6 Products the US Can't Make Fast Enough to Replace Chinese Imports

The story of 2025 US trade policy, told in aggregate, sounds like progress. Imports from China fell to their lowest level since the 2009 financial crisis. China's share of US imports dropped from roughly 13% to around 7%. The bilateral trade deficit narrowed by nearly a third, to $202 billion. These are real numbers, and the administration has been right to cite them.

But there is a second story embedded in the same data, and it's less comfortable. A significant portion of what moved out of China didn't move into American factories. It moved to Vietnam, Taiwan, Mexico, and India — countries that, in many cases, are themselves dependent on Chinese components, Chinese processing, and Chinese raw materials to make the goods they now export to the United States. The supply chain shifted origin. The dependency, in many critical categories, did not.

Researchers at the Peterson Institute have put this plainly: the United States has now unwound the supply chains that were easy to unwind. What remains are the hard ones — the categories where China's dominance is so deep, so structurally embedded, and so reliant on accumulated expertise and scale that no amount of tariff pressure can manufacture a domestic alternative on any near-term timeline. Here are six of the most consequential.

In June 2025, US automakers sent a stark warning to Washington: imminent shortages of rare earth element magnets would shut down their production lines "within weeks." That wasn't hyperbole. It was a direct consequence of China placing export licensing restrictions on the seven rare earth minerals that go into the neodymium-iron-boron (NdFeB) magnets found in every modern electric vehicle motor, every wind turbine generator, every hard drive, and dozens of critical defense systems.

The scale of Chinese dominance here is genuinely difficult to overstate. China controls over 94% of global permanent magnet production, according to the International Energy Agency. It controls more than 60% of the mining of the underlying rare earth elements and more than 80% of the refining capacity that turns raw ore into usable oxide.

When Beijing imposed export controls on these materials in April 2025 — requiring special licenses for all exports, with particular scrutiny of US-bound shipments — there was simply no alternative supply to draw on. The export controls were paused through a bilateral negotiation in late 2025, but they remain available to Beijing as leverage, with the suspension set to expire in November 2026.

The US government has responded with unprecedented investment. The Pentagon took a 15% equity stake in MP Materials, the only fully integrated domestic rare earth producer, and committed to a $1 billion lending facility to build a new magnet manufacturing plant in Texas. USA Rare Earth launched its Stillwater, Oklahoma, facility in early 2026, with the capacity to meet roughly 17% of US domestic demand. These are real steps.

But even the most optimistic projections put domestic magnet capacity at a small fraction of US demand through the end of the decade. The number of workers enrolling in mining and materials science programs is declining, compounding the challenge. Building a rare earth supply chain that took China thirty years and hundreds of billions in state investment cannot be reverse-engineered in a congressional appropriations cycle.

The numbers here are alarming in a way that has attracted genuine bipartisan urgency, and almost no meaningful resolution. China is the United States' largest foreign supplier of active pharmaceutical ingredients — the chemical compounds that make drugs work — by volume, accounting for roughly 40% of what flows into US generic drug manufacturing. For antibiotics specifically, the concentration is even higher: the US relies on China for more than 90% of antibiotic API imports, including penicillin, streptomycin, and amoxicillin. USAntibiotics, based in Bristol, Tennessee, is the last remaining end-to-end domestic manufacturer of amoxicillin — the most prescribed antibiotic in America.

A 2026 analysis by the API Innovation Center found that of the top 100 generic medicines prescribed in the United States, 83 have no domestic source of active ingredient whatsoever. For one in ten critical pharmaceutical inputs, China's market share exceeds 99%.

These are not niche products. The list includes APIs for drugs that treat type 2 diabetes, HIV/AIDS, high blood pressure, and common bacterial infections. In March 2025, prices for metformin hydrochloride — the first-line oral therapy for type 2 diabetes taken by millions of Americans — surged in China due to supply tightness and rising production costs. There was no domestic alternative to absorb the shock.

The structural problem is layered. Even when pharmaceutical companies try to "China Plus One" their sourcing by shifting to Indian manufacturers, they typically find that India's own API industry relies on China for roughly 70% of its chemical intermediates — the upstream building blocks from which APIs are synthesized. A drug that is officially "Made in India" may still be 70% dependent on Chinese chemistry. Testimony before the House Select Committee on China in March 2026 put the situation plainly: "For most of American history, when a doctor prescribed a medication, it was made in our country, or somewhere we trusted. That's no longer true."

No single country has ever dominated a global manufacturing sector as thoroughly as China dominates solar. Chinese manufacturers produce approximately 80% of the world's solar panels.

More importantly, they control the overwhelming majority of every upstream step in the manufacturing chain: the polysilicon that starts the process, the ingots and wafers cut from it, the solar cells assembled from those wafers, and the finished modules shipped to installation projects worldwide. A Rhodium Group analysis found that China's role in solar manufacturing is "dominant and still increasing along most of the supply chain."

The US government has attempted to use tariffs and trade law to protect domestic solar manufacturers for over a decade, with limited results. The core problem is cost. Chinese manufacturers, backed by decades of state investment, optimized supply chains, and the scale advantages that come from producing the vast majority of the world's supply, can produce panels at prices that are extraordinarily difficult to compete against.

American and European producers have consistently found themselves making the same argument — that Chinese pricing reflects subsidized competition rather than genuine market efficiency — with the same outcome: Chinese panels continue to dominate.

The secondary tactic — sourcing panels from Southeast Asian assemblers in Vietnam, Cambodia, and Malaysia — has run into two walls. First, the US applied tariffs to those countries, too, under the 2025 package. Second, and more fundamentally, those Southeast Asian assembly operations depend heavily on Chinese polysilicon, wafers, and cells.

They are finishing lines for Chinese supply chains, not independent alternatives. Building genuine solar manufacturing independence would require rebuilding the entire upstream chain in non-Chinese locations, a project that industry analysts estimate would take a decade and hundreds of billions of dollars even with sustained government support.

The semiconductor conversation in American policy is often framed as a Taiwan dependency problem, and that framing is largely correct for the most advanced chip fabrication. But there is a distinct and underappreciated China dependency embedded in the semiconductor supply chain at the packaging and assembly layer — and it matters more for everyday commercial electronics than the headline TSMC story does.

Advanced chip packaging — the process of encasing fabricated chips in their final protective housing, connecting them to circuit boards, and assembling multi-chip modules — is a critical step that China has been expanding aggressively. Chinese integrated circuit exports rose by 26.8% in 2025, accounting for roughly one-fifth of the total growth in China's overall export value.

The country is not just a market for chips designed elsewhere; it is increasingly a manufacturer and packager of semiconductor products that flow back into global electronics supply chains, including American ones.

The CHIPS Act has directed roughly $52 billion toward domestic semiconductor manufacturing, and Intel, TSMC, and Samsung have all announced US fab investments. But those facilities are focused on leading-edge fabrication — the most advanced nodes that power AI accelerators and high-performance computing. The mature nodes that go into cars, appliances, industrial equipment, and consumer electronics are manufactured in enormous volume in China and Taiwan, and the US has no near-term path to replicating that capacity domestically.