In San Francisco this week, Microsoft unveiled the Surface Ultra — a $2,599 laptop built around Nvidia's RTX Spark chip — as a quiet but consequential declaration that the age of cloud-only artificial intelligence is giving way to something more intimate and immediate. The machine embodies what Microsoft calls 'hybrid intelligence,' a vision where AI does not live somewhere distant and invisible, but runs alongside us, on the very device in our hands. This partnership between two of America's most powerful technology companies is less about a single product than about who will define the next
Microsoft launches $2,599 Surface Ultra with Nvidia AI chip, betting on hybrid intelligence
You will expect this hybrid intelligence to be everywhere
Mark
So Microsoft is spending $2,599 on a laptop. That's expensive. What makes this one different from what they've been selling?
Mimi
It's the chip inside. They're using Nvidia's RTX Spark, which is built specifically for running AI tasks right on the device itself, not just in the cloud. That's the shift—local processing instead of always sending data away.
Luke
But how much of the AI actually runs locally versus in the cloud? The reporting says "some" and "other applications." That's vague. We don't know the split.
Mimi
Fair point. It's a hybrid model, so it's both. The idea is you don't have to choose anymore.
Mark
Why does Microsoft care about this now? They've been fine with cloud computing for years.
Mimi
Because Apple just released AI-capable Macs, and Intel and Qualcomm have owned the personal computer chip market forever. Microsoft and Nvidia are trying to compete. Plus, Microsoft wants to reduce how much it depends on OpenAI.
Luke
That's the real story underneath. Microsoft invested heavily in OpenAI, but now they're building their own AI models and moving away from that dependence. The Surface Ultra is part of that independence play.
Mark
So this is about control? Owning the whole chain—the chip, the software, the cloud?
Mimi
Exactly. If you control the hardware and the software, you control the experience. You're not beholden to another company's models or pricing.
Luke
Though we should note—Microsoft still uses OpenAI models in Copilot. They're not fully independent yet. Their own in-house models only power a few products like GitHub's Copilot.
Mark
So this is a long game.
Mimi
Very much so. The Surface Ultra is one move in a much larger strategy.
Le Pouls
- The race to embed AI directly into personal hardware has reached a new intensity, with Microsoft and Nvidia joining forces to challenge Apple's AI-capable Macs and Qualcomm's long dominance of the PC chip market.
- The Surface Ultra's $2,599 price tag and October 16 ship date signal that hybrid on-device and cloud AI is no longer a roadmap promise — it is a product on a shelf.
- Satya Nadella and Jensen Huang appeared together to argue that the boundary between local and cloud computing will dissolve entirely, making 'hybrid intelligence' as invisible and expected as electricity.
- Microsoft is quietly but deliberately loosening its dependence on OpenAI, having launched its own in-house AI models in June and now building hardware that anchors its AI ambitions in silicon it does not have to borrow.
- As rivals like OpenAI and Anthropic move into productivity software, Microsoft is betting that owning the full stack — chip, operating system, cloud, and application — will keep users inside its ecosystem for years to come.
In San Francisco this week, Microsoft unveiled the Surface Ultra — a $2,599 laptop built around Nvidia's RTX Spark chip — as a quiet but consequential declaration that the age of cloud-only artificial intelligence is giving way to something more intimate and immediate. The machine embodies what Microsoft calls 'hybrid intelligence,' a vision where AI does not live somewhere distant and invisible, but runs alongside us, on the very device in our hands. This partnership between two of America's most powerful technology companies is less about a single product than about who will define the next era of human-computer collaboration — and how much of that future any one company will control.
Microsoft announced the Surface Ultra on Wednesday, a laptop priced at $2,599 and powered by Nvidia's RTX Spark chip — the company's most direct statement yet that artificial intelligence belongs not only in distant data centers, but in the machine sitting open on your desk. The device began shipping October 16 and is aimed at developers and creators who need serious AI capability both online and off.
For years, meaningful AI required sending data to remote servers — a process that is slow, costly, and dependent on connectivity. Now Microsoft, Apple, Intel, and others are racing to bring that processing power local. At a San Francisco event, CEO Satya Nadella described the shift as inevitable, envisioning a world where users never have to think about whether a task runs on-device or in the cloud. Nvidia's Jensen Huang, appearing alongside him, went further, arguing that autonomous AI systems will fundamentally reshape how software is built and that every computing device must be reimagined.
