I tried the Surface Laptop Ultra at Computex, and it’s clear: Microsoft means business

Microsoft and Nvidia are redefining the PC with the RTX Spark chip—a 1-petaflop AI superchip that turns laptops into “personal AI computers,” capable of running massive AI models locally while delivering AAA gaming and creative workloads. The Surface Laptop Ultra, unveiled at Computex 2026, is the first device to showcase this leap, but questions remain about adoption, performance, and whether the industry can keep up.

Why the RTX Spark chip is a turning point for AI on Windows

The RTX Spark isn’t just another GPU. It’s a full-stack reinvention of the PC, combining a 20-core ARM-based CPU, an AI-optimized GPU with up to 6,144 CUDA cores, and up to 128GB of unified memory—all designed to run large AI models, edit 12K video, and play games at 144FPS, all on a laptop. Nvidia calls it “the most efficient PC chip ever built,” while Microsoft frames it as the foundation for a new era where PCs don’t just run apps—they understand them. The Surface Laptop Ultra, the first device to ship with RTX Spark, is Microsoft’s boldest bet yet to prove this vision works in the real world.

But the stakes go beyond specs. For years, AI has relied on cloud servers, forcing users to upload data for processing. RTX Spark flips that script: it lets creators, developers, and gamers run AI agents—like OpenClaw or Nvidia’s own OpenShell—locally, without sending data to the cloud. That’s a game-changer for privacy, latency, and even creative workflows. As Nvidia’s CEO Jensen Huang put it: “The PC is being reinvented. For forty years, you launched apps. Click. Type. With RTX Spark and Windows, you ask—and the PC does the work.”

“The PC is being reinvented. For forty years, you launched apps. Click. Type. With RTX Spark and Microsoft Windows, you ask—and the PC does the work.”

What the Surface Laptop Ultra delivers—and what’s still unproven

At Computex 2026, Microsoft pulled out all the stops with the Surface Laptop Ultra, positioning it as the flagship RTX Spark device. The specs are staggering: a 15-inch mini-LED PixelSense Ultra display with 262 PPI and 2,000 nits of HDR brightness, a dual-fan cooling system to handle the heat, and a design that blends aluminum, glass, and a MacBook-inspired keyboard. But the real star is the RTX Spark chip itself, which ZDNET’s hands-on review called “a beast” capable of smooth gaming (demonstrated with *Indiana Jones and the Great Circle*) and professional-grade video editing.

Yet, as impressive as the demo was, real-world performance remains untested. The Surface Ultra’s unified memory architecture—up to 128GB shared between CPU and GPU—is designed to handle tasks like running 120-billion-parameter AI models, but whether that translates to usable battery life or practical workflows for most users is still unknown. Microsoft’s bet is that the RTX Spark’s efficiency will make this feasible, but The Verge notes that Arm-based chips like this still require emulation layers for legacy x86 software, which could eat into performance.

Key specs compared: RTX Spark vs. traditional PC chips

Feature	RTX Spark (Flagship)	Traditional High-End Laptop (e.g., RTX 5070)
CPU Cores	20 (Grace CPU)	8–16 (x86)
GPU Cores	6,144 (Blackwell RTX)	4,608 (RTX 5070)
Unified Memory	Up to 128GB	Up to 32GB (separate)
AI Performance	1 petaflop	~20 teraflops
Battery Life	All-day (claimed)	Varies (typically 6–10 hours)

The table above shows why RTX Spark isn’t just an upgrade—it’s a leap. Traditional laptops with discrete GPUs like the RTX 5070 struggle with unified memory and AI workloads, forcing users to choose between performance and portability. RTX Spark, by contrast, is designed to do both: run AAA games at 1440p/100FPS, edit 12K video, and host AI agents—all in a slim, battery-powered form factor. But whether this holds up outside controlled demos is the million-dollar question.

Photo: The Verge

Who stands to win—and who might get left behind

The RTX Spark ecosystem is already shaping up to be a two-tiered market. On one side, you have Microsoft’s Surface Laptop Ultra and other premium devices from ASUS, Dell, HP, Lenovo, and MSI—all targeting creators, AI developers, and power users. These machines will cost a premium, with the Surface Ultra likely starting north of $2,500. On the other side, Nvidia has hinted at lower-end RTX Spark variants with as little as 16GB of RAM, aiming to broaden adoption. But even these won’t be cheap.

