NVIDIA称Vera Rubin AI平台已量产并在客户数据中心运行

이전 기사보기
다음 기사보기
NVIDIA Begins Shipping Vera Rubin AI Platform to Customers
Character Size Settings
가
The body content of the article will be changed to this letter size.
스크롤 이동 상태바
Semiconductor
Next-generation rack-scale system delivers up to 10x higher AI throughput per megawatt than Blackwell
Lee Jun
Published 2026.07.23 07:42
Reply 0
이 기사를 공유합니다
Nvidia Vera Rubin AI platform. (Photo: Nvidia)
Nvidia has begun shipping its next-generation Vera Rubin AI platform to customers, marking the commercial rollout of the company's latest rack-scale computing architecture for AI factories.
The company announced on July 21 that Vera Rubin has entered volume production and is already running in customer data centers. Nvidia had previously announced Vera Rubin's entry into mass production on June 1. Customers include Google, Microsoft, Oracle and CoreWeave.
Vera Rubin is a rack-scale AI infrastructure platform that tightly integrates 36 Nvidia Vera CPUs, 72 Rubin GPUs, NVLink 6 interconnects and storage into a single system. The compute tray eliminates internal cables, fans and hoses, reducing assembly time from several hours to about one minute.
The Vera CPU is Nvidia's first server processor designed in-house specifically for AI infrastructure. It features 88 custom Olympus CPU cores and is designed to deliver up to 1.8x the performance of conventional x86 processors while providing three times greater inter-core bandwidth and 40% lower memory latency.
Nvidia said Vera Rubin significantly reduces both power consumption and AI inference costs compared with the previous generation. Benchmark testing on CoreWeave's cloud platform showed the system delivers 10 times more AI throughput per megawatt than Grace Blackwell NVL72, substantially lowering the cost per generated token. The company said tokens per megawatt has become a key metric for AI infrastructure because it measures how much AI work can be performed within a fixed power budget.
The platform also supports liquid cooling with coolant inlet temperatures of up to 45 degrees Celsius, enabling chiller-free dry cooling in suitable environments. Nvidia said the design can save millions of gallons of water annually per megawatt of deployed computing capacity.
2jun0504@thelec.kr
View other news
Copyright © The Elec Inc. No unauthorized transfer and redistribution
Comments
Name
Password
Enter comments
0
/ 400
Submit
Sort the comments.
Newest
Recommended
Top Comments
Best comment
It is automatically exposed by adding up the number of replies and recommendations.
close
더보기
Delete comments
Deleted comments cannot be recovered.
Do you still want to delete it?
Delete
Modify the comments
You can modify comments only within 1 minute after writing.
Text
Modify
My comment
Most Read
1
[Opinion] Mass Production Is Only the Beginning for Solid-State Batteries
2
LG Energy Solution Expands Tesla Battery Production at Nanjing Plant
3
SK hynix Prices ADR Offering at $149, Raising Record $26.51 Billion
4
ASML Posts Second-Quarter Earnings Beat on Strong AI Demand, Raises Full-Year Outlook
5
Micron Raises US Semiconductor Investment to $250 Billion, Breaks Ground on New York Fab
6
US Commerce Department Says Nvidia's H200 Shipments to China Have Begun
7
Samsung Electronics, Samsung Display Join Forces on Glass Interposer Technology
8
Samsung Wins Tesla AI5 Chip for 2nm Production
9
Samsung Weighs Outsourcing Back-End Design for Google's 2nm TPU I/O Die
10
Intel Foundry Deploys High-NA EUV Tools in High-Volume Production
Editor’s Picks
Samsung Plans 50-Tool Hybrid Bonding Production Line in Pyeongtaek
LG Electronics to Shut Down Its First China Manufacturing Plant After 34 Years
Samsung-Backed Pino Seeks Capital Raise as It Prepares for US Battery Rules
Google to Open Play Store to Third-Party App Stores in US
CMTX Produces Aftermarket Single-Crystal Silicon Ground Ring for Semiconductor Equipment
More Stories
Display Industry Calls for Government Support to Foster Inorganic Light-Emitting Display Sector
Government Approves Lotte, Hanwha and DL's Yeosu Petrochemical Restructuring
SK Hynix Accelerates Cheongju P&T7 Project, Reapproves 7.1 Trillion Won Investment
Samsung Expands Galaxy Ecosystem Beyond Foldables With AI Smart Glasses