Highlights

• Billions Spent, Still Dependent: America’s bid to reshore semiconductor manufacturing – backed by the $52 billion CHIPS Act – has yielded shiny new fabs in Arizona, but at staggering cost overruns (TSMC’s investment ballooned from $12 billion to $40 billion (opens in a new tab)) and glaring supply-chain gaps. Advanced chips made in the U.S. still must be flown back to Taiwan for final packaging, adding expense and underscoring how incomplete the domestic ecosystem remains (opens in a new tab).
• “Made in USA” Missing Pieces: The U.S. underestimated the inputs needed to operate high-tech facilities. Taiwan’s TSMC, for example, replicated its production environment in Arizona but had to import essentials – from hundreds of skilled engineers to ultra-pure water infrastructure – to keep the project on track, cites Apple Insider. Labor shortages, water scarcity in the desert, and reliance on foreign equipment reveal that simply building a factory doesn’t guarantee self-sufficiency.
• Rare Earth Déjà Vu: Now, a similar blind spot looms in rare earth elements (REEs). Washington, DC, under President Trump, is racing to fund mines and separation plants for critical minerals (via the legacy Defense Production Act and Inflation Reduction Act, and more imminently a series of emergency orders). Still, without a full plan with real industrial policy, new U.S. rare earth facilities could stall or remain dependent on Chinese chemicals and know-how. The refining of REEs requires hundreds of solvent extraction steps with specialized reagents as cited via REEx – guess who dominates those? China.
• Lightweight Policy Without a Plan: From the CHIPS Act to the Inflation Reduction Act to the even more transactional, emergency, and tariff-driven Trump 2.0, U.S. leaders touted bold investments to reclaim supply chains, yet failed to map out an end-to-end industrial strategy. The CHIPS Act poured subsidies into fabs but neglected downstream steps like advanced packaging, as but one example (hence the ongoing reliance on Taiwan). Likewise, the IRA incentivized EV battery minerals but overlooked permanent magnets and motor components that rely on rare earths, leaving that vulnerability unaddressed (opens in a new tab) back in 2022 under Biden.
• The Cost of Neglect: Lack of a comprehensive industrial policy means paying more for less security. America risks repeating the semiconductor fiasco with rare earths – spending vast sums on facilities that, in practice, still depend on imports or face delays. Experts warn, as identified in REEx, it will take 5–10 years of concerted effort to replicate China’s capabilities. Until then, the U.S. remains one supply chain shock away from crisis, having built half an industry at full price plus overtime costs.

---

Silicon Desert Mirage: Fabs Built, Supply Chain Broken

The effort to onshore semiconductor production was supposed to be a triumphant return of high-tech manufacturing to U.S. soil. Instead, it’s become a case study in how not to do industrial policy. Taiwan’s TSMC and Intel’s new chip fabs in Arizona – heavily subsidized by the CHIPS Act – have encountered soaring costs, delays, and unforeseen infrastructure needs. As Reuters reported (opens in a new tab), TSMC had to more than triple its planned Arizona investment (from $12 billion to $40 billion) just to get the first fab up and running. Construction has been beset by cost overruns and schedule slips, pushing initial production from 2024 into 2025.

Why the trouble? It turns out building a fab is one thing; operating it is another. These ultra-advanced plants require a whole ecosystem that the U.S. had allowed to wither. For instance, Arizona’s desert environment posed a basic challenge: water. A modern fab uses millions of gallons of ultra-pure water per day to clean silicon wafers. TSMC had to invest (opens in a new tab) in a 15-acre water reclamation plant to recycle and treat water on-site, aiming for “near-zero” liquid discharge. (One industry joke quips that they might as well import Taiwan’s rainwater to Arizona – a jab at how poorly matched the location is to the resource needs.)

On top of that, the highly specialized workforce needed wasn’t readily available in Phoenix. In 2023, TSMC flew in hundreds of veteran engineers from Taiwan to Arizona to train local staff and keep the project on schedule. American labor unions bristled at this, but it laid bare the reality: the U.S. lacked enough skilled semiconductor talent after decades of offshoring. REEx has warned of comparable challenges forthcoming involving REE build-up and operational realities.

