RTX Spark: Nvidia unveils a Blackwell chip for Windows PCs, taking on Qualcomm
At Computex 2026, Nvidia unveiled RTX Spark, a Blackwell-based chip for Windows PCs with 20 CPU cores and 6144 GPU cores. Microsoft will release the Surface Laptop Ultra and a Dev Box, while Asus, Dell and Lenovo are preparing their own devices. Nvidia has the software ecosystem that Qualcomm lacked when it tried and failed in 2024.
AI-processed from IEEE Spectrum AI; edited by Hamidun News
At Computex 2026 in Taipei, Nvidia unveiled its long-awaited RTX Spark, a version of the Blackwell superchip for ordinary Windows PCs. The Arm-based design arrives with official backing from Microsoft and from manufacturers such as Asus, Dell and Lenovo.
What is inside RTX Spark
RTX Spark adapts the DGX Spark chip, which shipped in late 2025 as a mini workstation. Its official name is N1X. It is a system-on-chip with 20 Arm CPU cores, 6144 GPU cores and support for up to 128 GB of LPDDR5X memory.
The main difference from the mini workstation is power draw. DGX Spark runs at 140 watts, while RTX Spark laptops will draw less, which will affect performance. Exact figures depend on each manufacturer.
RTX Spark also includes an NPU (neural processing unit), which qualifies the system for Microsoft's Copilot+ certification. The NPU handles background AI tasks such as Windows Recall, while the GPU does the heavy lifting: large language models and image generation.
Microsoft is already preparing the Surface Laptop Ultra and the Surface RTX Spark Dev Box. Windows for DGX Spark desktops will arrive in the third quarter of 2026.
Why Nvidia, and not Qualcomm
History is repeating itself. In June 2024, Qualcomm and Microsoft announced Copilot+ PCs built on Arm chips, an alternative to x86 processors from Intel and AMD. The commercial result was mixed: Intel still dominates the laptop market. But RTX Spark may follow a different path. Nvidia simply carries more weight in the industry.
"Nvidia has more leverage to attract game and software developers in the AI space.
Qualcomm was not able to do that at the start," said Ryan Shrout, analyst at Signal65.
Nvidia's main advantage is its absolute dominance of the GPU software stack. Its graphics cards are the standard for gamers and professionals. Nvidia holds more than 90% of the GPU market. Game and AI software developers write for Nvidia GPUs first.
- GPU performance on par with the mobile RTX 5070
- Nvidia holds 90%+ of the GPU market
- Mature drivers and a huge software ecosystem
- Professionals value stability and compatibility
The Windows on Arm challenge
Nvidia has a strong hand, but the challenge is the same as it was with Qualcomm: Windows on Arm simply has not caught on in the mass market. At Build 2026, Microsoft talked loudly about AI and introduced Microsoft Execution Containers (MXC), containers for AI agents that run autonomously but are isolated from other functions. For now, though, the focus is on AI, not on the general-purpose PC.
The real test is whether RTX Spark will be a good general-purpose PC. Just a PC, not a revolution.
What this means
RTX Spark shows that AI in the PC is ceasing to be a gadget: the GPU becomes a core component rather than an add-on. For developers, running LLM inference locally on a laptop is no longer science fiction.
But the main question remains: can Windows on Arm escape Qualcomm's niche past and take root in the mass market?
- Nov 1, 2026Voice bots built on LLMs turn to Jev, a fast classifier, for decisions in 200–500 ms
- Nov 1, 2026MoneyGate safety layer blocked all 42 fraudulent requests to local AI models
- Oct 31, 202622B-Parameter LTX-2.5 Runs on a 10 GB RTX 3080, but Takes 5 Minutes per Shot
- Oct 30, 2026Mistral opens Large 4 weights: strong at code, but Chinese models score 32 to its 26
Key Questions Answered
What is inside RTX Spark?
RTX Spark adapts the DGX Spark chip, which shipped in late 2025 as a mini workstation. Its official name is N1X. It is a system-on-chip with 20 Arm CPU cores, 6144 GPU cores and support for up to 128 GB of LPDDR5X memory. The main difference from the mini workstation is power draw. DGX Spark runs at 140 watts, while RTX Spark laptops will draw less, which will affect performance. Exact figures depend on each manufacturer. RTX Spark also includes an NPU (neural processing unit), which qualifies the system for Microsoft's Copilot+ certification. The NPU handles background AI
Why Nvidia, and not Qualcomm?
History is repeating itself. In June 2024, Qualcomm and Microsoft announced Copilot+ PCs built on Arm chips, an alternative to x86 processors from Intel and AMD. The commercial result was mixed: Intel still dominates the laptop market. But RTX Spark may follow a different path. Nvidia simply carries more weight in the industry. "Nvidia has more leverage to attract game and software developers in the AI space. Qualcomm was not able to do that at the start," said Ryan Shrout, analyst at Signal65. Nvidia's main advantage is its absolute dominance of the GPU software stack. Its
What are the key details on The Windows on Arm challenge?
Nvidia has a strong hand, but the challenge is the same as it was with Qualcomm: Windows on Arm simply has not caught on in the mass market. At Build 2026, Microsoft talked loudly about AI and introduced Microsoft Execution Containers (MXC), containers for AI agents that run autonomously but are isolated from other functions. For now, though, the focus is on AI, not on the general-purpose PC. The real test is whether RTX Spark will be a good general-purpose PC. Just a PC, not a revolution.
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.