76report

October 22, 2025

Building the Future Within Our Own Borders

“Semiconductors are the backbone of the 21st century economy…. We must be able to build the chips and semiconductors that we need right here in American factories with Americans skill and American labor.”

President Trump spoke these words in March 2025, but he has been pushing to onshore the semiconductor supply chain ever since his first term.

Last week, NVIDIA (NVDA), the undisputed leader in advanced AI chips, announced that it is now officially producing its cutting-edge Blackwell chip at a facility in Arizona.

This marks a historic milestone in America’s effort to manufacture the most strategically important technology within its own borders.

But this achievement did not happen overnight.

Well before “AI” became a household term—and more than two years before the launch of ChatGPT in November 2022—President Trump was laying the groundwork.

When the pandemic struck in early 2020, the fragility of global supply chains was laid bare. The world discovered just how dependent it was on Asian semiconductor production for everything from laptops to cell phones to cars.

Americans faced shortages of basic goods, like disinfectant and toilet paper. Soon, they could not even buy new vehicles because automakers were unable to secure the chips required by modern automobiles.

The rapid rise of remote work had led to a spike in demand for laptops and consumer electronics. Chipmakers prioritized those markets, leaving automakers stranded.

This was bigger than a temporary supply crunch. Essentially all modern technology relies on semiconductors, and the need for semiconductors would only intensify over time as Web 3.0 and other new technologies took hold.

Recognizing the long-term implications of this crisis, President Trump moved to rebuild the U.S. semiconductor base. His administration directly engaged chipmakers, urging them to establish production in America and reduce reliance on Taiwan and China.

In May 2020, after months of negotiation involving the White House and the Commerce Department, Taiwan Semiconductor Manufacturing (TSM) announced plans to build a $12 billion advanced chip fabrication plant in Arizona “with the mutual understanding and commitment to support from the U.S. federal government.”

The Arizona fab became the flagship example of the administration’s reshoring agenda. It also provided the foundation for the broader semiconductor revival later codified in the 2022 CHIPS Act, which had bipartisan support.

A prescient move

Semiconductor fabs are truly a modern miracle. They require extreme precision, with “cleanrooms” that are 10,000 times cleaner than hospital operating theaters.

A fab can take three to five years to complete. Inside, lithography machines fire microscopic laser pulses—up to 50,000 times per second— to etch complex circuits onto silicon wafers.

The goal of the first Trump administration in bringing TSM production capabilities into the United States was to start addressing the general risk of technological reliance on Asian supply chains.

The Arizona fab now has a more defined and even greater purpose. It will function as the first domestic facility producing the world’s most advanced Graphic Processing Units (GPUs).

GPUs are the specialized semiconductor chips that make AI possible. NVDA is the world leader in GPUs. It has also become the most valuable business in the world, representing an astounding 8% of the S&P 500 Index.

NVDA is arguably now the single most important company in the world, and it describes the Blackwell GPU that is being produced in Arizona as its “single most important chip.”

This past week was historic. For the first time, we manufactured the world’s most advanced AI chips in the most advanced fab, right here in America. It all began with President Trump’s effort to reindustrialize the United States. His tariff policies accelerated the process, and now, less than a year later, we’re producing cutting-edge AI semiconductors on U.S. soil…. Over the next three to four years, we expect to manufacture roughly half a trillion dollars’ worth of AI supercomputing technology for deployment across the U.S. - Jensen Huang (10/19/2025)

NVDA’s Jensen Huang on domestic production of AI chips

TSM’s American roots

A position within our American Resilience Model Portfolio, NVDA is a stock we have written about extensively, including, most recently, Implications of OpenAI/NVIDIA Mega Deal.

NVDA is the driving force behind the AI revolution. It is a business that one needs to understand well if one truly wants to understand what is happening within the economy as a whole.

NVDA has not been alone in this endeavor, however. NVDA designs its extremely complex chips, but it does not actually make them. For this, it relies primarily on TSM, a Taiwanese company with deep American roots.

TSM was founded in 1987 by Morris Chang, who is revered as the father of Taiwan’s semiconductor industry. (It is worth mentioning, and perhaps more than a coincidence, that Jensen Huang himself is from Taiwan and emigrated to the United States with his family in 1973 at the age of nine.)

Born in 1931 in Ningbo, China, Chang fled to Hong Kong, then a British colony, during the Chinese Civil War as the Communists came to power. He later studied in the United States, earning a Bachelor’s degree in Mechanical Engineering from MIT and a Ph.D. in Electrical Engineering from Stanford.

Chang spent 25 years at Texas Instruments (TXN), rising to senior management before being recruited by Taiwan’s government in the 1980s to help build its technology industry. In 1987, he founded TSM, pioneering the “pure-play foundry” model, where a company manufactures chips for others rather than designing its own.

The foundry model reshaped the global semiconductor industry and enabled the rise of “fabless” giants like Apple, NVIDIA, and Qualcomm. These companies dreamt up the designs, while Chang’s TSM made them a reality.

By specializing only on chip production, TSM achieved a degree of technological superiority that is now basically unchallenged. TSM has approximately 90% share of the advanced chip manufacturing market.

