EDITORIAL 🔗 Part of: Trump Administration Policies (43) · 21 Jul 2026 · 17:44 GMT ·13 min read

Beyond the US Tariffs: The Mechanics of Reshoring. Why Rebuilding American Factories and Chip Plants May Demand Decades of Blueprints, Not Just Pressure

The United States is trying to rebuild what it spent forty years sending offshore.

NEUTRALITY 74 · BALANCED· 10 sources

CNBC

The United States is trying to rebuild what it spent forty years sending offshore. That means factories, and above all the chip plants that make the physical world run. Two presidents bet on opposite tools. Joe Biden signed a law in 2022 that put 39 billion dollars of subsidies on the table. Donald Trump, back in office, has leaned on tariffs and pressure instead. He even took a government stake in Intel. Yet by 2026 factory construction spending is falling. Manufacturing has lost jobs for three years running. The hard question is whether either lever can beat the clock.

The debate begins with a clock. A leading-edge semiconductor fabrication plant, designed to manufacture the most advanced chips, takes years to build and even longer to staff. Because the work is heavily automated, these facilities employ only a few thousand people even when operating at full capacity. For example, Taiwan Semiconductor Manufacturing Company (TSMC) began construction in Arizona in 2020, yet its second plant there is not expected to reach high-volume production until the latter half of 2027. Similarly, Intel's flagship Ohio site was initially promised for 2025, but its timeline has now slipped to 2030 or 2031. These are the daunting timelines that any reshoring plan must bend, whether it relies on subsidies or tariffs.

US Government Invested In Intel

The argument over how to accelerate these schedules is real and cuts across party lines. One faction argues that the country faces a national security emergency, noting that the pandemic exposed a deep dependence on chips made in a handful of Asian plants. From this perspective, only hard leverage, such as strict tariffs and the threat of losing access to the American market, can force companies to move quickly. Conversely, the opposing side contends that factories are built on certainty rather than threats. They warn that whipsawing tariff policies have actually frozen the very investments they were meant to unleash.

Ultimately, economists and industry analysts offer a thesis worth considering. They claim that durable reshoring will likely require decades of steady industrial policy, infrastructure development, and worker training. In this view, tariffs and political pressure are merely one lever among many, rather than a magic switch that can instantly alter the outcome on its own.

Context

The story starts with a shock. Beginning in 2020, the COVID-19 pandemic disrupted global supply chains,

Washington responded with financial backing. In August 2022, then-President Joe Biden signed the CHIPS and Science Act, setting aside 39 billion dollars in direct incentives to build, expand, or modernize semiconductor plants on American soil. The law also included a separate 25 percent investment tax credit for chipmaking equipment and facilities. The Congressional Research Service, Congress's nonpartisan research arm, describes this structure in detail in its public report on the legislation.

Under the Biden administration, these funds were distributed in large, named awards. TSMC was offered up to 6.6 billion dollars for its Arizona plants, while Intel was awarded up to 7.86 billion dollars, slightly down from an earlier 8.5 billion dollar figure. Micron received up to 6.2 billion dollars for facilities in New York and Idaho, and Samsung's award was set at approximately 4.7 billion dollars for its site in Taylor, Texas. By July 2025, the Commerce Department had awarded about 30.9 billion dollars across 40 projects run by 19 companies, with the plants expected to be completed by 2033.

However, the factories rose more slowly than the initial announcements suggested. Although TSMC's first Arizona plant began high-volume production in late 2024, its second plant was pushed to the second half of 2027, and its third only broke ground in April 2025. Intel's Ohio site, once nicknamed the "Silicon Heartland," was delayed from a planned 2025 opening to 2030 or 2031 as the company sought to align construction timelines with market demand. Samsung slowed progress on its Texas plant into 2026 amid reports citing a lack of customers for the specific chips it was originally designed to manufacture. Micron similarly pushed the start of construction at its largest New York plant toward late 2026. While the specific reasons varied, a clear pattern emerged: building leading-edge chip capacity is a slow, expensive process that relies heavily on fluctuating market demand.

The political lever then changed hands. When Donald Trump returned to the presidency in January 2025, he made it clear that he preferred tariffs over subsidies. In a March 4, 2025, speech to Congress, he called the CHIPS Act a "horrible, horrible thing," arguing that the government did not need to pay companies to build domestic facilities when it could simply tax the chips they manufactured abroad. On August 6, 2025, he announced that the United States would impose a 100 percent tariff on imported semiconductors. This tariff included a notable carve-out: companies that were already building or had committed to building in the United States would pay nothing. To qualify for this exemption, firms had to file their commitment with the Commerce Department and agree to be audited during construction.

