Jensen Huang in GTC Berlin said that the Vera Rubin platform has entered full production
BIBIBI
AT A GLANCE
PANews said that Jensen Huang, in GTC Berlin 2026 announced that Vera Rubin has entered full production and released DSX POD designs.
Article
Jensen Huang, Berlin GTC Speech: NVIDIA shifts from “selling GPU” to “selling AI factories,”Vera Rubin enters full production
PANews October 22 reported, citing the NVIDIA official blog, that NVIDIA CEO Jensen Huang October 21 delivered a speech at Berlin’s Tempodrom concert hall GTC Berlin 2026 keynote (approximately 2 hours), with the central theme that “Europe should shift from AI consumers to AI producers,AI and factories are the next-generation infrastructure.” Jensen Huang said that “AI is driving the largest infrastructure buildout in human history,” comparing AI it to a “five-layer cake”: energy → compute →AI factories → models → applications. Jensen Huang announced that the Vera Rubin platform has entered full production; the architecture is 100%liquid-cooled, with a single rack connecting 72 units GPU into a supercomputer, with 2 performance expected to increase further within the year by 30 times (beyond Moore’s Law). It also released DSX POD modular AI factory reference designs—complete pre-integrated AI factories in containers, for on-site deployment...
PANews said that its report was based on the NVIDIA official blog.
02
The report stated that NVIDIA CEO Jensen Huang October 21 delivered a speech at Berlin’s Tempodrom concert hall GTC Berlin 2026 keynote, which lasted approximately 2 hours.
03
Jensen Huang said that Europe should shift from AI consumers to AI producers,AI and factories are the next-generation infrastructure.
04
Jensen Huang compared AI it to a “five-layer cake” composed of energy, compute,AI factories, models, and applications.
05
Jensen Huang announced that the Vera Rubin platform has entered full production; the report described the architecture as 100%liquid-cooled, with a single rack connecting 72 units GPU。
06
The report stated that Vera Rubin is expected to 2 further increase performance within the year by 30 times.
07
Jensen Huang also released DSX POD modular AI factory reference designs, which the report described as complete pre-integrated AI factories in containers.
AI-assisted interpretation
The following is analysis, separate from reported facts. Verify important claims independently.
This news describes an infrastructure approach that integrates energy, computing equipment, models, and applications into so-called “AI factories.”Vera Rubin is the platform announced during the speech as having entered full production,DSX POD is referred to as a modular AI factory reference design.
Why it matters to readers
The report characterized NVIDIA’s focus as a more complete AI infrastructure solution, rather than merely standalone GPU; production status, rack scale, and performance improvements are noteworthy product information.
Beginners should distinguish among concepts:GPU is a computing chip,Vera Rubin is a platform,AI and a factory is the term for an overall infrastructure comprising energy, compute, models, and applications.
Risks and unknowns
The supplied text is truncated after “on-site deployment,” so the complete release details cannot be confirmed.
The evidence does not provide the NVIDIA original official blog link, so key statements such as “entering full production” cannot be independently verified based solely on this material.
“2 The methodology, comparison baseline, and conditions for achieving the “further performance increase within the year by 30 times” are not explained in the supplied text.
DSX POD partners, actual deployments, shipments, and commercial contribution are not explained in the supplied text.
Related Developments
Loading event timeline…
Related concepts
AI Factories
This term is not in the glossary yet. Browse related concepts in the glossary.
