Nvidia CEO Jensen Huang announced at the company’s GTC conference in Washington, D.C., that its latest Blackwell AI chips are now in full production at a facility in Phoenix, Arizona. This marks a major shift, as Nvidia’s fastest GPUs were previously manufactured exclusively in Taiwan. Huang credited former President Donald Trump for pushing the company to bring production back to the U.S., recalling that Trump had asked him nine months prior to do so, citing national security and economic benefits. The first Blackwell wafers—thin silicon discs used to fabricate chips—were produced at a new facility in Phoenix by Taiwan Semiconductor Manufacturing Company (TSMC), in partnership with Nvidia. The company also confirmed that systems based on Blackwell chips will now be assembled domestically. This move underscores Nvidia’s broader strategy to strengthen its U.S.-based semiconductor supply chain amid growing geopolitical tensions. The conference, held in Washington, was strategically timed to engage policymakers and emphasize Nvidia’s role in national technological leadership. Huang stated that the event was arranged to allow Trump to attend, though the former president was abroad at the time. Reports indicate Trump planned to meet with Huang the following day. Demand for Nvidia’s GPUs remains exceptionally strong. Huang revealed that 6 million Blackwell chips have been shipped in the past four quarters. The company projects $500 billion in revenue from Blackwell and the upcoming Rubin chip generation combined. In another significant development, Nvidia announced a strategic partnership with Finland’s Nokia to develop next-generation telecommunications infrastructure. As part of the deal, Nvidia will take a $1 billion stake in Nokia and supply its chips for 5G and future 6G base stations. Huang emphasized that the U.S. communication network should not rely on foreign technologies, particularly those from China’s Huawei, which was banned from U.S. markets in 2018. The new system, called Nvidia ARC, integrates the Grace CPU, Blackwell GPU, and networking components to power smarter, AI-driven networks capable of supporting advanced applications like robotics and precision weather modeling. Huang’s message to Western leaders is clear: U.S. technological dominance in AI and communications must be secured. He argued that by making U.S. chips the standard globally, American innovation can shape the future, even in markets like China. He believes it’s better for Chinese developers to adopt U.S. technology than to build their own isolated systems. However, U.S. export controls continue to challenge Nvidia’s global reach. The company is currently “100% out of China,” Huang said, with no market share there. Although the U.S. government has approved licenses for the H20 chip under strict conditions—including a 15% royalty to the U.S. Treasury—Nvidia has not yet launched a Blackwell-based chip for the Chinese market. The company lost an estimated $10.5 billion in potential H20 sales due to licensing delays. Despite these hurdles, Nvidia continues to position itself as a cornerstone of American technological strength, with its presence in Arizona symbolizing a broader national effort to secure critical AI and semiconductor infrastructure.