Nvidia’s Ambitious Plans for Supercomputers
Nvidia CEO Jensen Huang has announced the company’s intention to construct seven supercomputers for the Department of Energy, each featuring up to 100,000 chips, all manufactured in the United States. Huang mentioned that one of President Trump’s first requests was to “bring back manufacturing.”
During his keynote speech at Nvidia’s recent GTC event in Washington, DC, Huang shared several significant updates reflecting the company’s dedication to advancing AI technology. This was the first event of its kind in the U.S. capital, which underscores Nvidia’s commitment to working closely with local governments and contractors.
Huang detailed that these supercomputers will serve multiple functions, including managing and developing the U.S. nuclear arsenal and exploring alternative energy solutions like nuclear fusion. Notably, the largest of these supercomputers is set to be developed in partnership with Oracle and will utilize 100,000 Nvidia Blackwell chips.
In another remarkable highlight, Nvidia has reportedly secured a vast $500 billion in reservations for its Blackwell and Rubin chips over the upcoming five quarters, indicating a strong demand for its AI offerings. This news comes amid ongoing challenges posed by the U.S.-China trade war, which could influence the global tech landscape.
Additionally, Nvidia announced a new collaboration with Nokia, a Finnish communication equipment manufacturer, targeting the AI communications sector. The company plans to invest $1 billion to acquire a 2.9% stake in Nokia and launch a new product line called Arc, aimed at enhancing the energy efficiency of Nokia’s 6G base stations.
Moreover, Nvidia’s partnership with Palantir Technologies, which collaborates closely with the U.S. government, will focus on improving logistics for commercial clients like home improvement giant Lowe’s.
In the automotive realm, Nvidia introduced a new self-driving technology platform named Hyperion and revealed a partnership with Uber to develop a network of robotaxis. Huang expressed optimism about the potential success of this new computing platform.
As the U.S. government prioritizes the expansion of AI and computing capabilities, competition is heating up. AMD recently announced a $1 billion deal with the Department of Energy to create two supercomputers aimed at tackling major scientific challenges, including nuclear power and national security.
The ongoing flow of advanced technology between the U.S. and China is anticipated to be a pivotal topic in the impending trade discussions between President Trump and Chinese President Xi Jinping. Huang has consistently supported accessibility to the Chinese market, suggesting that about $50 billion in revenue is essential for funding U.S.-based R&D and maintaining a competitive edge in the AI sector.
Despite the hurdles presented by the trade conflict, Nvidia continues to produce chips at its TSMC facility in Arizona, assemble servers in Texas, and manufacture networking gear in California. Huang commended President Trump’s push for domestic manufacturing, noting, “We’re making things in America again, which is remarkable. The first thing he asked of me was to ‘bring manufacturing back.’”
As Nvidia continues to innovate across various sectors, it seems poised to strengthen its global leadership in AI technology.
