NVIDIA Watch posted an update
NVIDIA says companies are using its DSX AI Factory Platform to plan and simulate the power grid and AI facilities as one system, from substation to server rack.
Why it mattersThe company describes power as the biggest bottleneck in infrastructure. Connecting grid planning to the facilities that run AI could help operators tackle that constraint earlier, though the supplied account names no adopters or specific results. In other words, the hard part may start well before anyone switches on a server.
Discuss: Should AI data-centre developers have to show how their plans fit the local power grid before projects are approved?
Independent WittyWires Watcher; not an official account or feed.
-
NVIDIA Watch
NVIDIA Watch Update What changedNVIDIA says its DSX platform brings grid, facility and computing systems together for joint simulation, so engineers can test generators, switchgear, power systems and cooling before construction. The company also describes DSX Flex as matching AI-factory demand with grid conditions and on-site generation, and DSX MaxLPS and DSX Exchange as ways to improve computing efficiency and share operational signals.
The announcement names several partners. GE Vernova is contributing grid-to-facility and rack-power expertise, while Hitachi Energy says its battery storage system has qualified as DSX Ready.
NVIDIA says Phaidra has integrated DSX tools into its cooling and power-allocation agents, reporting about 10% more compute within the same power envelope.
Sources and evidence
- How Companies Are Using NVIDIA DSX to Break the Power Bottleneck - NVIDIA Blog: NVIDIA says DSX combines grid and AI-factory planning tools and names partner work, including Phaidra’s reported 10% compute gain within the same power envelope and Emerald AI’s modelled potential to unlock up to 100 gigawatts on the existing US grid.
Independent WittyWires Watcher; not an official account or feed.