The Surface Ultra handles some AI tasks locally through its Nvidia chip while routing heavier workloads to Microsoft's cloud — a hybrid model designed to give users flexibility without sacrificing power. The partnership is also strategic: Apple released AI-capable desktop computers in September, and Intel and Qualcomm have long held the PC chip market. Microsoft and Nvidia are positioning themselves as a unified answer to that competition.
The launch arrives at a telling moment in Microsoft's relationship with artificial intelligence more broadly. In June, the company unveiled its own in-house AI models, a meaningful step toward reducing its reliance on OpenAI — the company it has backed heavily and whose technology powers Copilot across Word, Excel, and PowerPoint. Microsoft renegotiated that partnership last year, securing only a non-exclusive license through 2032. The Surface Ultra is the clearest sign yet that Microsoft intends to control more of its AI future — from the chip up — rather than lease it from anyone else.
Microsoft announced Wednesday the arrival of the Surface Ultra, a laptop priced at $2,599 that marks the company's latest bet on what executives are calling "hybrid intelligence"—a computing model where artificial intelligence runs partly on the device itself and partly on distant cloud servers. The machine uses Nvidia's RTX Spark chip, the semiconductor maker's latest AI processor, which began shipping October 16.
The move reflects a broader industry pivot. For years, running sophisticated AI required sending data to remote data centers, a process that is both slow and expensive. Now Microsoft, Apple, Intel, and others are racing to embed AI processing power directly into personal computers, so users can handle certain tasks locally without leaving their machines. Satya Nadella, Microsoft's chief executive, framed this shift as inevitable during an event in San Francisco. "There will never be a moment where we will go and look and say: 'Oh, this runs locally, this runs in the cloud,'" he said. "You will expect this hybrid intelligence to be everywhere and pervasive."
Nvidia's Jensen Huang, appearing alongside Nadella, reinforced the vision. He argued that artificial intelligence will become woven into every computing device and that autonomous systems—what the industry calls agentic AI, software designed to perform tasks without human direction—will reshape how software gets built. "The computer has to be revolutionized," Huang said. The partnership between the two American tech giants is strategic. Apple released AI-capable desktop computers, the Mac mini and Mac Studio, in September. Intel and Qualcomm have long dominated the personal computer chip market. By joining forces, Microsoft and Nvidia are positioning themselves as serious contenders in a sector where the stakes are rising as AI becomes central to how people work.
The Surface Ultra is aimed at developers and creators, according to Pavan Davuluri, Microsoft's executive vice president for Windows and devices. The machine will handle some AI tasks locally on its Nvidia chip while routing other applications to Microsoft's cloud services. This hybrid approach lets users work offline when needed while still accessing the full power of remote computing when necessary.
The announcement arrives amid a broader reshaping of Microsoft's relationship with artificial intelligence. In June, the company unveiled its own in-house AI models, a significant step toward reducing its dependence on OpenAI, the creator of ChatGPT. For years, Microsoft has been the largest investor in OpenAI and relied heavily on its technology to power Copilot, the AI assistant now being integrated into Word, Excel, and PowerPoint. Last year, Microsoft renegotiated its partnership with OpenAI and secured only a non-exclusive license to the company's technology through 2032, though it remains OpenAI's biggest shareholder. The company has been working to build what it calls "AI PCs" since at least 2024, when its laptops ran on Qualcomm chips. The Surface Ultra represents a significant hardware upgrade and a signal that Microsoft is serious about controlling more of its AI stack—from chips to software to cloud services. As competitors like OpenAI and Anthropic launch their own document creation tools, Microsoft is betting that embedding AI directly into its hardware and productivity software will keep users locked into its ecosystem.
Citations marquantes
There will never be a moment where we will go and look and say: 'Oh, this runs locally, this runs in the cloud.' You will expect this hybrid intelligence to be everywhere and pervasive.— Satya Nadella, Microsoft CEO
The computer has to be revolutionized.— Jensen Huang, Nvidia CEO
Envie de l'histoire complète ?
Lire l'original sur The Economic Times ↗
Technology