This is a two part series on the conflict between the United States and China over Taiwan, a topic that I acknowledge has been quite exhaustively written about. I thought I’d do a nice twist on the genre. Part I is non-fiction and sets up the second part, a fictional sort of alt-future-history, extended into the future.

Part I will discuss the nature of the conflict over Taiwan, to show that it is primarily a conflict over regional hegemony. It will then discuss what are the best strategic options for the United States. Part II is a fictionalization of what might happen if we fumble the ball on that strategy.

Introduction to Part I

Let’s establish, what I believe, to be the main crux of the argument: Countries desire to be regional hegemons. It is a lucrative and powerful spot to be in. It’s good to be the king. Being the regional hegemon of East Asia gives you a springboard to global hegemony. It’s a natural desire for China, with great historical precedent, and it is the main reason for China’s quest to take Taiwan.

China’s goal in taking Taiwan is to retake the mantle of regional hegemon. Taiwan is simply a Schelling/focal point about which regional hegemony and signals of credible commitment now revolve, creating a mechanism to show that the United States’ security architecture is not relevant—effectively removing the competing power in the region.

Conversely, the United States has long adhered to a foreign policy of preventing the rise of a hostile hegemon in East Asia or Europe, understanding that such dominance would expose its strategic flanks. China’s drive for dominance is therefore pitted against America’s longstanding role as a regional balance. In response to this assertion of Chinese power, the United States has made its own military strength the backstop and foundation stone for an anti-Chinese regional coalition (in Elbridge Colby’s terminology, an anti-hegemonic coalition).

China is currently threatening the regional architecture of East Asia, which places the United States as a counterweight to its ambitions. Apart from the barebone realities that define such a contest—military power projection and economic power—the United States and China both face choices, responses, risks, and rewards to manage. The evolution of this contest could result in a variety of outcomes, ranging from a vague reset of the status quo (a “grand compromise”) to a more pessimistic “China Triumphant.”

While I do believe that regional hegemony underpins China’s ambitions in Taiwan, two other plausible explanations for its focus on Taiwan are often cited: chips (e.g., TSMC) and “national unification.” These arguments are ultimately not convincing.

China’s domestic chip capacity, including lower-node chips, is advancing significantly. It could achieve technological dominance without risking a military invasion of Taiwan. National unification and domestic causes are certainly more convincing than any others, but if China wanted national unification, they would be doing quite a bit differently.

Another meta-argument is that China is not an expansionist power and only seeks to “reunite” Taiwan, consistent with a national policy rather than regional domination. This is patently false. China has been the regional hegemon in Asia almost continuously since the Han. When its dominance was unimpeachable, few challenged it. When it was not, China historically acted—often forcefully—to reassert control.

Why be a regional hegemon?

The allure of regional hegemony extends far beyond mere prestige—it is about power, security, and economic domination. It’s the foundation stone for achieving global hegemony.

The United States’ rise as the undisputed regional hegemon in the Americas during the 19th and early 20th centuries offers a nice blueprint. The United States established a distinctly American form of dominance over its weaker neighbors—sometimes beneficent, sometimes coercive.

Neighboring countries like Mexico and Cuba became economically reliant on the United States, supplying raw materials while importing U.S. manufactured goods. This dynamic fueled the United States’ industrial expansion. Crucially, the regional security this dominance provided served as a springboard for the United States to project power globally. No competitors in the region meant they could look further afield.

China’s strategy in East Asia is no different. Unlike Latin America during the United States’ rise, which primarily supplied raw materials, East Asia is home to established industrial economies (and plenty of raw materials, like those in SE Asia) that both compete with and depend on China’s manufacturing dominance. China can both extract resources but also integrate regional economies into its industrial and trade networks, barrier free.