This pressure generated massive pledges. Apple announced an expanded United States investment commitment, which the White House valued at 600 billion dollars. TSMC raised its total planned Arizona investment to 165 billion dollars and successfully secured an exemption from the tariff, marking the largest foreign investment in a new project in American history. The administration also restructured one of the largest CHIPS awards. On August 22, 2025, following a meeting between the President and Intel Chief Executive Lip-Bu Tan, the federal government converted Intel's remaining grant money into equity. By purchasing 433.3 million shares for about 8.9 billion dollars, the government acquired a roughly 10 percent stake in the company. While the government described this as a passive holding with no board seat, Senator Elizabeth Warren and others raised concerns about the federal government owning a slice of a private chipmaker.

The administration then formally implemented the chip tariffs. On January 14, 2026, the White House announced that the President had signed a proclamation under Section 232 of the Trade Expansion Act of 1962, a national security trade law. This signing followed a Commerce Department investigation concluding that reliance on imported chips posed a threat to national security. The proclamation imposed a 25 percent tariff on certain advanced computing chips, including the NVIDIA H200 and AMD MI325X, though it exempted chips imported specifically to build up domestic manufacturing capacity. An accompanying fact sheet indicated that broader semiconductor tariffs could soon follow.

By 2026, the data revealed a complicated story, as the chip tariffs were just one part of a much larger tariff program. On November 17, 2025, the Budget Lab at Yale University, a nonpartisan policy research center, reported that all 2025 tariffs combined had pushed the average effective United States tariff rate to 16.8 percent, the highest level since 1935. While manufacturing construction spending had surged by more than 200 percent during Biden's four years in office, peaking in 2024 at an annual rate of about 235.6 billion dollars, it eventually began to fall. In February 2026, FactCheck.org, a project of the University of Pennsylvania's Annenberg Public Policy Center, reported that this spending declined 6.7 percent from the last quarter of 2024 through the third quarter of 2025, adding that forecasters expected further declines in 2026 and 2027.

Manufacturing jobs did not see an increase, either. Drawing on Bureau of Labor Statistics data, FactCheck.org reported that the economy lost 63,000 manufacturing jobs during Trump's first 11 months in office, which followed a loss of 98,000 jobs in the 11 months prior. Consequently, the manufacturing sector posted three consecutive years of net job losses. Despite this, supporters point to a slower-moving reality that even data analysts concede: the largest CHIPS-funded plants are still under construction, meaning the jobs they eventually create may not materialize for several years.

Supportive perspectives

The Trump administration and its backers
i. The dependence on imports was a genuine national security threat.
The administration's core argument is about security, not just trade. The White House proclamation of January 14, 2026, invoked Section 232. That section is the national security clause of trade law. The proclamation rested on a Commerce Department finding. American capacity to manufacture, especially semiconductors and chip-making equipment, was insufficient to meet demand. That left the country reliant on foreign supply. Chips run the military as well as the economy. The pandemic shut assembly lines. The administration argues that leaving the most advanced production concentrated near a possible China conflict was a risk the United States could not accept.
ii. Tariffs and pressure move companies faster than subsidies alone.
The administration's case is that leverage works where checks did not. It points to the scale of the pledges that followed the tariff threat. Apple made a commitment that the White House valued at 600 billion dollars. TSMC decided to raise its planned Arizona investment to 165 billion dollars and seek the Build in America exemption. President Trump argued in his March 2025 remarks that companies would rather build in the United States than pay a steep import tax. He argued the government need not hand out subsidies to get the same result. Supporters add that the Intel equity stake of August 2025 turned a giveaway into an investment the taxpayer could profit from.
Manufacturers and the business sector
i. Subsidies and certainty are what actually get factories built.
The business case starts from how capital decisions are made. A leading-edge plant can cost tens of billions of dollars and take most of a decade to pay back. So firms commit only when the rules are stable. Companies took the CHIPS money because it lowered the upfront cost and signaled a durable national commitment. TSMC, Intel, Micron, and Samsung all expanded United States plans during the subsidy era. The industry states its preference through its trade bodies. It wants predictable incentives and clear timelines over shifting threats.
ii. Demand drives business not patriotism
Manufacturers argue the slipped timelines are a feature of the business, not defiance. By its own account and press reporting, Samsung lacked customers for the chips its Texas plant was first built to make. It slowed the plant and chose to upgrade it to a more advanced process instead. Intel slowed Ohio to align with market demand. The business view is that the most useful thing government can do is provide stable rules and let companies build capacity into real orders. On this view, government should not force plants to open before there is a market to feed them.
Labor and workers
Reshoring is a chance to rebuild a path into the middle class(if it lasts).
For workers and the unions and training programs that represent them, the prize is real jobs. The Reshoring Initiative is a group that tracks the return of manufacturing. It has counted hundreds of thousands of jobs announced through reshoring and foreign investment each year since 2010. Computer and electronics work leads the recent figures. The CHIPS Act drove that growth. Even with heavy automation, a single chip plant employs thousands. The suppliers around it employ more. Labor's strongest case is that a sustained national effort, whatever the tool, could rebuild skilled blue-collar career paths that two generations of offshoring erased. That holds provided the training pipeline is funded to match.
Economists, trade partners, and allies view
Allied investment can be a win for everyone(if the terms hold).
Some economists and the governments of Taiwan and South Korea argue the reshoring push can serve all sides if it is handled through negotiation. To secure lower tariffs, Taiwan agreed to invest 250 billion dollars in United States industries. It offered up to 250 billion dollars more in credit guarantees. That helped cut its tariff rate down to 15 percent, according to reporting on the United States and Taiwan trade deal. The rate had first been announced at 32 percent and later set at 20 percent. South Korea reached its own investment-for-tariff-relief deal. The allied case is that channeling Asian chip expertise and capital into American plants strengthens a shared supply chain against China. Walling that expertise out would not. The approach also keeps these economies inside the American market.