The semiconductor industry was invented in the United States. Thanks to the support of American tech companies and the American government over the decades, that industry has blossomed in Taiwan, propelling both Taiwanese and American tech businesses to new heights.

NVDA’s success would never have been possible without TSM and the vast ecosystem of suppliers within Taiwan that support the chip industry.

Now, through the partnership between NVDA and TSM, and with some strong encouragement by the American government, the knowledge forged across continents is converging again—this time on American soil.

TSM was formed decades ago as an act of resilience and self-determination. Taiwan needed an engine of economic growth that might allow the small island nation to keep its independence from China. The U.S. facilitated this.

It is only fitting that TSM is now helping the United States secure its economic independence in a precariously interconnected and rapidly changing world.

The limits of comparative advantage

Trump has been criticized, often by conservatives, for his tendency to intervene in the free functioning of the global economy.

The main economic principle in question, which Trump has allegedly violated, is known as “comparative advantage.” It refers to the ability of one nation to produce a good or service more efficiently than another.

Adam Smith laid the intellectual groundwork for this understanding of trade back in 1776 when he wrote The Wealth of Nations, the bible of Enlightenment capitalism.

Smith wrote: “If a foreign country can supply us with a commodity cheaper than we ourselves can make it, better buy it of them with some part of the produce of our own industry, employed in a way in which we have some advantage.”

British economist Robert Torrens is credited with developing the concept more fully, when he illuminated the benefit to England of trading with France and exchanging broadcloth for lace.

If I wish to know the extent of the advantage which arises to England from her giving France a hundred pounds of broadcloth in exchange for a hundred pounds of lace, I take the quantity of lace which she has acquired by this transaction, and compare it with the quantity which she might, at the same expense of labour and capital, have acquired by manufacturing it at home. - Robert Torrens, The Economist Refuted (1808)

The idea was later refined by economists like David Ricardo, who actually introduced the term comparative advantage in 1817 and provided more theoretical context.

Taiwan has certainly become very good at manufacturing semiconductors and related equipment. For that reason, one might be tempted to allow Taiwan to continue to focus on its area of specialization, while the U.S. pays attention to other tasks that it does well.

While theoretically sound under certain circumstances, this free trade approach assumes stable trade, perfect mobility of goods, and, most importantly, no geopolitical risk. In reality, supply chains can be weaponized or broken, exports can be restricted, and costs can be distorted by subsidies.

In the particular case of Taiwan, the threat of an eventual Chinese invasion creates an unacceptably dangerous scenario. Economic theory goes out the window as soon as the People’s Liberation Army’s amphibious vehicles start to land on Taiwan’s beaches.

Private enterprises, when functioning properly, do what is in the best interests of their shareholders. Major global companies will not redirect billions of dollars of resources into new geographies simply because you ask them nicely. It takes pressure.

As Jensen Huang said, it was only because of Trump’s pressure through tariffs that TSM originally built the Arizona fab and, more recently, that the fab was reoriented around production of advanced AI chips.

Powering the next phase

The Trump administration has recognized AI as a critical area of technological innovation. It is pursuing a multi-pronged industrial policy around achieving global AI dominance.

Making the first Blackwell chips in Arizona is an important initial step toward establishing AI supply chain resilience. Ideally, over time, a critical mass of production will be located within U.S. borders.

But having reliable access to AI equipment is only half the story. The other half is energy. Jensen Huang understands this more than anyone.

The way to think about AI is that it requires factories — massive data centers — and those factories consume energy. One of the most important things President Trump did for the U.S. technology industry was to recognize that energy is essential to AI growth. His pro-energy initiatives made it possible for us to sustain this expansion. You take energy, put it into these AI factories, and what comes out is digital intelligence. - Jensen Huang (10/19/2025)

China may or may not invade Taiwan at some point, but what is certain is that China is building its electrical infrastructure at an extraordinary pace—adding generation capacity, grid storage, and industrial power corridors that will fuel its AI ambitions for decades to come.

Here again, the free market, left to its own devices, cannot solve the problem. Energy is too capital-intensive, too entangled with regulation, and too strategically vital to be left entirely to market forces. It requires national direction—policy clarity, permitting reform, and long-term planning and coordination across many layers of government and private enterprise.

If the United States hopes to remain the global leader of the AI revolution, it must pair semiconductor onshoring with a parallel energy infrastructure build-out of enormous proportions.

Investing in American AI

As we recently described in our Guide to Investing in the Age of AI, the AI revolution is multi-faceted—driven by companies like NVDA at the very epicenter of the movement, but extending far beyond the semiconductor sector itself.

It encompasses energy, industrial, and materials firms that will power and supply the physical backbone of the AI economy—from data centers to next-generation power grids. It also includes enterprises across finance, healthcare, logistics, and manufacturing that will harness AI to improve efficiency, reduce costs, and unlock new revenue streams.

As America re-industrializes around compute capacity and energy, capital is gravitating toward companies that combine resilience, productivity, and technological leverage. These are the long-term compounders we target across all of our Model Portfolio strategies.

Click HERE to learn more about our Model Portfolio subscription plans.

Disclaimer

FOR SUBSCRIBER USE ONLY. DO NOT FORWARD OR SHARE.