For China, the path to achieving regional hegemony runs through Taiwan, a linchpin with immense symbolic and strategic value. Taiwan is not only a challenge to the United States’ credibility but also a decisive test for the anti-hegemonic coalition in the Indo-Pacific. Successfully taking Taiwan would deliver a crippling blow to this coalition, discrediting the United States’ ability to defend its allies and uphold its security commitments. Such a failure would erode trust among regional partners like Japan, South Korea, and the Philippines, undermining the collective resistance to Chinese dominance.

Taiwan’s fall might trigger a bandwagoning effect, where states realign with the dominant power to secure their interests, allowing China to reshape the regional order to its advantage. Taiwan represents the fulcrum for consolidating Beijing’s regional supremacy.

Taiwan is not merely a territorial dispute, an issue of national pride, or a matter of semiconductor production. Instead, it is the centerpiece of China’s strategy to reset the historical order and secure its future as a global superpower.

Chips are not the goal

First, let’s consider the notion that China’s potential conquest of Taiwan is primarily driven by a desire to acquire advanced semiconductor technology, particularly from Taiwan Semiconductor Manufacturing Company (TSMC). The main reason this argument falls short is that it overlooks China’s substantial investments in developing its own domestic chip production capabilities.

China has been aggressively investing in its semiconductor industry to reduce reliance on foreign technology and enhance self-sufficiency. In 2024 alone, China allocated approximately $41 billion to wafer fabrication, marking a significant increase from previous years. Leading Chinese companies, such as Semiconductor Manufacturing International Corporation (SMIC) and Hua Hong Semiconductor, are spearheading these efforts. Beyond SMIC, China has also been expanding its domestic semiconductor equipment manufacturing sector, with firms like Naura Technology Group—China’s largest semiconductor equipment manufacturer—making significant strides. These companies are steadily advancing their capabilities and are on track to achieve technological parity with TSMC, including the ability to produce chips using EUV (extreme ultraviolet) lithography.

Given these substantial investments and developments, it is implausible that China would undertake the immense risks associated with a military invasion of Taiwan solely to acquire semiconductor technology. If semiconductor access were truly the goal, China could negotiate a deal with Taiwan over independence in exchange for this technology—a far less risky and more practical approach.

National Unification and Nothing More is Not the Goal

The assertion that China's interest in Taiwan is solely about national unification, without broader ambitions for regional hegemony, is simply ahistorical hogwash. Throughout its history, China has engaged in both direct territorial expansion and the establishment of hegemonic influence over neighboring regions.

China has, quite to the contrary, used violent coercive force, sometimes invasion, sometimes simple threats of force, to expand its own directly ruled territorial boundaries and, far more often, draw neighbors into a regional hegemonic order, often called “The tributary system”. Additionally, in a not so surprising historical observation, whenever Asia became more competitive among national polities, China has become more aggressive and expansionist. The Qing seem to have quite the record of violent invasion of their neighbors.

Let’s start actual territorial expansion through history, since that one is seen as rarer. A map might help show the silliness of the “only national unification goal”. The Han Empire’s borders are on the right (note Vietnam) and the Qing’s borders are on the left.

I am not a geography expert, but this does look like territorial expansion. Such territorial expansion was a consistent feature of Chinese history, especially in the Qing. To list a few:

Southern Expansion: Various Chinese dynasties extended their control southward, often into regions like Northern Vietnam. For instance, during the Ming Dynasty, China launched invasions to assert dominance over Vietnam.

Central Asian and Mongolian Campaigns: China's historical expansion also extended into Central Asia and Mongolia. The Tang Dynasty notably expanded control into Central Asia, reaching as far as present-day Kazakhstan. This expansion was marked by significant military campaigns, including the Battle of Talas in 751 AD, where Tang forces confronted the Abbasid Caliphate.

Qing Dynasty Conquests: The Qing Dynasty expanded China's territories into areas such as Manchuria, Xinjiang, and Tibet, frequently through military conquest. These expansions were driven by strategic and security considerations, aiming to consolidate China's influence over vital frontier regions.

Beyond direct territorial control, China has historically sought to establish a hierarchical regional order. This system positioned China as the central authority, with neighboring states acknowledging its supremacy in exchange for political legitimacy and economic benefits (such as the right to trade in China).