Critical perspectives

Of the Trump administration
i. The numbers so far show factories slowing, not roaring back.
The sharpest objection is empirical. Manufacturing construction spending fell 6.7 percent from late 2024 through the third quarter of 2025. It is forecast to keep falling in 2026 and 2027. Those figures showed the economy lost 63,000 manufacturing jobs in the administration's first 11 months. The Wall Street Journal, in its own reporting, joined other critics. They argue the tariffs have raised manufacturers' input costs. They also argue that stop-and-start tariff announcements created what executives called a lost year for investment.
ii. Tariffs are a tax that families and the broader economy pay.
Economists who study tariffs warn the cost lands at home. The Budget Lab at Yale University estimated on November 17, 2025, that all 2025 tariffs raised consumer prices by about 1.2 percent in the short run. That works out to a loss of roughly 1,700 dollars per household on average. The lowest-income households lose about 900 dollars because they spend more of their income on goods. The same analysis projected that manufacturing output might expand about 2.9 percent in the long run. But it found that gain would be more than offset by a 4.1 percent contraction in construction and a 1.4 percent decline in agriculture. It also put payroll employment about 490,000 lower by the end of 2025. Critics argue a tool that shrinks the whole economy to grow one slice is a poor bargain.
iii. Government ownership and audited build pledges raise their own risks.
The Intel equity stake drew bipartisan unease. Senator Elizabeth Warren pressed the Commerce Secretary on a deal that made the government a major shareholder in a private chipmaker. The Information Technology and Innovation Foundation, a technology policy group, argued publicly that the administration should not take equity in semiconductor firms. Critics warn that a government owning a stake in one competitor distorts the market against the others. The tariff exemptions are granted company by company, through filed and audited commitments. Critics say that process invites lobbying and favoritism rather than clear, neutral rules.
Of the manufacturers and business sector
i. Subsidies without delivery look like corporate welfare.
The objection aimed at industry is that public billions have not yet bought the promised plants or jobs. By 2026 the marquee projects had slipped years past their first targets even as awards were announced. Manufacturing employment kept falling. Critics on the populist right and parts of the left make the same point from different directions. Chipmakers received grants and tax credits while the plants ran late. The finished plants will ultimately employ relatively few people because of automation. That can read as a transfer to large corporations rather than a deal that delivers for the public that funds it.
ii. Automation means the job payoff will be smaller than the slogans suggest.
A structural critique cuts against everyone who sells reshoring as a jobs engine. Modern fabrication plants are among the most automated facilities on earth. They are designed to run with minimal human intervention. So each finished plant supports far fewer permanent workers than the factories of the past. Most of the labor is concentrated in the construction phase, which ends. Critics argue that even a fully successful reshoring of chip production would not restore mass factory employment. Promising it oversells what the technology will deliver, they say.
Of labor and the workforce pipeline
There are not enough trained workers to staff the plants being built.
The hardest constraint on labor's own hopes is a shortage of skilled people. A 2023 study by the Semiconductor Industry Association and Oxford Economics projected the chip workforce would need to grow by about 115,000 jobs by 2030. It found roughly 67,000 of those positions risked going unfilled at current rates of relevant degree and certificate completion. TSMC's Arizona plant cited a shortage of trained United States staff among its early delays. Critics call for a far larger and faster investment in technical training, community colleges, and engineering pipelines. Without it, they warn, the plants will rise faster than the country can staff them. The jobs will not translate into the careers labor is counting on.
Of trade partners and allies
Pressure on allies risks retaliation and may not bring the jobs home.
Trade partners and many economists question both the fairness and the effect of the strategy. Allies such as Taiwan and South Korea were pushed to pledge hundreds of billions of dollars under tariff threat. That strains relationships with the very partners whose expertise the United States needs. It also invites retaliation. The Budget Lab at Yale University factored that retaliation into its estimate of lower United States exports. Economists also warn that the deep uncertainty created by shifting tariffs deters the long-term investment reshoring requires. The Joint Economic Committee minority staff estimated that a prolonged period of trade-policy uncertainty could cut United States manufacturing investment by about 13 percent a year. That is roughly 490 billion dollars in foregone investment by 2029. The objection is that the pressure may scatter the capital it was meant to attract.