NVIDIA announced that its AI supercomputer Vera Rubin NVL72, edge AI platform Jetson Thor and open autonomous-driving foundation Alpamayo took four honors at the Best Choice Awards (BCA) of COMPUTEX, one of Asia's largest technology trade shows[1]. The company will also hold GTC Taipei from June 1 to 4 in Taipei, where founder and CEO Jensen Huang will deliver a keynote on the opening day[1].

Four awards at the COMPUTEX 2026 Best Choice Awards

The Best Choice Awards is a leading technology competition held alongside COMPUTEX, with entries evaluated on functionality, innovation and market potential[1]. This year NVIDIA was recognized across AI computing, integrated circuits and autonomous vehicle (AV) development, earning four awards in total[1].

The breakdown is as follows: the rack-scale AI supercomputer Vera Rubin NVL72 won a Golden Award and the Sustainable Tech Special Award; Jetson Thor, built for edge AI and robotics, won a Golden Award; and Alpamayo, an open platform for autonomous-driving development, won the Vehicle Technology and Smart Cockpit Category Award[1]. The recognition spans the full range from large data-center systems to AI running in the field, such as in robots and vehicles.

Rack-scale AI supercomputer: Vera Rubin NVL72

Vera Rubin NVL72, which won a Golden Award and the Sustainable Tech Special Award, connects 36 NVIDIA Vera CPUs and 72 NVIDIA Rubin GPUs through the sixth-generation NVIDIA NVLink Switch[1]. On the scale-out side, it combines ConnectX-9 SuperNICs and Spectrum-X Ethernet with BlueField-4 DPUs that handle storage and security processing[1].

In terms of performance, it delivers up to 10x higher inference performance per watt and can cut cost per token by up to a factor of 10[1]. Designed for agentic AI, reasoning and long-context workloads, it can scale intelligence from inside the rack to across the entire data center[1].

The energy- and labor-saving design that led to the Sustainable Tech Special Award is also notable. A cable-free, hose-free and fanless modular tray design reduces assembly time per compute tray from two hours to five minutes[1]. The power shelves store 6x more energy to smooth power fluctuations, protecting both the rack and the power grid from steep load swings[1]. Its 100 percent liquid-cooled architecture also operates at 45 degrees Celsius, so it drops easily into existing liquid-cooled data centers and lets power once used for cooling be redirected into token generation[1].

Edge AI and robotics: Jetson Thor

Jetson Thor, which received a Golden Award, is an edge AI compute platform built for physical AI and autonomous robots operating in the field[1]. Powered by the NVIDIA Blackwell GPU architecture, it packs up to 2,070 FP4 teraflops of AI performance into a compact module configurable between 40 and 130 watts[1].

Compared with the previous Jetson Orin generation, it reaches 7.5x the compute and 3.5x the energy efficiency[1]. Already in production across hundreds of applications, it brings generative AI to smart robots, industrial systems, medical devices and autonomous machines[1].

Open autonomous-driving foundation: Alpamayo

Alpamayo, which won the Vehicle Technology and Smart Cockpit Category Award, is an open platform for reasoning-based autonomous-driving development[1]. It targets complex situations that rarely appear in training data, such as how to read a pedestrian's ambiguous hand signal, which to prioritize when traffic lights and road markings contradict each other, and how to safely pass an emergency vehicle parked partly in the lane[1].

The platform includes Alpamayo 1.5 and Alpamayo 1, 10-billion-parameter chain-of-thought vision language action models for AV research; AlpaSim, an open-source end-to-end simulation environment for high-fidelity development; and NVIDIA Physical AI Open Datasets, which hold more than 1,700 hours of driving data across different regions and conditions[1].

GTC Taipei runs June 1 to 4 in Taipei

These products will be showcased in full at GTC Taipei, held from June 1 to 4 at the COMPUTEX venue in Taipei[1]. Jensen Huang is scheduled to deliver a keynote at 11 a.m. Taipei time on June 1[1].

Ahead of the event, NVIDIA unveiled a new Taipei campus called NVIDIA Constellation[1]. Located in the Beitou-Shilin Technology Park in northern Taipei, the site spans roughly 4 hectares and is designed to house about 4,000 employees; once operational, it will become one of the largest AI research and development hubs in the Asia-Pacific region[1]. Taipei Mayor Chiang Wan-an attended the announcement and presented Huang with a key to the city[1].

Huang pointed to the major growth opportunity created by today's agentic AI, and cited physical AI that will transform manufacturing as the next stage[1]. Speaking to the press, he described Vera Rubin as "probably the largest product launch in the history of Taiwan," explaining that a single system consists of nearly 2 million parts and is built with the cooperation of 150 ecosystem partners in Taiwan[1].

Summary

At the COMPUTEX 2026 Best Choice Awards, NVIDIA won four honors with three products: Vera Rubin NVL72, Jetson Thor and Alpamayo. What stands out this time is that the recognition spans not only large data-center AI systems but also AI in the field, including robotics and autonomous driving. The details are expected to become clear at GTC Taipei starting June 1, with attention focused on Jensen Huang's keynote.

出典:https://blogs.nvidia.com/blog/nvidia-gtc-taipei-computex-2026-news/