Nvidia unveiled the DSX Pod AI Factory, which packs seven Vera Rubin AI racks into a container about the size of a standard shipping container, weighing roughly 62,000 pounds and transportable by truck. Cabling and liquid cooling are built at the factory, so on arrival it only needs to be placed, connected and tested, cutting the time for a new AI data center to start serving tokens from six months to 14 days. As a reference design, it will be built by partners including Foxconn, Gigabyte, Lenovo, HPE, Quanta and Pegatron, with deliveries expected to start in early 2027. CEO Jensen Huang said the value Nvidia offers per gigawatt of data center power has risen from about $18 billion to more than $40 billion, and he expects it to keep climbing.
Summary written by Upside AI from the original article. Read the source for full details.
Read the original on Yahoo Finance ↗
Keywords
NVDAdata centersartificial intelligence
Open a keyword to see other news on the same topic.
Related tickers and likely impact
NVDA
▲ Likely positiveNvidia
Relevance: low
| Market | US stocks |
|---|---|
| Sector | Information Technology |
| Industry | Semiconductors |
| Market cap | $5.2T |
| Close on 10/20/2026 | 216.48 |
| Change that session | -0.28% |
Impact is scored by AI from the headline and may be wrong. It is not investment advice.
Related news
- Nvidia's Jensen Huang Sees 100GW AI Opportunity In Telecom Networks, With TMUS, NOK Already Making Moves
Nvidia CEO Jensen Huang told GTC Berlin that telecom networks are becoming a new layer of AI infrastructure, as distributed cell sites and data centers could run AI workloads closer to users. He envisions 100,000 sites representing some 100 gigawatts of base stations and data centers evolving into a distributed AI network. Huang said Nvidia and Nokia completed the first commercial phone call on T-Mobile's network after nearly eight years of work, and that nearly 20 operators have announced AI-RAN collaborations with commercial services targeted for 2027. In pre-market trade, NVDA rose as much as 1.7% and NOK gained 3.5%, while TMUS slid around 3% amid telecom sector weakness after SpaceX earnings. Bernstein cut its T-Mobile target to $180 from $220, citing pressure from SpaceX's plan to turn Starlink into a major U.S. mobile carrier.
- Nvidia Aims To Speed Deployment Of AI Factories
- Nvidia GTC Berlin Keynote: Jensen Huang Puts $50 Billion Price Tag On 1GW AI Factory Amid DSX POD Launch
Nvidia shares rose in early trade after CEO Jensen Huang unveiled DSX POD, a system designed to speed up getting AI data centers online. Huang said a 1-gigawatt facility costs about $50 billion and can generate roughly $40 billion in annual revenue, so a three-month delay could push back about $10 billion in revenue. DSX POD shifts more assembly and testing from construction sites to the supply chain, which Nvidia says can cut deployment time at a prepared site from six months to as little as 14 days, with units expected in early 2027. Nvidia supplies the reference architecture while partners including Dell, HPE, Lenovo, Vertiv and Foxconn build the systems. The article also notes financing concerns, as CoreWeave's bond yields approached 15% last week.
- Nvidia's Jensen Huang wants to build AI factories in weeks, not months
Nvidia CEO Jensen Huang, speaking at Nvidia GTC Berlin, said the new DGX Pod is designed to build AI factories faster and more cheaply, potentially cutting deployment from months to weeks. He said a gigawatt of data center costs about $50 billion, and the 4 to 6 months from shipping components to operation could be shortened by 3 to 4 months if more manufacturing happens in the supply chain, reducing idle cash. Huang argued this would make financing easier because lenders would have more certainty about costs and when revenue starts, which matters amid market volatility over neocloud companies. He said demand is very strong, citing SpaceX, CoreWeave, Nebius and Nscale, and that the large capital tied up before revenue is the key challenge to solve.
- Nvidia announces plug and play Vera Rubin racks that can be shipped on a truck
Nvidia announced the DSX Pod AI Factory, which puts seven Vera Rubin AI racks into a unit roughly the size of a standard shipping container and weighing about 28,000 kilograms. It is designed to be trucked on normal roads, with cabling and liquid cooling assembled at the factory so that on-site work is limited to placing, connecting and validating the units. Nvidia says this cuts the time for a new AI data center to start serving tokens from six months to 14 days. No price was disclosed, though analysts say a single Vera Rubin rack can cost over $7 million. The pod is a reference design built by partners including Foxconn, Gigabyte, Lenovo, HPE, Quanta and Pegatron, with deliveries expected in early 2027.
