NVIDIA Releases Modular AI Factory DSX POD, Production in as Fast as 14 Days
NVIDIA CEO Jensen Huang unveiled the modular AI factory DSX POD at GTC Berlin, with plans for global delivery in early 2027. The product completes computing, networking, cabling, piping installation, rack integration, and functional testing in the factory in advance; once the container arrives on site, it can operate after connecting to power, cooling, and network infrastructure. NVIDIA claims it can compress deployment time from about 6 months under the traditional approach to as short as 14 days.
Why read this
AI score 60
The original text provides a comparison of delivery time and deployment cycle for modular AI factories, which readers can use to gauge changes in the pace of data center construction.
Topics
Body · AI translation
Related topics top start Related topics top end Futures promotion begin Futures promotion end Miaopai begin Miaopai end Video player start Video player end Market chart begin Market chart end
21st Century Business Herald, Peng Xin
On October 21, Nvidia CEO Jensen Huang unveiled the modular AI factory DSX POD at GTC in Berlin, with plans to deliver it globally in early 2027.
DSX POD moves computing, networking, cabling, piping installation, rack integration and functional testing—work that previously had to be done on site—entirely into the factory for integration and validation in advance. Once the container arrives on site, it only needs to be connected to ready power, cooling and network infrastructure to begin operation.
Nvidia says DSX POD can compress deployment time from about 6 months under the traditional approach to as little as 14 days, thereby accelerating customers' speed from infrastructure readiness to producing production-grade tokens. news_keyword_pub,stock, iframe below the main text iframe below the main text General statement General statement QR code at the end start QR code at the end start
QR code at the end start QR code at the end start Editor name and work code
Editor name and work code