Nvidia to manufacture AI chips in U.S. for first time

Nvidia founder and CEO Jensen Huang has led the company’s transformation into a pillar of artificial intelligence.

Nvidia, the semiconductor giant known for its graphics processing units (GPUs), announced a historic plan to manufacture its artificial intelligence chips in the United States for the first time. The company, considered the undisputed leader in AI hardware with more than 70% of the AI accelerator market, will turn around its traditionally Asia-centric production model. This decision marks a milestone in the technology industry and is part of the growing U.S. strategy to boost local manufacturing of critical components.

Nvidia: from graphics card to AI leader

Founded in 1993 in California, Nvidia began as a manufacturer of graphics cards for video games, but its parallel processing capabilities catapulted it to the center of the AI revolution. Its GPUs proved ideal for accelerating deep learning algorithms, becoming the standard for training artificial intelligence models in the last decade. Today Nvidia dominates the high-performance AI chip segment, to the point where it “controls almost the entire industry,” according to the U.S. president. The company supplies its processors to cloud giants, data centers and AI supercomputing projects, consolidating a position of technological leadership and a market value close to a trillion dollars.

This leadership is supported not only by powerful hardware but also by a solid software ecosystem (CUDA, AI libraries, etc.), which has made it difficult for competitors such as AMD, Google or startups to snatch significant market share. With the explosion of applications such as generative AI (e.g. advanced chatbots, generative art), demand for Nvidia chips such as the A100 and H100 has skyrocketed, reinforcing its central role in the current AI “gold rush”.

The announcement: AI chip fabs in Arizona and Texas

Nvidia surprised the industry by announcing it will open manufacturing facilities in the United States, a first for its nearly 30-year history. According to the official statement, the company has enabled more than 90,000 square meters of space (approximately 1 million square feet) to build and test its advanced chips in the US. Below, we summarize the key details of the plan:

Investment and scope: Nvidia has committed to producing AI infrastructure valued at $500 billion in the United States over the next four years. This colossal investment will be made in partnership with several leading global manufacturing partners.
Locations and partners: Production will be primarily split between Arizona and Texas. In Arizona, Nvidia is collaborating with Taiwan Semiconductor Manufacturing Co. (TSMC) to manufacture chips at its new Phoenix plant. Additionally, it has partnered with semiconductor packaging and testing companies such as Amkor Technology and SPIL for packaging stages in the same state. Meanwhile, in Texas, the company will build AI supercomputer integration plants in partnership with electronics manufacturers Foxconn (in Houston) and Wistron (in Dallas).
Timeline: Manufacturing is already partially underway. Next-generation Nvidia Blackwell chips have begun production at TSMC’s Phoenix plant. Meanwhile, supercomputer assembly plants in Texas are under construction, with the goal of starting mass production in 12 to 15 months.

This means that these facilities will be fully operational by mid-2026, significantly increasing the supply capacity of AI hardware from North American soil.

Project Scope: In total, it is estimated that more than 1 million square feet of new facilities will be dedicated to this effort. Nvidia plans for approximately half of all its AI infrastructure produced in the coming years to come from the US, which would represent a notable shift from its current production concentrated in Asia.

This move is made possible by a network of alliances: “These world-leading companies (TSMC, Foxconn, Wistron, Amkor, SPIL) are deepening their partnership with Nvidia, expanding their global footprint and strengthening supply chain resilience,” the company noted. In other words, Nvidia won’t be building factories from scratch on its own, but will instead leverage contractor and partner plants already operating or being set up in the US, integrating its chips and systems within those facilities.

Blackwell chips and AI supercomputers

The new US plants will primarily manufacture Nvidia’s Blackwell architecture chips, the company’s most advanced. Blackwell is Nvidia’s next-generation GPU, the successor to the Hopper (aimed at AI data centers) and Ada Lovelace (advanced graphics) architectures. These processors are designed to handle massive machine learning computational workloads, with significant performance improvements for deep learning and accelerated computing tasks.