Perhaps most damning, even once TSMC’s Arizona “Fab 21” produces advanced chips, those chips cannot stay in the U.S. supply chain. There is no cutting-edge packaging facility domestically to assemble the silicon into finished AI accelerators or smartphone processors. As a result, Tom’s Hardware (opens in a new tab) reports that wafers from Arizona are being shipped 7,000 miles back to Taiwan for packaging before they can go into Apple or NVIDIA products. This ridiculous loop – Made in America, then mailed to Asia for final vital touches – adds cost and time, eroding the very supply chain security that the CHIPS Act was meant to achieve. By one account, the Arizona-made NVIDIA chips will end up “more expensive than those produced in Taiwan” and yield mostly symbolic political wins, rather than real self-reliance.  Time for the commonsensical to wake up.

In short, the silicon comeback touted by U.S. politicians under the Biden administration has been undermined by poor planning and no real inherent understanding of industrial policy. We built the fab but forgot the“fab ecosystem.” As an Apple Insider (opens in a new tab) investigation bluntly noted, TSMC replicated much of its Taiwan setup in Arizona, yet “the U.S. still relies heavily on foreign equipment, materials and expertise” to run it. Lithography machines must come from the Netherlands (ASML), critical chemicals and gases from overseas suppliers, and the highly advanced know-how from TSMC’s Taiwanese workforce. This is what passes for industrial policy: throw money at one link in the chain, assume the rest will magically follow. It hasn’t. And that should set off alarm bells far beyond the chip sector. Where is the U.S. media?

Rare Earths: Same Story, Different Element

The rare earth element (REE) supply chain is eerily poised to repeat these mistakes. Rare earths – as the REEx community knows, it is a group of 17 critical metals like neodymium, dysprosium, and praseodymium – are essential for electric vehicle motors, wind turbine generators, missiles, and smartphones, to name just some more notable products. The U.S. has felt the pinch of dependence here for years. China controls about 85–90% of global rare earth refining capacity and over 90% of permanent magnet production, giving Beijing a chokehold power over technologies of the future.

In response to Trump 2.0’s aggressive trade program, Washington has finally started throwing money at the problem: funding new mines, processing facilities, and magnet factories through Department of Defense grants, the Energy Department, and tax incentives in recent legislation. On paper, it looks like a race to build a “mine-to-magnet” supply chain on U.S. soil.  Under Trump, the Department of War (formerly defense) has taken it a step further, making actual investments in companies such as MP Materials, owner of the only true operational rare earth mine in the U.S.

But here’s the rub: mines and money alone don’t make a supply chain. The bottleneck is in the chemistry and the extended ecosystem, just as it was with semiconductors. Digging ore out of the ground is relatively easy; turning that ore into high-purity rare earth oxides, metals, and then magnets is the hard part. Today, even a good deal of the ore from Mountain Pass, California – America’s sole rare earth mine – has, until recently, been shipped to China for processing because we lacked any full-scale separation plant domestically. That is slowly starting to change (Mountain Pass began refining some light rare earths on-site, and new facilities are planned), but the U.S. is miles behind.

Rebuilding rare earth refining is a far more complex task than many appreciate. And MP Materials deserves a lot of credit for stepping up to the government’s aggressive edicts. After all, separation of rare earths is an arduous,“dirty” process involving hundreds of stages of solvent extraction (SX). Imagine dozens upon dozens of mixer-settlers, acid baths, and chemical tweaks, separating elements that are chemically near-twins. Minor missteps in acidity or temperature can ruin purity, so deep expertise is required. As REEx often reminds, China spent decades mastering this art, tolerating the toxic waste and environmental damage that came with it. Western nations, by contrast, offloaded this sector to China long ago – one reason it’s so dominant now.

