At GTC Berlin, Nvidia CEO Jensen Huang the $50 billion estimate includes land, power, computing equipment, networking, storage, cooling and other infrastructure.
- According to Huang, a 1GW facility could generate about $40 billion in annual revenue.
- The Nvidia CEO forecast that delays could cost operators roughly $10 billion in foregone revenue.
- Nvidia aims to cut the time required to move equipment through the supply chain and begin producing AI tokens from roughly four to five months to about one month with DSX POD.
Nvidia (NVDA) shares rose in early morning trade on Wednesday after CEO Jensen Huang unveiled DSX POD, a new system designed to get AI data centers up and running faster. Huang said a 1-gigawatt facility costs about $50 billion, making delays costly for operators.
“Each gigawatt data center is $50 billion.”
– Jensen Huang, CEO, Nvidia
Such a facility can generate about $40 billion in annual revenue, he said, meaning a three-month delay could push back roughly $10 billion in revenue. “10 billion worth of revenues will make anybody anxious,” Huang said.
DSX POD is Nvidia’s latest effort to speed up the deployment of AI computing infrastructure by shifting more assembly and testing from construction sites to the supply chain. Nvidia says the approach can cut deployment time at a prepared site from six months to as little as 14 days. Units are expected to be available in early 2027.
NVDA stock rose as much as 2% in pre-market trade and was among the top trending tickers on Stocktwits at the time of writing. Retail sentiment around the AI bellwether shifted to ‘bullish’ from ‘neutral’ territory over the past day.
Nvidia Targets Idle AI Computing Capacity
Huang said the system builds on Nvidia’s earlier move to assemble its NVLink 72 rack-scale computers within the supply chain rather than on-site. DSX POD extends that approach to a larger system, integrating computing, networking, storage, software, power and cooling into a single AI factory reference design.
“Great chips alone don’t make a great AI factory,” Huang said. “NVIDIA DSX has a full-stack AI factory reference architecture, co-innovated with the world’s leading infrastructure companies.”
The aim is to reduce the work required after equipment arrives at a data center, when computing capacity may remain idle while systems are assembled, connected and tested.
“The wasted idle compute, the wasted idle capex will reduce, and everybody becomes more profitable.”
– Jensen Huang, CEO, Nvidia
Huang said moving more of that work upstream could reduce the time between the arrival of computing equipment and the start of token production, when AI systems begin generating output that customers can use or pay for.
“We think something along the lines of four to five months, down to about a month,” he said.
$50B Data Centers Raise The Stakes For Operators
Huang’s $50 billion estimate for a 1-gigawatt facility includes land, power, the building shell, computing equipment, networking, storage, power systems and cooling.
The scale of that investment is particularly important for companies financing expansion with debt. CoreWeave (CRWV), Nebius (NBIS) and IREN (IREN) are among the companies building out AI computing capacity.
However, the industry’s spending plans are unfolding alongside rising concerns about financing costs and the time required to turn capital spending into revenue. CoreWeave’s bond yields approached 15% last week, while CRWV stock dropped below the $60 mark in morning trade on Tuesday before recovering.
CRWV stock and NBIS stock were up around 2% each in pre-market trade, while IREN stock gained more than 3%. On Stocktwits, retail sentiment around all three neoclouds was trending in ‘extremely bullish’ territory over the past day.
Nvidia Enlists Global Partners For DSX POD
Nvidia is working with a broad group of technology and infrastructure companies on the design, including Dell Technologies (DELL), Hewlett Packard Enterprise (HPE), Lenovo, Vertiv (VRT), Schneider Electric, Eaton (ETN), Pegatron, Wistron, Foxconn, Crusoe and DayOne. Nvidia supplies the reference architecture, while its partners build the systems.
Huang said Nvidia had invested nearly $10 billion in Europe over the past year. The DSX POD rollout adds to the company’s efforts to expand AI infrastructure in the region, with the broader goal of helping customers bring computing capacity online sooner and generate returns on their investments faster.
Read also: AI Trade Rotation Favors Mag 7 Over Neoclouds, Analysts Say Ahead Of SPCX, META, NFLX Results
For updates and corrections, email newsroom[at]stocktwits[dot]com.<