These chips will primarily be used to build complete AI supercomputers. Nvidia will integrate the chips into high-performance systems (such as its well-known DGX platforms, or AI pods) in its Texas factories, assembling true turnkey “AI factories.” Nvidia AI supercomputers are considered “the engines of a new type of data center built for the sole purpose of processing artificial intelligence,” the company explains. These machines combine hundreds or thousands of interconnected GPUs to train massive AI models (e.g., GPT-4 language models and higher, large-scale computer vision, scientific simulations, etc.). Nvidia predicts that dozens of dedicated AI data centers will emerge in the coming years—which it calls “gigawatt AI factories” due to the enormous computing power they consume. New US-made chips and supercomputers will supply these centers, which form the fundamental infrastructure of the burgeoning global artificial intelligence industry.

In short, Nvidia will manufacture both the chips and the complete systems: from the semiconductor level (Blackwell GPUs) in Arizona to the final assembly of AI servers and supercomputers in Texas. These products will then be shipped to customers such as technology companies, universities, research labs, and governments building AI data centers. The goal is to meet the growing global demand for AI computing power with less reliance on foreign inputs. “The engines of the world’s AI infrastructure are being built in the United States for the first time,” said Jensen Huang, founder and CEO of Nvidia, referring to the fact that such systems have never before been manufactured entirely in the country.

Impact on the US technology industry

Nvidia’s initiative promises to have a significant impact on the US technological and economic landscape:

Reindustrialization and Employment: Washington hailed the announcement as part of a “renaissance in American manufacturing.” Hundreds of thousands of highly skilled jobs are expected to be created in the coming years, associated with these factories and their supply chains. Engineers, technicians, and specialized personnel will be needed in chip production, supercomputer assembly, and supporting industries (suppliers of materials, manufacturing equipment, logistics, etc.). This represents a boost to advanced manufacturing employment after decades of offshoring.
Strengthening the Supply Chain: Producing such complex chips locally strengthens the resilience of the US technology chain against external disruptions. Huang emphasized that incorporating US manufacturing will help “better meet the incredibly growing demand for AI chips and supercomputers, strengthen our supply chain, and increase our resilience.” By diversifying Geographically, Nvidia and its customers will be less vulnerable to events such as international conflicts, natural disasters, or trade restrictions affecting Asia. Delivery times are also shortened by being able to assemble systems directly in the US market.
National technological leadership: This move helps the United States regain ground in cutting-edge semiconductor manufacturing, considered a strategic sector. Although many of the underlying technologies continue to originate in Silicon Valley and other innovation hubs, in recent decades, physical chip production had shifted almost entirely to East Asia. Having part of the manufacturing of the most advanced AI chips in the US consolidates US leadership in the key industry of the 21st century (AI), not only in design but also in production. Furthermore, other domestic companies such as Intel, AMD, and Google will be able to benefit from the industrial ecosystem generated around these new factories (local suppliers, trained labor, infrastructure, etc.).
Ripple effect in the private sector: Nvidia’s commitment joins a wave of technological investments in manufacturing. local. For example, rival or partner companies could announce similar projects to maintain competitiveness and maintain government support. In fact, massive commitments are already being seen, such as the OpenAI, SoftBank, and Oracle initiative to invest $500 billion in AI data centers in the US (the so-called “Stargate” Project) or Microsoft’s plan to spend $80 billion on AI infrastructure in the country. Nvidia’s move could turn regions like Texas into AI manufacturing hubs, attracting suppliers and other large technology companies. Cities like Houston anticipate thousands of jobs and an industrial transformation thanks to the arrival of these AI plants.

Of course, the real impact will depend on the successful execution of the project. Experts point out that the United States faces challenges in realizing this manufacturing resurgence, especially a possible shortage of specialized personnel. Operating semiconductor plants requires engineers in physics, chemistry, materials science, and electronics, who are currently in short supply in the country. Training and attracting that talent will be crucial for the investment to be successful. pay off. Moreover, production costs in the U.S. are higher than in Asia, posing long-term competitiveness challenges that Nvidia and its partners will have to balance with efficiency, automation, and government support.

Between the CHIPS Act and reducing dependence on Asia

Nvidia’s announcement does not come in a vacuum, but aligns with recent US industrial policies to regain technological sovereignty in semiconductors. In 2022, under the Joe Biden administration, the CHIPS and Science Act was enacted, providing more than $52 billion in subsidies and tax credits to encourage the construction of chip factories in the United States. This initiative triggered “an unprecedented wave of investment in microprocessors in the United States.” Companies such as Intel (in Arizona and Ohio), TSMC (in Arizona), and Samsung (in Texas) announced new mega-manufacturing projects taking advantage of public incentives. The government’s goal is for at least 20% of the world’s advanced chips to be manufactured in the US in the next decade, doubling the current share.

