Apple is admitting it cannot do everything alone
For the first time in its history, Apple is preparing to build a flagship product experience on the infrastructure of its rivals — Nvidia's Blackwell chips and Google's Gemini models — to power a reimagined Siri unveiled at WWDC 2026. The company that long defined itself through total control of its technology stack is now conceding that the velocity of the generative AI era demands a different kind of humility. It is a quiet but consequential admission: that speed and capability, in this moment, matter more than self-sufficiency.
- Apple has fallen visibly behind in the generative AI race, and the gap has grown too wide to close through internal development alone.
- The decision to rely on Google's language models and Nvidia's data centre hardware marks a rupture with decades of vertically-integrated identity.
- Privacy — Apple's most powerful brand promise — is under pressure, with user data now flowing through Google's cloud and Nvidia's processors, however encrypted.
- Nvidia's Confidential Computing and Apple's Private Cloud Compute are being deployed together as a trust architecture designed to hold skeptical users.
- The new Siri is set to arrive with ChatGPT-rivaling features: context awareness, multi-step task execution, file uploads, and a Dynamic Island presence — a product transformation years in the making.
Apple is preparing to unveil a fundamentally redesigned Siri at its Worldwide Developers Conference — and for the first time, it will be built on someone else's foundation. According to reporting by The Information, the new assistant will run on Nvidia's Blackwell B200 processors and draw its intelligence from Google's Gemini language models, a striking departure for a company whose identity has long rested on owning every layer of its technology.
The generative AI race forced the reckoning. Rather than build from scratch and fall further behind, Apple is borrowing from two of its most formidable competitors. Google supplies the underlying models, hosted on its own cloud infrastructure. Nvidia supplies the computational muscle — the Blackwell B200 being among the most powerful AI chips currently available, offering meaningful gains in speed and operating efficiency over its predecessor.
Apple is not walking away from its privacy commitments. Nvidia's Confidential Computing technology will encrypt data within the hardware as it processes sensitive requests, complementing Apple's existing Private Cloud Compute system, which routes complex AI tasks through Apple-controlled servers. The combined approach is designed to reassure users that even as their requests travel through external infrastructure, meaningful protections remain in place.
The Siri that emerges will be nearly unrecognizable. Long-promised features — personal context awareness, on-screen reading, multi-step task completion across apps — are finally expected to arrive. A standalone chatbot experience, built to compete directly with ChatGPT and Claude, will support voice and text, conversation history, and mixed media inputs. Visually, Siri will gain an animated presence inside the Dynamic Island, with a central search bar anchoring the AI experience.
The deeper significance is philosophical. Apple is acknowledging, however quietly, that it cannot move fast enough alone. Whether users will extend their trust to a pipeline running through Google's servers and Nvidia's chips is the question the company will begin answering next week.
Apple is about to announce something it has never quite done before: a major product overhaul built on someone else's chips and someone else's artificial intelligence. The company that built its reputation on controlling every layer of its technology stack is preparing to unveil a revamped Siri at next week's Worldwide Developers Conference, and according to reporting by The Information, it will run on Nvidia's Blackwell B200 processors while drawing intelligence from Google's Gemini models.
This is a striking reversal. For decades, Apple has prided itself on vertical integration—designing its own silicon, writing its own software, keeping the entire experience within its walled garden. But the generative AI race has forced a reckoning. The company fell behind in the chatbot space, and rather than build everything from scratch, it is borrowing from two of its fiercest competitors to catch up. Google will provide the underlying language models, hosted on Google's cloud infrastructure. Nvidia will supply the raw computational muscle—the Blackwell B200 is among the most powerful data centre chips available for AI work, offering faster performance and lower operating costs than Nvidia's previous generation hardware.
Apple is not abandoning its privacy principles in the process. The company plans to enable Nvidia's Confidential Computing technology, which encrypts data directly within the hardware as it processes sensitive information. This matters because privacy has become Apple's primary selling point in the AI era. At last year's developer conference, Apple introduced Private Cloud Compute, a system that lets complex AI tasks run in the cloud on Apple-controlled silicon, reassuring users that their personal data stays protected even when requests leave the device. The new Siri strategy extends that logic: sensitive processing happens on Nvidia hardware, but with encryption safeguards built in.
The Siri that emerges from this partnership will be unrecognizable compared to the voice assistant users have known for over a decade. Reports suggest it will finally gain features Apple has been promising for years: the ability to understand personal context, to read what is displayed on your screen, to complete multi-step tasks across both Apple apps and third-party services. Beyond voice, Apple is preparing a standalone chatbot experience designed to compete directly with ChatGPT and Claude. This chatbot will support both voice and text, let users view conversation history, upload files and images, and submit requests that mix text and visual inputs. There is also a visual redesign in the works—Siri will gain an animated presence within the Dynamic Island, the pill-shaped cutout at the top of newer iPhones, with a central search bar for AI interactions.
What makes this moment significant is not just the technical capability but the philosophical shift it represents. Apple is admitting that it cannot do everything alone, at least not quickly enough to matter in a market moving at the speed of generative AI. By outsourcing the heavy lifting to Nvidia and Google, the company is trading some control for speed and capability. Whether users will accept this new arrangement—trusting their data to a pipeline that runs through Google's servers and Nvidia's hardware—remains an open question. But Apple is betting that the features and performance gains will outweigh the discomfort of relying on rivals. The company will have its answer next week.
Citações Notáveis
Apple plans to enable Nvidia's Confidential Computing technology to protect sensitive information while it is being processed by encrypting data directly within the hardware environment— The Information reporting
A Conversa do Hearth Outra perspectiva sobre a história
Why would Apple, of all companies, hand over control of Siri to Google and Nvidia? That seems to contradict everything Apple stands for.
It does, but the generative AI moment forced Apple's hand. They fell behind. ChatGPT and Claude were already in users' hands while Apple was still talking about on-device processing. Sometimes you have to swallow your philosophy to stay relevant.
But doesn't this create a privacy nightmare? Your Siri requests now flow through Google's servers?
That's the tension Apple is trying to solve with encryption and Confidential Computing. The data gets encrypted at the hardware level, so theoretically Google can't see what you're asking. Whether users believe that is another matter.
What's actually new about this Siri compared to what exists now?
Everything. It understands context, can see your screen, chains together multiple actions, and works like a real chatbot—not just a voice command parser. It's the Siri that should have existed five years ago.
And Nvidia's Blackwell chip is doing what exactly?
The heavy lifting. Running the language models, processing the requests, handling the compute that makes all those features possible. Nvidia's hardware is built for this kind of work.
So Apple is essentially admitting it can't compete in AI alone?
Not admitting—adapting. There's a difference. Apple is still controlling the user experience, the privacy layer, the integration with your device. It's just borrowing the brain.