Jensen Huang unveils DSX POD containerized AI infrastructure at GTC Berlin

Nvidia CEO Jensen Huang unveiled DSX POD, a new reference design for containerized AI infrastructure, during his keynote address at GTC Berlin. The announcement came as environmental protesters briefly interrupted the presentation at the Tempodrom venue, demanding the tech giant address its environmental impact. Huang joked with the demonstrators before security removed them, allowing him to continue pitching his company's hardware and software vision.

The DSX POD combines seven liquid-cooled Vera Rubin compute racks using Nvidia’s Spectrum-X networking technology. Partners will cable, integrate, and test these systems in modular containers before delivery, with availability expected in early 2027. Nvidia claims that deployment at a prepared site could shrink from six months to as little as 14 days, though this timeline excludes the time required to secure power and connectivity infrastructure.

Context and Background

Huang emphasized that Vera Rubin systems are designed for closed-loop liquid cooling with coolant temperatures up to 45°C. Outdoor dry coolers release heat into the air without evaporating water, a feature Huang highlighted in response to the protesters. He claimed that at 2 gigawatts, the DSX system can save 54 million liters of water a day. This focus on efficiency aligns with his earlier statements at GTC Berlin, where he urged European industries to accelerate AI adoption to realize the technology's broader benefits.

Who Is Affected

The announcement targets datacenter operators and enterprises seeking rapid AI infrastructure deployment. Nvidia’s grand plan involves "AI trailer parks," or containerized datacenters, to speed up installations. This strategy supports the company’s $10 billion commitment to Europe’s technology sector, announced earlier in the keynote. The modular approach allows partners to pre-integrate systems, reducing on-site construction time significantly for organizations ready to adopt high-performance computing solutions.

What should you do?

Organizations planning AI infrastructure should evaluate their site readiness for rapid deployment models. The 14-day deployment claim assumes power and connectivity are already in place, so verifying utility capacity is critical. Companies should also consider the environmental benefits of liquid cooling systems, which may align with sustainability goals. Monitoring Nvidia’s partner ecosystem will be essential for accessing these pre-integrated containerized solutions when they become available in early 2027.

What to do and how to stay safe: Nvidia

• Audit your datacenter site’s power and connectivity capacity to determine if it meets the prerequisites for rapid modular deployment.
• Review your organization’s sustainability metrics to see if liquid-cooled, water-saving infrastructure aligns with corporate environmental goals.
• Monitor announcements from Nvidia’s partner network for integration timelines and availability of DSX POD systems in early 2027.
• Assess current deployment timelines for AI infrastructure to compare against the potential 14-day installation window offered by containerized solutions.

Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation

General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

What is DSX POD and when will it be available?

DSX POD is a reference design for factory-built, containerized AI infrastructure using Vera Rubin systems, with availability expected in early 2027.

How does Nvidia claim DSX POD reduces deployment time?

Nvidia claims deployment can shrink from six months to 14 days at a prepared site because partners pre-cable, integrate, and test systems in containers.

Did the protest interrupt the entire keynote?

No, environmental protesters briefly interrupted the presentation, but security removed them quickly, allowing Jensen Huang to continue his speech.