Although the chip was invented in the United States, the country has lost ground as a manufacturer in recent decades. Today, it manufactures barely 10% of the world’s semiconductors, “and none of the most advanced ones,” which come primarily from Taiwan and South Korea. This dependence on Asia is worrying not only for economic reasons but also for geopolitical ones: Taiwan—where Nvidia has produced most of its chips through TSMC—faces tensions with China, posing a risk to the global supply chain. In fact, leading US chip design companies (such as Nvidia itself in AI, Qualcomm in communications, and Apple in mobile devices) have historically outsourced manufacturing to Asian partners, concentrating the physical production of even the most sophisticated chips in that region.

Given this scenario, Nvidia manufacturing in the US for the first time represents a strategic shift. The White House has actively sought such announcements to move toward self-sufficiency in semiconductors. In addition to incentives like the CHIPS Act, the government has imposed export controls to prevent cutting-edge technologies from falling into the hands of geopolitical rivals. There are currently strong restrictions on the sale of the most advanced AI chips to China. At the same time, the new administration in 2025 has employed protectionist policies to pressure companies to produce locally. President Donald Trump threatened punitive tariffs of up to 32% on technology imports from Taiwan (the source of Nvidia’s GPUs) and 145% on products from China, seeking to reduce the technological deficit. At the same time, he temporarily exempted semiconductors and other strategic devices from these tariffs, apparently as a gesture to encourage announcements like Nvidia’s.

In this climate, it is no coincidence that Nvidia decided to announce its plan right now. According to analysts, the company is balancing the balance between attracting government support and avoiding sanctions that harm its market access or supply chain. The promise to invest heavily in the US allows Nvidia to curry favor with American authorities – and could have helped it circumvent potential export restrictions on certain next-generation chips – while gaining greater legal and logistical security in the long term.

Reactions: Government support and vision for the future

Nvidia’s announcement received widespread praise from US officials. The White House called the initiative part of the “Trump effect,” attributing the decision to national policies that prioritize domestic manufacturing. At an Oval Office event, the president himself celebrated the news, saying that “this is one of the most important announcements you’ll ever hear,” highlighting Nvidia’s dominant position in the market. Trump, true to form, noted that “the reason they did it is because of the November 5th election, and because of something called tariffs, the most beautiful word in the dictionary after love, God, and relationships,” alluding to the pressure that his protectionist measures and the election year would have exerted to precipitate the company’s decision.

Government officials emphasized that this is exactly the kind of outcome sought by the strategy to revitalize the industry. “Trump has prioritized U.S. chip manufacturing as part of his relentless pursuit of a manufacturing renaissance, and it is paying off: trillions of dollars in new investment secured in the technology sector alone,” the White House said in an official statement. Indeed, in addition to Nvidia, numerous investments have been announced recently that aim to return the United States to the forefront of advanced technology production.

Reactions in the private and academic sectors have also been generally positive. Semiconductor experts see Nvidia’s move as “a step in the right direction” to mitigate risks in the global chip supply chain and ensure supply in the face of explosive demand for AI hardware. However, they emphasize that challenges remain: “Subsidies alone do not guarantee a sustainable industry; it requires customers, a complete supply chain, and, above all, a specialized workforce.” In other words, the United States will need to train tens of thousands of new professionals to operate these cutting-edge facilities and maintain competitive costs. Public-private initiatives in education and immigration incentives for foreign talent may be necessary to fill this skills gap.

Ultimately, investors and the stock market greeted the news with cautious optimism. While Nvidia is already a deeply globalized company, this strategic move reinforces confidence that it will be able to navigate international tensions without production disruptions. Furthermore, it sends a powerful signal: technological innovation centers also want to be manufacturing centers, integrating the entire value chain from design to production. In the words of one analyst, “it’s a paradigm shift that just a few years ago seemed remote, but is now beginning to materialize.”

With this project, Nvidia not only consolidates its leadership in the era of artificial intelligence, but also becomes the spearhead of the US technological manufacturing resurgence. It remains to be seen how its implementation evolves in the coming months, but the announcement undoubtedly sets a precedent: the manufacturing of AI chips “Made in the USA” will soon be a reality, with profound implications for the global semiconductor industry and the global technological balance.