Now consider what it will take for the U.S. to stand up its own REE separation and magnet industry. It’s not just about building one shiny refinery. Where will the specialized organic solvents and acids come from? The SX process relies on reagents like P507 (a phosphoric acid ester) and various kerosene-based diluents – products largely made in Asia. Without foresight, a U.S. plant might find itself importing large volumes of Chinese-made extraction chemicals, ironically depending on China to reduce dependence on China. And then there’s the waste disposal: rare earth processing can generate radioactive and toxic waste. Will local communities accept a waste pond, or have we streamlined permits for disposal? Not really – environmental regulations remain strict (for good reason). Does a comprehensive national plan with state buy-in exist for handling the pollution aside from hoping expensive new tech can mitigate it?

The skill gap looms large, too. Only a handful of non-Chinese engineers today have hands-on experience running commercial-scale rare earth separation, according to our sources. You can’t conjure expert chemists out of thin air or fast-track 30 years of know-how with money alone. Just as Arizona’s chip fab needed Taiwanese talent, a future rare earth facility in, say, Texas or California may quietly need Chinese technical consultants or imported expertise to debug all those solvent extraction circuits.  And REEx continues to chronicle how the Chinese government is cordoning off its talent. Even a trip among friends in China today is not the same as a few years ago. The mood is far more apprehensive, tense, and paranoid.

So this would be the ultimate irony – and a real possibility – if we charge ahead without training a new generation of metallurgists and chemical engineers specialized in REEs.

Early signs of trouble are already visible. One flagship project, for example, is MP Materials’ Mountain Pass mine. The company received government grants to build a refining and magnet-making operation. But after two years of effort, magnet output is still nascent, and according to some experts on condition of anonymity, heavy rare earths (like dysprosium) remain unaddressed by domestic capability. To MP Materials' credit, they inked unprecedented deals this summer with the U.S. government, Apple, as well as Goldman Sachs and JP Morgan for a large credit line. But it will take a lot more than one company to rebuild an entire supply chain.

Another is Lynas Rare Earths’ planned Texas separation plant, supported by the Pentagon. It hopes to process heavy rare earths for the U.S., but Lynas will likely import feedstock from its Australian mine, and it may need to import chemical reagents and equipment from its Malaysian operation (or China) to get started. The firm has quietly started shipping the far more difficult heavy rare earths to Japan, yet we are not even out of the first inning of at least a nine-inning game. The point is that each “solution” is only partial—yes, even the purported “mine-to-magnet” deals. Without coordination, we risk siloed projects that don’t connect into a full supply chain.

Billions on the Table, No One at the Wheel

The core issue is the absence of a coherent U.S. industrial policy – an end-to-end strategy to rebuild an industry, not to mention a whole sector. Washington’s approach so far has been a scattershot of subsidies and executive orders, without knitting together all the pieces. Biden’s CHIPS Act and Inflation Reduction Act (IRA) injected huge sums into manufacturing and clean tech. The CHIPS Act focuses on semiconductors; the IRA includes funds and tax credits for critical minerals and EV supply chains.

Yet these initiatives, while laudable in intent, don’t close the loop.

Take the CHIPS Act: it allocated $52.7 billion for domestic chip manufacturing and R&D. Companies eagerly lined up for grants to build fabs. But the legislation and its implementation largely ignored follow-on stages like packaging and testing. Only belatedly has the Commerce Department started to encourage “ecosystem” investments (e.g., Amkor building a packaging plant (opens in a new tab)) after it became embarrassingly clear that U.S. fabs had nowhere local to send their wafers for the next step. In effect, we funded cutting-edge silicon production but not the ability to fully use that silicon in final products – a partial industrial revival at best.

The Inflation Reduction Act, meanwhile, heavily incentivizes domestic sourcing of battery materials (lithium, nickel, etc.) and critical minerals production through tax credits and grants. It even set aside $500 million under the Defense Production Act for essential mineral projects. Yet, when it comes to rare earth magnets – absolutely vital for EV motors and wind turbines – the IRA dropped the ball. The law ties EV tax credits to battery content and final assembly in North America, but imposes no such requirements on the rare-earth magnets in motors. An automaker could meet the IRA’s criteria and still use Chinese-made magnets in the drivetrain, facing no penalty. In other words, the biggest piece of legislation addressing the energy transition ignored the rare earth supply chain under Biden. That’s a glaring omission if the goal were to lessen dependence on China across the board.  While Biden proudly took credit for any wins, the mounting losses, well, not a word.

