NVIDIA Unveils DSX AI Factory Framework at Computex 2026

NVIDIA CEO Jensen Huang unveiled the company's DSX infrastructure framework at Computex 2026, presenting a blueprint for building gigawatt-scale AI factories alongside a roster of ecosystem partners. During his June 2, 2026 keynote, Huang called Marvell Technology "the next trillion-dollar company," triggering a 241% year-to-date stock gain for the semiconductor firm. The DSX framework represents NVIDIA's expansion beyond chip sales into comprehensive AI factory design, with the company projecting 100 gigawatts of such facilities by 2030. NVIDIA introduced DSX as a response to the complexity of modern AI infrastructure projects, which now cost between $30 billion and $100 billion per gigawatt of capacity. The announcement reflects the semiconductor industry's shift toward integrated infrastructure solutions as hyperscalers race to deploy large-scale AI computing facilities.

Marvell Technology Gains 241% YTD Following Huang Endorsement

Marvell Technology stock rose approximately 241% year-to-date following Huang's June 2, 2026 statement at Computex. The stock traded between $50 and $100 as recently as April 2026 and reached around $300 by mid-June, with an all-time high near $316. According to the analysis published June 18, 2026 by Ziven.io, the company re-rated by roughly a quarter-trillion dollars following the single endorsement. The stock movement prompted investor interest in identifying other companies Huang might highlight in future presentations.

NVIDIA Introduces Three-Tier Framework: RTX, DGX, and DSX

Huang outlined NVIDIA's product hierarchy during his Computex keynote. RTX represents the GPU level, described as "the worker" performing individual computing tasks. DGX constitutes the system level, described as "the team," combining multiple chips into a single machine. DSX forms the infrastructure level, described as "the factory," encompassing power, cooling, networking, and software required to operate thousands of systems continuously. Huang stated that DSX is "a blueprint, a reference design for building and operating AI factories at maximum efficiency and profitability." The DSX toolkit includes DSXSim for digital twin design and testing, DSX OS for facility operation, and DSX Max LPS and DSX FLEX for power optimization and grid flexibility.

DSX Ecosystem Includes Dozens of Infrastructure Partners

NVIDIA published a roster of partners designated as the "AI Factory Ecosystem" to cover infrastructure layers beyond semiconductor manufacturing. The coalition includes companies handling construction, high-voltage electrical work, cooling systems, and utility negotiations. The ecosystem spans both publicly traded and private companies across multiple countries. According to the source analysis, U.S.-listed partners include firms with varying degrees of AI infrastructure exposure, from companies where AI factory demand represents their entire business model to diversified industrial conglomerates where AI infrastructure constitutes a minor revenue segment. Private companies with planned IPOs mentioned in the ecosystem include Lambda (U.S.), Nscale (U.K.), Firmus (Australia), and Yotta (India).

AI Factory Projects Cost $30-100 Billion Per Gigawatt

Huang stated that a one-gigawatt AI factory now represents a $30 billion to $100 billion project. At this scale, the facilities function as infrastructure comparable to refineries or power stations rather than traditional data centers. NVIDIA does not handle construction, high-voltage electrical installation, chiller manufacturing, or utility negotiations directly. The company positioned DSX as a coordination framework because components including chips, racks, networking, power, and cooling require integrated design from project inception. NVIDIA projects that 100 gigawatts of AI factory capacity will come online before 2030.

FAQ

What did Jensen Huang announce at Computex 2026? Jensen Huang announced NVIDIA's DSX infrastructure framework at Computex 2026, presenting a blueprint for building gigawatt-scale AI factories. During his June 2, 2026 keynote, he also called Marvell Technology "the next trillion-dollar company" and outlined NVIDIA's three-tier product hierarchy of RTX, DGX, and DSX.

How much did Marvell Technology stock gain following Huang's statement? Marvell Technology stock gained approximately 241% year-to-date following Huang's June 2, 2026 endorsement. The stock traded between $50 and $100 in April 2026 and reached around $300 by mid-June, with an all-time high near $316.

What does NVIDIA's DSX framework include? NVIDIA's DSX framework includes DSXSim for digital twin design and testing, DSX OS for operating AI factories, and DSX Max LPS and DSX FLEX for power optimization and grid flexibility. Huang described DSX as "a blueprint, a reference design for building and operating AI factories at maximum efficiency and profitability."

Disclaimer: The information on this page may come from third-party sources and is for reference only. It does not represent the views or opinions of Gate and does not constitute any financial, investment, or legal advice. Virtual asset trading involves high risk. Please do not rely solely on the information on this page when making decisions. For details, see the Disclaimer.
Comment
0/400
No comments