The bigger question is whether this shift benefits consumers or just tech giants. For now, the advantages are clear for professionals: Adobe is already rearchitecting Photoshop and Premiere for RTX Spark, promising 2x faster AI performance. Developers can run local AI agents like OpenClaw or Nvidia’s OpenShell without cloud dependencies, while gamers get buttery-smooth performance in demanding titles. But for the average user? The value proposition is less obvious. If you’re not editing 4K video or training AI models, the RTX Spark’s capabilities might feel overkill—especially at a high price.

“Running solutions like OpenShell and the Microsoft security primitives on RTX Spark will enable users to leverage a fully integrated stack for private, personal agents running on device.”

Nvidia and Microsoft are pushing hard on the “personal AI computer” angle, but adoption hinges on two things: software and security. The RTX Spark’s local AI capabilities require new tools—like Nvidia’s OpenShell runtime and Microsoft’s Windows security primitives—to run agents safely. Companies like Nous, which developed the Hermes Agent, are already on board, but widespread adoption will depend on whether developers build for this platform. If the ecosystem grows, we could see a future where your laptop isn’t just a tool but a collaborator—automating tasks, answering queries, and even controlling smart home devices. But if the software doesn’t follow, the hardware might just sit on the shelf.

What happens next: Release dates, rivals, and risks

RTX Spark-powered devices will hit shelves this fall, with Microsoft’s Surface Laptop Ultra leading the charge. Other manufacturers like ASUS, Dell, and HP will follow with their own models, though exact pricing and availability remain under wraps. The bigger question is how this shakes up the PC industry.

Intel and AMD are already watching closely. Both have invested heavily in AI-optimized chips (Intel’s Meteor Lake, AMD’s Strix Point), but neither has matched Nvidia’s vertical integration—combining CPU, GPU, and AI acceleration in a single chip. Qualcomm, too, is a player in the Arm-based PC space, but its chips lack Nvidia’s AI and graphics prowess. The RTX Spark’s arrival could force these companies to accelerate their own AI strategies—or risk being left behind.

Then there’s the question of software. For RTX Spark to succeed, developers need to build for it. Adobe’s commitment to Photoshop and Premiere is a strong signal, but other major apps—like Final Cut Pro or Blender—haven’t yet announced support. Without broad adoption, the RTX Spark’s potential could go untapped. Microsoft and Nvidia are betting that the allure of local AI will drive adoption, but history shows that hardware alone doesn’t guarantee success.

Finally, there’s the risk of fragmentation. The RTX Spark’s Arm-based architecture means legacy x86 software will run through emulation, which could slow down performance. While Microsoft has spent years optimizing Windows for Arm, some apps—especially older or poorly optimized ones—might not run smoothly. If users find the experience clunky, it could dampen enthusiasm for what’s otherwise a groundbreaking leap.

The bottom line: A bold bet with big implications

The RTX Spark isn’t just another chip—it’s a bet on the future of computing. By merging AI, gaming, and creative workloads into a single, efficient package, Nvidia and Microsoft are pushing PCs into uncharted territory. The Surface Laptop Ultra is the first taste of what’s possible, but whether this vision becomes mainstream depends on three things:

• Software adoption: Will developers build for RTX Spark, or will it remain a niche for power users?
• Performance in the real world: Can the Surface Ultra and other devices deliver on their promises outside controlled demos?
• Price and accessibility: Will the high cost limit adoption, or will lower-end variants make this tech more accessible?

One thing is clear: the PC industry is at a crossroads. For decades, we’ve relied on cloud-based AI and discrete GPUs. RTX Spark flips that script, putting the power back in the user’s hands. If it succeeds, we could see a future where laptops aren’t just tools—but intelligent, always-on assistants. If it fails, we might just get another expensive, underutilized piece of hardware.

The Surface Laptop Ultra’s debut at Computex was a masterclass in hype and hardware. But the real test comes when users take it home. Until then, the RTX Spark remains a tantalizing promise—one that could redefine computing, or fade into the background as just another high-end chip.

“At Nous, we expect tasks to increasingly run on device as personal agents like our Hermes Agent become more capable and ubiquitous.”