technology partners could accelerate Siri’s recovery, but it also exposes how difficult it has become to build frontier AI systems entirely in-house.
The move would mark a notable shift for a company that has long preferred tight control over its hardware, software and cloud infrastructure. Siri, once a defining feature of the iPhone, has fallen behind rival assistants built around generative AI. Apple now faces pressure to show that its delayed assistant overhaul can handle complex requests, understand personal context and act across apps without undermining the company’s privacy commitments.
The upgraded Siri is expected to be tied to iOS 27, with a preview possible around Apple’s annual developers conference. The system is being designed to deliver more conversational responses, interpret what is visible on a user’s screen and carry out deeper in-app actions. Those were among the capabilities first promised under Apple Intelligence, but the company pushed back delivery after internal testing showed the experience was not reliable enough for broad release.
The reported architecture would place Google and Nvidia at the centre of Apple’s AI catch-up effort. Google’s Gemini models could provide the reasoning layer for more advanced Siri requests, while Nvidia’s Blackwell B200 chips inside Google Cloud data centres would handle demanding inference workloads. The arrangement would allow Apple to tap large-scale AI computing capacity without waiting for its own server infrastructure to match the performance required for a widely deployed assistant.
Apple’s challenge is not merely technical. The company has marketed Apple Intelligence as a privacy-first system built around on-device processing and Private Cloud Compute, where more complex tasks are handled on Apple silicon in secure servers. Using Google-hosted infrastructure and Nvidia hardware would require Apple to reassure users that sensitive personal data remains protected. Nvidia’s confidential computing tools, designed to encrypt data during processing, are expected to be part of that answer.
The stakes are high because Siri has become a visible weakness in Apple’s broader AI strategy. OpenAI, Google, Microsoft, Anthropic and Amazon have all pushed voice and chat interfaces towards more capable digital agents. Google has embedded Gemini across Android and Workspace, Microsoft has built Copilot into Windows and Office, and Amazon has been reworking Alexa around generative AI. Apple, by contrast, has moved more cautiously, favouring controlled features such as writing tools, image generation, notification summaries and selected ChatGPT integration.
That caution has carried costs. The first wave of Apple Intelligence features did not fully answer investor concerns that the iPhone maker was lagging in generative AI. Delays to the personalised Siri upgrade added to doubts over execution. A stronger assistant could help Apple defend the iPhone’s premium position, deepen user dependence on its ecosystem and create a new route for services revenue if advanced AI functions are eventually bundled into paid offerings.
Hardware requirements could also become part of the commercial equation. More advanced AI features are likely to demand newer chips and larger memory capacity on devices. That may encourage iPhone upgrades, especially among users with older models that cannot run the full Apple Intelligence suite. At the same time, limiting the best functions to newer devices risks frustrating customers who expected major software improvements on existing iPhones.
A partnership with Google would not be unprecedented. Google already pays Apple heavily to remain the default search engine on Safari, a relationship that has drawn regulatory scrutiny in the United States. A deeper AI arrangement could attract further examination, particularly if Apple positions Siri as a gateway to third-party services, search, shopping and app actions. Regulators are already assessing the influence of dominant technology platforms over digital markets, and AI assistants may become the next area of concern.
For Google, the reported role would strengthen Gemini’s position as a core supplier of AI infrastructure beyond its own products. For Nvidia, another large-scale deployment on Blackwell chips would reinforce its grip on the AI computing market, where demand for high-performance processors continues to outstrip supply. For Apple, the calculation is more delicate: relying on two major technology partners could accelerate Siri’s recovery, but it also exposes how difficult it has become to build frontier AI systems entirely in-house.
Topics
Technology