Surface RTX Spark Dev Box: Price, Specs and Who It's For

Microsoft built a small aluminum desktop for people who run AI models all day. It costs $5,999, and it isn't for everyone.

Running large AI models usually means renting cloud GPUs by the hour. Microsoft wants developers to do more of that work on their own desks instead. The Surface RTX Spark Dev Box, introduced at Microsoft Build in June, is a compact developer PC built around NVIDIA's RTX Spark chip. Microsoft opened preorders this week, with shipping set for November 2026.

This article covers what the Dev Box offers, how its security works, what it costs, and who should actually buy it.

What Is the Surface RTX Spark Dev Box?

It's a desktop-sized AI developer PC designed for local-first AI work, meaning you fine-tune and run models on your own hardware instead of in the cloud. It's the second Surface built on NVIDIA's RTX Spark platform, after the Surface Laptop Ultra. Engadget's hands-on compared it to Apple's Mac Studio rather than to a traditional PC tower.

Inside the Box: Key Specs

Chip: NVIDIA RTX Spark, with a 20-core Arm CPU and a Blackwell GPU
GPU cores: 6,144 CUDA cores
Memory: 128GB of unified memory shared by the CPU and GPU
Storage: 2TB SSD
AI compute: Up to 1 petaflop
Power: A 100W thermal envelope
Ports: USB-C, USB-A, HDMI, Ethernet, and a headphone jack, but no card slot

Unified memory is the key idea. The CPU and GPU draw from one large pool, so big models don't need to be copied between separate memory banks. Microsoft says that's enough to run models with more than 120 billion parameters locally, a vendor claim worth testing once independent reviews arrive.

Design and Cooling

The chassis is anodized aluminum with 1,000 air vents, a nod to its 1,000 teraflops of AI compute. Microsoft says the body doubles as a cooling system built to keep performance steady through long training runs. Some coverage describes the design as passively cooled, but Engadget couldn't see the machine powered on, so real-world heat and noise remain unconfirmed. It weighs at least five to six pounds, so it's a desk machine, not a portable one.

Ready to Code on Day One

The Dev Box ships with a developer-tuned version of Windows 11 Pro. VS Code, GitHub Copilot, WSL, PowerShell 7, Git, Python, and Node.js come preinstalled. Because RTX Spark supports NVIDIA's CUDA and TensorRT tools, developers can use familiar AI frameworks without a separate compatibility layer.

Security and Local AI Data Privacy

Security is a core part of Microsoft's pitch, since many teams can't send proprietary code or sensitive data to the cloud. The Dev Box uses chip-to-cloud protections aligned with Microsoft's Zero Trust approach:

Secured-core PC architecture for hardware and firmware protection
BitLocker drive encryption
Microsoft Defender endpoint protection
Entra ID and Intune support, so IT teams can manage identity and policy at scale

Keeping models and data on a local machine reduces exposure, but it doesn't solve every compliance problem. Local AI work can still create risks, such as unapproved models, unmanaged Python environments, and sensitive datasets stored on a single device. IT teams still need clear policies.

Dev Box vs Cloud GPUs and Other AI Workstations

Against the cloud, the Dev Box trades flexibility for predictability. You pay once instead of per hour, and you keep data local. But it won't match the raw throughput of large cloud GPU clusters, and it can't scale up on demand. Against Apple's Mac Studio, it offers CUDA support and Windows, which Apple doesn't. Against NVIDIA's own DGX Spark, it differs mainly in running a full Windows 11 Pro environment with Microsoft's management tools.

One caveat applies to the whole RTX Spark family: it runs Windows on Arm. Most developer tools now run natively, but some x86-only software may need emulation.

Price and Availability

The Surface RTX Spark Dev Box starts at $5,999.99. That's more than twice the $2,599 starting price of the Surface Laptop Ultra. Preorders opened on October 7, 2026, with shipping in November. At launch, Microsoft is selling it only in the United States through Microsoft.com, and it hasn't confirmed international availability.

Who Should Buy It?

The Dev Box fits AI engineers who fine-tune or run large models daily, and enterprise teams that need sensitive work to stay on managed local hardware. It also suits developers tired of unpredictable cloud GPU bills. Skip it if you only run AI models occasionally, since cloud rentals will likely cost less. If you need portability, the Surface Laptop Ultra is the better fit. And if your toolchain depends on x86-only software, test Arm compatibility before you buy.