Then there’s the recent flurry of Trump 2.0 actions. President Trump, during his latest term, initiated a trade war (Liberation Day) that brought critical minerals to the forefront. His administration has invoked emergency measures to encourage domestic mining and even considered using Section 232 tariffs (national security tariffs) on imported critical minerals. President Trump declared a national emergency on critical mineral imports, directing investigations and threatening high tariffs to force a shift.

China’s response was immediate – it temporarily banned exports of certain heavy rare earths and magnets to the U.S., a stark reminder of who holds the cards. Trump’s hardball tactics have indeed sparked a race to build capacity in America (no company wants to be caught dependent if tariffs or embargoes hit). Dozens of rare earth and battery metal projects have been announced or accelerated in the West since the trade war ramped up. And a certain amount of credit must go to President Trump.

But fear, money, and a fresh dose of Trump-era swagger (and yes, troubling early signs of Don-style crony capitalism creeping in) still do not amount to strategy. Without true coordination, we will see the same chaotic pattern repeat: a pilot processor in one state, a magnet plant in another, an extraction demo somewhere else — none of it guaranteed to connect into a functioning, resilient supply chain.

Maybe, this time, likely MP Materials has enough political, economic, and financial gravity to keep Mountain Pass alive through the next crisis. But if China turned off the tap tomorrow (or even next year), would America’s patchwork of early-stage projects keep the country supplied? Not even close.

As Rare Earth Exchanges has emphasized repeatedly, rebuilding an industrial ecosystem is a 5–10 year minimum undertaking — even with unlimited funding. Just permitting a new chemical facility can eat up years. Developing a skilled workforce with real separation and metallurgical expertise takes years. Turning pilot plants into commercially reliable operations takes years. This is a marathon, not a ribbon-cutting ceremony.

Yet Washington keeps behaving as if cutting a check — first under Biden (2022–2024), now under Trump 2.0 (2025+) — magically solves the problem. True Trump 2020 (Operation Warp Speed) did a lot. Many things are possible.  Yes money does not magically solve the problem. It doesn’t. Without a comprehensive, end-to-end strategy, America is simply pouring money into isolated islands of activity and calling it a supply chain.

The High Price of Half Measures

The consequences of this fragmented approach — and what many inside Washington quietly suggest is becoming a disturbingly politicized and insider-driven process — are becoming painfully clear.

American taxpayers are subsidizing factories that cost far more to build and operate than their overseas counterparts – and which still leave us dependent. By some estimates, making chips in the U.S. can be 20–50% more expensive (opens in a new tab) than in Asia once you factor in higher labor and construction costs, plus the inefficiencies of a less developed local supply chain.

We’re essentially paying a premium for sovereignty, yet not getting full sovereignty in return. At this rate, we never will.  Similarly, producing rare earth magnets outside China will be significantly costlier (due to pricier inputs, environmental safeguards, lack of the scale factor when compared to the state-owned entities, etc.), meaning Western manufacturers might struggle to compete unless permanently propped up by subsidies or tariffs. And what’s the probability of that occurring?

What do we get for these extra costs? If done right, it could result in a more secure, resilient supply base and high-tech jobs. But done halfway, we risk the worst of both worlds: exorbitant spending with minimal security improvement. The semiconductor saga already shows this. Despite the pomp around “Made in America” chips, key portions of the chain are still abroad, so the geopolitical risk (e.g., from a Taiwan conflict) isn’t eliminated – it’s just slightly shifted. We’ve built a very expensive fab in Arizona. However, if a crisis hit, we’d discover that we still need Taiwan and others to actually deliver a finished chip to our military or industry. That is not supply chain independence by any stretch.

The rare earth sector, we predict, will be a repeat.

Imagine 2028: several U.S. rare earth mines are operating, and a couple of separation plants have finally opened with government help. Yet if those plants import Chinese solvents, or if their output still has to be sent to foreign facilities for final purification or alloying, then a Chinese export ban involving not only ore, metals, and magnets but also chemical inputs and processes know-how would still cripple us. We’d have spent billions to domesticate REEs and still end up scrambling in a crisis. It’s a very real scenario if each link (mining, refining, alloying, magnet-making) isn’t connected domestically or with trusted allies.

