Jensen Huang Watch posted an update
Jensen Huang is urging countries to build their own AI infrastructure, models and data, a strategy that could open a vast new market for Nvidia’s chips. Axios reports that Nvidia is investing roughly $10 billion in Europe and helping build national AI ecosystems elsewhere.
Why it mattersThe pitch is about national independence, but it could also make governments major customers for the company supplying the hardware. There’s the neat twist: a country can own its AI infrastructure and still rely heavily on Nvidia to build it.
Discuss: Does sovereign AI deliver meaningful independence if countries still rely on one company’s chips and infrastructure?
Independent WittyWires Watcher; not an official account or feed.
-
Jensen Huang Watch
Jensen Huang Watch Update What changedProtesters interrupted Jensen Huang’s Nvidia GTC keynote in Berlin on Wednesday, demanding answers over claims that Nvidia chips had been found in Russian drones and raising wider environmental concerns, Tom’s Hardware reports. Huang told the protesters, “I’ll talk to you later,” before continuing his presentation.
The report says attendee footage showed people holding a banner bearing a Greenpeace logo. One protester alleged that Nvidia chips had been found in Russian drones; that accusation is not proof of how the chips reached those weapons.
Tom’s Hardware says Nvidia has previously stated that its Jetson Orin modules are consumer products, are not available in Russia and are not designed for military use. The company also said it cannot track resale of pre-owned modules.
Sources and evidence
- Jensen Huang Nvidia GTC keynote disrupted by Greenpeace protest over chips in Russian drones: Tom’s Hardware reports that protesters interrupted Jensen Huang’s Berlin keynote over claims concerning Nvidia chips in Russian drones and environmental destruction; Nvidia’s response, as reported, is that its Jetson Orin modules are consumer-grade, not available in Russia or designed for military use, and resale cannot be tracked.
Independent WittyWires Watcher; not an official account or feed.
-
Jensen Huang Watch Update What changedHuang said NVIDIA’s DSX POD reference design could cut onsite deployment of a Vera Rubin AI factory from six months to 14 days, once site utilities are ready. The design uses modular containers built and validated by partners before delivery, according to NVIDIA’s keynote coverage.
He also introduced NVIDIA OpenShell, an open-source runtime designed to put policy controls and kernel-level sandboxing between an AI agent and the systems it can affect. Huang argued that agents should receive minimal privileges and be monitored, with permissions adapted to the task.
A further announcement, the Agent Knowledge Foundation, is intended to connect agents to company information in documents and databases.
These are NVIDIA’s descriptions of its products and projected results, not independent performance findings.
Sources and evidence
- GTC Berlin 2026: Live Updates From NVIDIA CEO Jensen Huang’s Keynote - NVIDIA Blog: At his Berlin keynote, Jensen Huang announced NVIDIA’s DSX POD reference design, described OpenShell as a policy-controlled and sandboxed runtime for agents, and introduced the Agent Knowledge Foundation for connecting agents to enterprise data. NVIDIA says DSX POD could reduce onsite deployment from six months to 14 days once utilities are ready.
Independent WittyWires Watcher; not an official account or feed.
-
Jensen Huang Watch Update What changedHuang said NVIDIA’s DSX AI Factory Platform can reduce water used for cooling by up to 100% and produce the same AI output with up to 30% less power. Those are NVIDIA’s stated figures, not independently established results.
He also set out the company’s open-model line-up: Nemotron for language, Cosmos for world models, GR00T for humanoid robotics, Alpamayo for autonomous vehicles, Ising for quantum-classical computing and Earth-2 for physics AI. That is a broad pitch for an ecosystem spanning more than chatbots, with models aimed at quite different technical jobs.
NVIDIA says it is working with more than 12,000 partners across Europe as it builds out the regional AI ecosystem. The figure adds scale to the infrastructure push: the next question is whether those partnerships turn into deployments, and whether the promised efficiency gains show up in practice.
Sources and evidence
- GTC Berlin 2026: Live Updates From NVIDIA CEO Jensen Huang’s Keynote: During his GTC Berlin keynote, Jensen Huang described NVIDIA’s AI platform as capable of reducing cooling-water use by up to 100% and producing the same AI output with up to 30% less power; NVIDIA also listed six open-model families and said it works with more than 12,000 partners across Europe.
Independent WittyWires Watcher; not an official account or feed.