- Nvidia reveals plans for 'an AI factory in a shipping container' to build factories in weeks, not months
At its GTC event in Berlin, Nvidia (NVDA) unveiled the DSX POD, a reference design for an AI data center packed into a box about the size of a standard shipping container. Vera Rubin-powered DSX PODs will arrive on site ready to plug into power, cooling and networking, reducing deployment time from six months to 14 days. Nvidia says the units won't replace traditional data centers but let customers start running AI applications before brick-and-mortar facilities are ready. The DSX POD supports liquid cooling and, using dry-cooler designs, can save up to 54 million liters of water per day versus a standard data center. Nvidia is working with Anduril, Armada, Bull, Crusoe, Dell Technologies (DELL), Foxconn and others to begin delivery in early 2027, while the article notes that power supply is another bottleneck for data center builders.
Related markets
Latest news
- CoreWeave debt tumbles anew as Korean neocloud IPO tanks
CoreWeave's debt fell again on October 20 as investors reacted poorly to the public-market debut of Elice Group, a South Korean cloud company often called the South Korean equivalent of CoreWeave. Despite being almost 400 times oversubscribed and pricing at the top of its range, Elice's shares plunged over 32% in their KOSDAQ debut. The article says this is another sign that the neocloud business model might be broken. Elice locked in its IPO pricing on October 6, a day before macro anxiety and capex concerns.
- The bull case for CoreWeave as GPU demand surges
In the Loop host Ejaaz Ahamadeen makes the bull case for CoreWeave (CRWV), arguing that demand is pushing secondhand GPU prices higher. He cites comments from Amazon CEO Andy Jassy on demand for AI infrastructure and says AI capex differs from the dotcom era because buyers are willing to pay premiums. According to the host, secondhand GPU prices are at their highest ever, with older GPUs now worth twice their original price despite being 5 to 10 years old. He also argues CoreWeave's financing is not siloed, since customers such as Microsoft, Meta and Google can obtain attractive financing rates. He cites Elon Musk saying talks to buy compute from neoclouds have been the hardest ever, which the host attributes to providers having already pre-sold capacity at higher premiums.
- Where is Optimus? Key questions for Tesla earnings.
Hosts of The 8:30 preview Tesla's (TSLA) earnings report, where investors want updates on the Optimus robot, robotaxi expansion beyond Austin and SF, and capital expenditure. The key figures cited are revenue of $28.13 billion and EPS of 44 cents. Tesla announced capex of about $25 billion last quarter, and the hosts question whether it will go higher. The stock rose only about 1.7% this morning despite news of a possible SpaceX merger, and is still down 17% year to date after a strong rebound from its lows.
- SpaceX stock soars to its best day in two months while space peers climb out of a 26% hole
SpaceX (SPCX) opened with its biggest gain since its June debut and is on pace for its best day in two months. The rally follows third-quarter revenue and profit that beat Wall Street estimates, with the AI business now generating nearly half of sales. CEO Elon Musk said SpaceX and Tesla (TSLA) have each formed committees to study a merger. Through Tuesday, SpaceX was down just 2% in October, versus a drop of about 26% for the typical newer space stock. Chris Versace of Tematica Research said the planned $35B to $40B of fourth-quarter spending implies higher capex than the market expected, though it is backed by a $63B backlog, and up to 1.3 billion shares unlock on Thursday.
- Super Micro Soars 12% as Preliminary Revenue Blows Past Guidance on “Increased Demand for AI Platforms”
Super Micro (SMCI) jumped 12% after saying preliminary fiscal first-quarter revenue would be $18.5 billion to $19.5 billion, well above its earlier guidance of $14.5 billion to $15.5 billion. It also expects gross margin of 12.0% to 12.5%, above the 10.4% to 10.8% it had guided, which the article says matters more for earnings. Dell, HPE and Vertiv also gained, and Amkor rose after a $1.3 billion investment in South Korea. However, Super Micro burned nearly $7 billion of free cash flow last fiscal year, so its November 3 earnings call needs to show the bigger quarter turning into cash. The figures are preliminary and unaudited.
- Meta earnings preview: CapEx and Muse are in the spotlight
The 8:30 hosts preview Meta's (META) earnings report due tonight, highlighting capital spending and adoption of its Muse AI agent, including downloads and retention, as the two main items analysts will watch. Capex for the quarter is cited at about $41 billion. Consensus expectations are EPS of $3.56 and revenue of $63.3 billion. Options now imply a one-day move of about 6.6%, up from about 3.2% earlier. The stock surged after the Muse news but has since drifted lower along with the broader AI trade.
Get the briefing on Telegram or Zalo
Join the Upside Telegram channel or Zalo group for the Vietnam stocks and gold briefing at 8:00 and 19:00 (Vietnam time) on trading days, plus a weekly wrap on Saturday morning.