๐ฐThe VergeโขStalecollected in 10m
Nvidia to unveil new ARM processors at Computex
๐กNvidia's potential shift to ARM-based PC chips could redefine local AI hardware requirements for developers.
โก 30-Second TL;DR
What Changed
Jensen Huang to present GTC Taipei keynote at Computex.
Why It Matters
If successful, this move could challenge the x86 dominance in the laptop market and provide developers with more powerful ARM-based local inference capabilities for AI workloads.
What To Do Next
Monitor the Computex keynote for specifications on N1/N1X to evaluate if your local AI models can benefit from optimized ARM-based hardware acceleration.
Who should care:Developers & AI Engineers
Key Points
- โขJensen Huang to present GTC Taipei keynote at Computex.
- โขRumored N1 and N1X ARM processors targeting Windows PCs.
- โขPotential strategic partnership with Microsoft for Surface lineup.
- โขSignals a renewed push for high-performance ARM architecture in the PC market.
๐ง Deep Insight
Web-grounded analysis with 23 cited sources.
๐ Enhanced Key Takeaways
- โขThe N1X is reportedly a collaboration between Nvidia and MediaTek, built on TSMC's 3nm process, featuring a 2.5D package that integrates a CPU die and a Blackwell-architecture GPU die connected via NVLink C2C.
- โขThe N1X chip is rumored to include a 20-core ARM CPU (10 Cortex-X925 performance and 10 Cortex-A725 efficiency cores) and a Blackwell-architecture GPU with 6,144 CUDA cores, which is comparable to a desktop RTX 5070.
- โขNvidia's entry into the Windows on ARM PC market, potentially with Microsoft's Surface lineup, aims to challenge Qualcomm's previous exclusivity and provide a more competitive platform with strong GPU and AI performance, including support for the full CUDA software stack.
- โขLower-end N1 chip variants are also expected, featuring 10 to 12 ARM CPU cores and GPUs with 2,048 to 2,560 CUDA cores (comparable to an RTX 5050), targeting more affordable laptops with a lower power envelope of 18-45W.
- โขThe N1X is reportedly based on the GB10 Superchip, which Nvidia previously used in its AI-focused DGX Spark mini PC, indicating a strong foundation in AI workloads for the new PC chips.
๐ Competitor Analysisโธ Show
| Feature/Category | Nvidia N1/N1X (Rumored) | Qualcomm Snapdragon X Elite | Intel/AMD (x86) |
|---|---|---|---|
| CPU Cores | Up to 20 ARM v9.2 cores (10P+10E) | Up to 12 Oryon cores (all performance) | Varies widely (e.g., Intel Core Ultra 9 285HX with 22,104 multi-core score in Cinebench 2024) |
| GPU Architecture | Blackwell (up to 6,144 CUDA cores, RTX 5070-class) | Adreno GPU (integrated) | Integrated (e.g., Intel Iris Xe) or Discrete GPUs (Nvidia/AMD) |
| AI Performance (NPU) | Reportedly exceeds current Intel/AMD NPUs, meets Copilot+ certification (45+ TOPS) | Powerful NPU for on-device AI tasks | Integrated NPUs in newer generations (e.g., Intel Core Ultra, AMD Ryzen AI MAX+) |
| Manufacturing Process | TSMC 3nm | 4nm | Varies (e.g., Intel 4, TSMC for AMD) |
| Memory | Up to 128GB LPDDR5X (unified) | Unified LPDDR5X | DDR5/LPDDR5 (separate CPU/GPU memory for discrete) |
| Power (TDP) | 18-80W (N1: 18-45W, N1X: 45-80W) | 23W to 80W | Varies widely (e.g., 55W for Intel Core i9-13980HX) |
| x86 Emulation | Relies on x86 emulation layer for compatibility, potential gaming performance issues | Relies on x86 emulation layer (Prism optimized for Snapdragon) | Native x86 compatibility |
| Key Advantage | High-performance integrated GPU with full CUDA support, strong AI capabilities | Strong multi-core CPU performance, excellent battery life, established Windows on ARM presence | Broad software compatibility, established market dominance |
| Market Focus | Windows on ARM PCs (laptops, potentially Surface) | Windows on ARM PCs (laptops, Surface) | Traditional Windows PCs (laptops, desktops) |
๐ ๏ธ Technical Deep Dive
- Nvidia N1X Processor:
- Manufacturing Process: Built on TSMC's 3nm process.
