The popularity and unpopularity of artificial intelligence (AI) varies among individuals. Nvidia, with its massive market cap of $5.2 trillion, has significantly profited from the AI boom. Consequently, the company is striving to convince those who doubt the technology.
In the United States, the Trump administration generally supports AI and disapproves of excessive government regulation, preferring a self-regulatory approach that fosters innovation. In contrast, the European Union has a prolonged background of enacting laws to regulate the technology industry.
In Europe, AI safety is a central worry, especially considering recent accounts of AI systems from OpenAI and Anthropic malfunctioning. However, just like in other countries, even though governments generally support the expansion of data centers for their AI projects, local communities are strongly against it.
They mainly have concerns about energy and water usage. In such a situation, how can a company that relies on the continuous growth of data centers build relationships and gain acceptance in an area where its objectives are conflicting with both the government and society’s views?
Last week at GTC Berlin, Nvidia attempted to achieve this… with mixed outcomes… 21 Mar 133. In the realm of AI, Nvidia’s CEO Jensen Huang has expressed that currently, there are three primary concerns: Energy, safety, and financial resources. Nvidia made several sustainability- and safety-focused announcements at GTC Berlin, which appeared to be a wise strategy.
The company claimed that its DSX platform can save water and energy. Nvidia, which now identifies as a full-stack platform provider, stated that its DSX Platform would enhance the efficiency of facilities in terms of water and power, enabling AI data centers to reduce the water consumption of cooling technologies by up to 100% and use 30% less power for the same output.
Saving 54 million liters (14.27 million gallons) of water daily might seem impressive, but it should be noted that this figure corresponds to a 2GW data center deployment, which doesn’t currently exist, even though many gigawatt-scale campuses have been announced worldwide. Nvidia has not released statistics regarding the amount of water the platform can conserve in data centers functioning at a mere-megawatt scale.
Additionally, the company introduced a reference architecture at the conference, which is based on its OpenShell software. The complete solution ensures that AI agents are protected by security policies, network restrictions, and privacy measures at the system level. It now also includes support for Vera Rubin infrastructure to handle business tasks, Vera CPUs to manage agent tools, and BlueField DPUs for running security software architectures.
