NVIDIAWindowsJensn Huang is planning to develop robots like R2-D2 in the future while already planning the second and third generation of SoC for laptops.

NVIDIAWindowsJensn Huang is planning to develop robots like R2-D2 in the future while already planning the second and third generation of SoC for laptops.

On June 1, 2026, the SoC “RTX Spark” for laptops that enable local execution of high-performance AI was announced by NVIDIA. NVIDIA CEO Gスn Huang talks about this SoC in a few interviews, and they want to create a system like Star Wars’ R2-D2. We are already working on next-generation chip development for that goal.

Nvidia J n Huang tells us about RTX Spark CPU Tom's Guide

https://www.tomsguide.com/computing/i-tomke-to-nvidia-ceo-j.n-huang-about-rtx-spark-he-is-willing-to-work-on-an-rtx-gaming-h.-n2x-and-n3x-are-already-planned-and-the-chip-is-more-like-r2d2-than-a-laptop-cpu

The goal is the Star Trek computer

https://www.theverge.com/tech/942588/nvidia-rtx-spark-n2x-n3x-r2-d2-star-trek-star-wars-plan

ChatGPTHomeClaudeHigh performance AI is deployed as a cloud service, and it is required to promote the development of AI that can be run locally in the field of dealing with sensitive and private information. However, it is highly efficient to run high-performance AI locallyGPUIt requires a large amount of memory. RTX Spark is a SoC developed in mind to run local AI, with up to 128GB of memory to run the AI model of 12 billion parameters.

NVIDIA Announces SoC “NVIDIA RTX Spark” for Laptops, Locally Run High Performance AI with 128GB Memory & NVIDIA GPU & Arm Base CPU - GIGAZINE

According to Huang CEO, RTX Spark is looking for the second generation and third generation of RTX Spark, not a single-generation product. Before the launch of RTX Spark, the device named "N1" and "N1X" was revealed. According to Huan CEO, RTX Spark is N1X, and N1 is smaller than RTX Spark.

RTX Spark, the second generation and third generation, is named N2X and N3X. The roadmap is also published on this page, and after this Blackwell generation, it has been shown to release the Rubin generation in 2027 and the Feynman generation in 2029.

Huang CEO wanted to make a system like R2-D2 in the future and make it possible to operate PC with voice. "If I want to talk to my laptop now, I have to wait until I return to my room. If you want to do something on your laptop in the future, just send a message with WhatsApp. R2-D2 and PowerPoint slides have problems. 17th slide. The size or title of the image is incorrect. CX10 instead of CX9 Then R2-D2 will open PowerPoint, modify it, make it PDF and send it to me. Can you imagine it? It’s simple, CEO Huang said.

Huang CEO is committed to local processing rather than the cloud, and there are no people who rent a washing machine. "It should be locally processed on a computer that is used recently." “In the future, computers can use AI.” Everything is AI. I also talked to the vacuum cleaner and added, “I will say that I wipe it out.”

The first generation of RTX Spark features up to 128GB of memory and has enough performance to hold an AI agent of 1200 billion parameters. However, SoC with 128GB of memory is inevitably expensive. Huang CEO also agrees to the question “Is it a power user to buy the first generation?”

NVIDIA’s new SoC “RTX Spark”Microsoft ASUS ProArt P16

Huang CEO will realize the realization of physical AI in a few years. “We believe that preparing a laptop for that purpose will be very important for the future of computing.”

・Re ・ articles

Microsoft announces "Surface RTX Spark Dev Box" with NVIDIA RTX Spark, and the operation performance of local AI is up to 1PFLOPS - GIGAZINE

NVIDIA Announces “Cosmos 3”, a physical AI-based model including high-performance image generation model and video generation model - GIGAZINE

Microsoft Surface Laptop Ultra and ASUS ProArt P16 with NVIDIA's new SoC "RTX Spark" revealed - GIGAZINE

NVIDIA Announces New Open Model “N tron 3 Ultra” in the U.S. - GIGAZINE

NVIDIA announces DGX Station, a monster-class Windows desktop PC with GB300, memory can run up to 1 trillion parameters AI with up to 748GB - GIGAZINE

NVIDIA Announces N1 and N1X Arm-based New Laptop Chips, or Aims for Low- Laptop and Gaming Notebooks - GIGAZINE