Nvidia to Build AI Chip Plants in US for First Time

Nvidia to Manufacture AI Supercomputer Chips Entirely in the US for the First Time, signalling a Strategic Shift Triggered by Tariffs and Supply Chain Resilience

In a watershed moment for the global semiconductor industry, Nvidia has announced a sweeping plan to manufacture its cutting-edge AI supercomputer chips entirely within the United States—a first for the company. The move, backed by a staggering $500 billion investment in partnership with manufacturing giants like TSMC, Foxconn, Wistron, Amkor, and SPIL, marks a dramatic pivot from Nvidia’s long-standing dependence on overseas fabrication.

The decision comes amid the broader geopolitical reorientation of the global chip industry and is widely seen as a direct outcome of US President Donald Trump's tariff policy and industrial vision to restore American manufacturing dominance. The White House called Nvidia’s decision “the Trump Effect in action,” underscoring the former President’s efforts to use trade policy as a lever to bring advanced manufacturing back to US soil.

A Semiconductor Titan’s New Chapter

Nvidia, which commands between 70% and 95% of the AI chip market—a sector powering everything from OpenAI’s GPT models to data centre supercomputing—has been a dominant force amid the AI boom. With a gross margin of 78%, it is not only outpacing hardware peers like Intel (41%) and AMD (47%) but also redefining the economics of chipmaking.

Its flagship Blackwell AI chips, which retail at over $30,000 each, are already in production at Taiwan Semiconductor Manufacturing Company’s (TSMC) Arizona facilities. But for the first time, Nvidia will not rely solely on foundry partners for chipmaking. The company is spearheading the creation of supercomputer manufacturing plants in Houston (with Foxconn) and Dallas (with Wistron), aiming for full-scale production within 12 to 15 months. An additional one million square feet of manufacturing space has been commissioned in Phoenix, Arizona.

The Blackwell chips, revealed in March, feature 208 billion transistors and are fabricated using a custom-built 4NP process by TSMC. They are reportedly up to 30 times faster in generative AI applications than their predecessors.

Supply Chain Sovereignty Meets Manufacturing Might

Nvidia's US expansion reflects a growing imperative: to "harden supply chain resilience" as CEO Jensen Huang noted. “The engines of the world’s AI infrastructure are being built in the United States for the first time,” Huang said, emphasizing the company’s push to meet rising AI chip demand domestically.

The move aligns with a broader trend in Washington, where semiconductor self-sufficiency has become a bipartisan priority. The US government’s $54 billion CHIPS Act—passed under President Joe Biden—has helped catalyse 21 new semiconductor fabs and six major expansions. Yet, only three of those facilities are confirmed to manufacture leading-edge chips (sub-10nm nodes), a crucial benchmark for next-gen AI workloads.

Despite Nvidia’s commitment, challenges remain. While TSMC’s Arizona plant is handling the front-end production of Blackwell chips, final packaging still needs to be done in Taiwan due to a lack of CoWoS (chip-on-wafer-on-substrate) capacity stateside. This bottleneck highlights the complexities of establishing a fully localized supply chain for high-performance chips.

Semiconductor Race Accelerates

As of March 2025, Nvidia controls 7.3% of the global semiconductor market, upending the previous duopoly of Intel and Samsung that dominated pre-2020. The AI chipmaker generated $80 billion in revenue over the past four quarters, with Bank of America estimating $34.5 billion came from AI chips alone in 2023.

The semiconductor industry is projected to reach $702 billion in 2025, up $95 billion year-over-year. With governments worldwide racing to secure chip sovereignty, analysts believe the sector could surpass $1 trillion in value by 2030.

The Political Undercurrent

While Nvidia’s announcement has been championed by Trump and his allies, the political subtext remains fraught. Trump has promised additional tariffs on imported chips to bolster domestic capacity but has simultaneously called for the repeal of the CHIPS Act—legislation that has catalyzed much of the current investment wave.

Congressional Republicans have shown little appetite for dismantling the bipartisan law, creating an odd tension between the populist tariff agenda and the technocratic subsidy model. Meanwhile, in a notable reversal, the White House recently dropped plans to ban Nvidia’s H20 chips from being sold to China—a move likely designed to maintain revenue flows while the US domestic ecosystem catches up.

Has US Reindustrialisation begun?

Nvidia’s US manufacturing leap is more than just a corporate milestone—it’s a bellwether for the reindustrialization of America in the age of AI. As geopolitical risk, economic nationalism, and exponential technological change converge, the stakes for semiconductor leadership have never been higher. With $500 billion on the table, Nvidia isn’t just betting on chips—it’s betting on America.

Follow the Benefit News channel on WhatsApp