- Package: 2.5D package integrating two distinct chiplets: a CPU die (designed by MediaTek) and a GPU die (Blackwell architecture).
- Interconnect: Chiplets connected via Nvidia's NVLink C2C interconnect, offering 300 GB/s bidirectional bandwidth.
- CPU Configuration: Features 20 ARM v9.2 cores, split into two clusters of 10 performance (Cortex-X925) and 10 efficiency (Cortex-A725) cores.
- L3 Cache: Shared 32 MB of L3 cache.
- GPU Configuration: Blackwell architecture with 6,144 CUDA cores across 48 Streaming Multiprocessors, matching the shader count of a desktop GeForce RTX 5070. Includes fifth-generation Tensor Cores with NVFP4 precision and dedicated ray tracing cores.
- Memory: Unified LPDDR5X memory pool, supporting up to 128GB across a 16-channel interface.
- Thermal Design Power (TDP): Configurable from 45W to 80W.
- Connectivity: Includes 12 PCIe 5.0 lanes and 5 PCIe 4.0 lanes.
- Storage: Supports up to three M.2 SSDs.
- Nvidia N1 Processor (Lower-end variants):
- CPU Configuration: Available in variants with 12 ARM CPU cores (8 performance, 4 efficiency) or 10 ARM CPU cores (7 performance, 3 efficiency).
- GPU Configuration: Features 20 Streaming Multiprocessors (2,560 CUDA cores, comparable to RTX 5050) or 16 Streaming Multiprocessors (2,048 CUDA cores).
- Memory: Unified LPDDR5X memory, supporting up to 64GB with an 8-channel setup.
- Thermal Design Power (TDP): Configurable from 18W to 45W.
- Connectivity: Includes 8 PCIe 5.0 lanes and 3 PCIe 4.0 lanes.
- Storage: Supports up to two M.2 SSDs.
- Shared Foundation: The N1X is reportedly based on the GB10 Superchip, which was previously used in Nvidia's DGX Spark mini PC for AI workloads.
๐ฎ Future ImplicationsAI analysis grounded in cited sources
Nvidia's entry will significantly intensify competition in the Windows on ARM PC market.
With Nvidia's strong GPU and AI capabilities, it will directly challenge Qualcomm's established position and potentially attract more OEMs and developers to the ARM ecosystem, fostering innovation.
Gaming performance on Windows on ARM PCs may remain a significant challenge despite powerful Nvidia GPUs.
The reliance on x86 emulation layers for most PC games can introduce performance degradation, and current emulation solutions are primarily optimized for Qualcomm chips, potentially limiting the N1X's appeal to gamers.
The strategic partnership with Microsoft and potential integration into the Surface lineup could accelerate the mainstream adoption of AI PCs.
Microsoft's push for 'Copilot+ PC' and on-device AI agents gains substantial validation and hardware capability with Nvidia's participation, potentially making ARM-based AI PCs a more viable and attractive option.
โณ Timeline
2011-01
Nvidia announces 'Project Denver' to develop custom 64-bit ARMv8-A CPU cores, hinting at PC and supercomputing applications and potential Microsoft collaboration.
2012-10
Microsoft launches Windows RT, an ARMv7-based version of Windows 8, with Nvidia's Tegra 3 powering devices like the original Surface RT.
2014-10
Google's Nexus 9, featuring Nvidia's dual-core Tegra K1 SoC with Denver CPU cores, is announced as the first 64-bit Android device.
2020-09
Nvidia announces its intent to acquire ARM Holdings to unite Nvidia's AI computing with ARM's CPU reach across various markets, including PCs (acquisition later abandoned in 2022).
2024-01
Nvidia reveals the AI-focused DGX Spark mini PC, powered by its GB10 Superchip, which CEO Jensen Huang later confirms uses the same core silicon as the N1.
2026-05-29
Microsoft, Nvidia, and ARM post coordinated social media teasers ('A new era of PC') pointing to Computex, intensifying speculation about Nvidia's ARM PC chips.
๐ Sources (23)
Factual claims are grounded in the sources below. Forward-looking analysis is AI-generated interpretation.
๐ฐ
Weekly AI Recap
Read this week's curated digest of top AI events โ
๐Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: The Verge โ