Another “price” to consider is the opportunity cost of not thinking holistically. The U.S. could spend its finite dollars smartly. A comprehensive industrial policy might identify choke points and fund infrastructure for all critical inputs – e.g., incentivize chemical manufacturers to produce reagents domestically, fund training programs for specialty chemical engineers, streamline permits for facilities that handle waste, and build the end-use factories (like magnet plants and chip packaging plants) in tandem with the upstream.

Instead, we have siloed funding: CHIPS Act for fabs (with packaging as an afterthought), separate DOE grants for critical minerals, DoD contracts for magnets, etc., each marching to their own tune. No single entity is ensuring that when a new rare earth refinery opens, the acids, electricity, water, skilled staff, and offtake customers will all be in place and synchronized.

Without that coordination, the price we pay will be ongoing: higher consumer prices for products, continued vulnerability to supply shocks, and perhaps the failure of some heavily subsidized ventures that find they can’t operate competitively or have to pause for lack of inputs. The lack of industrial policy in America isn’t just an academic issue – it will hit our wallets and our national security. As one sobering example, experts estimate it will take at least five to ten years of focused effort for the West to replicate even a semblance of China’s rare earth supply chain dominance.

During that time, China will not sit idly; it’s already moving to tighten its grip, as seen with recent export restrictions and technology bans threats.  The Chinese are relentless in patenting downstream innovative use cases, planning to own the future while we play catch up today—in a half-effort manner. We could find ourselves always several steps behind, reacting to the latest move without a long-term game plan – a costly and dangerous position, that unfortunately is our reality in the USA today.

A Call to Action: From Catch-Up to Coordination

The cautionary parallels between semiconductors and rare earths send a clear message: America needs a real industrial strategy, or our “tech independence” efforts will remain a mirage. One that, unfortunately, we fear politicians and Beltway insider networks monetize along the way. It’s not enough to bankroll one factory or declare victory after one groundbreaking ceremony.

We must think in ecosystems. That means asking hard questions upfront: If we build it here, do we have everything to run it here? If not, how do we develop those missing pieces now, not as an afterthought?

Policymakers should urgently consider a coordinated approach – perhaps an “Industrial Policy Council” that bridges the siloed programs. A Critical Mineralczar to coordinate and direct across the vast sprawling federal government remains key. The CHIPS Act could be coupled with packaging and materials initiatives in a single supply-chain framework. Rare earth projects should be linked from mine to magnet, ensuring that funding and incentives cover each link (mining, refining, alloying, magnet manufacturing, recycling) and the connections between them—importantly, across companies both in USA and abroad in allied nations.

Environmental permitting for critical facilities might need streamlining so that plants can actually get built in time. Alliances with trusted countries (like Australia, Japan, and EU nations) should be leveraged not only for raw materials but also for technical knowledge exchange – e.g., bringing over rare-earth experts or partnering on reagent production, so we’re not secretly beholden to Chinese inputs.

Conclusion

The alternative is to continue stumbling forward with half-measures. The U.S. chip fiasco in Arizona is a $40 billion lesson in the importance of foresight. Let’s not learn the same lesson again five years from now when a much-hyped rare earth supply chain still can’t stand on its own. America pioneered the semiconductor and built the first rare earth magnets decades ago; we have the ingenuity. What’s lacking is the political will and long-range vision to orchestrate a full revival of these industries.

In a world where supply chains equal power, partial independence is no independence at all. It’s time to shed the complacency, the short-term politics, and the creeping insider favoritism now coloring key decisions — and finally plan our industrial resurgence down to the last bolt and last drop of acid. Otherwise, we’ll keep paying top dollar only to discover that, when it really counts, the supply chain’s weakest links were never fixed – and they lie somewhere over the Pacific.

 © 2025 Rare Earth Exchanges™ – Accelerating Transparency, Accuracy, and Insight Across the Rare Earth & Critical Minerals Supply Chain.