While some narratives portray this arrangement as a harmonious Confucian order, it was just a brutish regional hegemony under a different name. This arrangement allowed China to exert significant influence over its neighbors' political and economic affairs, ensuring a regional hierarchy that favored Chinese interests.

The so-called tributary system was less about mutual respect and more about consolidating China's regional supremacy, with Confucian ideals serving as a veneer for realpolitik strategies. And China has used force to maintain this system, many many times (many more times than it used force for territorial expansion). Korea and Vietnam, which fell in and out of under direct Chinese territorial influence and indirect rule, are the biggest “beneficiaries” of this sort of violence:

The Ming and Qing in Southeast Asia: The Ming invaded Vietnam in the 1400s as punishment for Vietnam’s attempts to conquer Champa and installed a puppet dynasty to enforce control. China used veiled threats just as often, for example, warning Thailand and the Majapahit against attacking the Malacca Sultanate, placing Malacca under Chinese protection as a protectorate. The Qing invaded Burma and Vietnam several times.

The Ming and Qing in Korea: The Ming don’t invade Korea to conquer it per se, but instead act to maintain a puppet relationship with them by protecting them from a Japanese invasion. The Qing however, do just straight up invade.

Often you hear critics of the idea of “aggressive China” look at this data and point out that China fights relatively few wars vs. their Western counterparts. This might seem to be the case. After all, Europeans conquered entire continents. This overlooks the fact that its overwhelming dominance often precluded the need for frequent military interventions. The mere presence of China's power served as a deterrent, ensuring compliance from neighboring states without the necessity of force. But when it needs to fight, it does.

It’s not that, as is often claimed, the historical record exonerates China from accusations of territorial and hegemonic ambitions, making Taiwan just an “Internal affair”, it does quite the opposite.

State of Play

Now that we understand the stakes, what can the United States do about it? How does this all play out? Chinese regional hegemony would be a net negative for the world. While the United States is an imperfect democracy, it has promoted free trade and liberalization across the globe. In contrast, China is an autocracy that seeks to crush dissent and, just speaking purely of self interest, likely deny the United States economic access to East Asia if it achieved dominance.

A lot of foreign policy thinkers want to give agency to the United States, but the truth is that the United States is the acting global hegemon, defending an increasingly embattled position. The United States can do a lot, but in many ways, it is white (China) to move and we are responding to whatever they do next.

I’ve created a nice handy chart to help try to conceptualize all the possible outcomes and choices.

To conceptualize the possible outcomes and choices, I’ve created a framework based on the interactions between China and the United States. China faces three key decisions, while the U.S. has two. Together, these choices lead to four potential outcomes.

The Choices

Choice 1 (US), Defend or Compromise: The United States must decide whether to defend the status quo or compromise with China. Backing down would involve negotiating a “grand compromise” with Beijing, conceding significant influence in the region, but still having something left to show for it. Defending the status quo, on the other hand, would mean deterring China through measures such as weapons sales to Taiwan and strengthening the existing security architecture in Asia.

Choice 2 (China), Be Deterred or Attack: If the United States chooses to deter, China must decide whether to maintain the status quo or escalate toward a kinetic solution. Choosing deterrence would mean accepting that regional hegemony is unattainable. Opting for military action, however, is a high-risk strategy that could result in achieving regional dominance if successful.

Choice 3 (Mutual), Who Will Back Down:If a conflict arises, both sides face the ultimate decision of whether to back down. In a war between two nuclear-armed powers, it is unlikely that either side could deliver a decisive defeat, but eventually, one must concede. This concession could occur due to economic strain, military stalemate, or a fait accompli. The forms of conflict could range from a blockade to an amphibious invasion, each with varying degrees of escalation and consequences.

Possible Outcomes

Outcome 1, Grand Compromise: If the United States determines that China cannot be deterred and that a war would be too painful, risky, or unwinnable, the best option may be a “Grand Compromise.” While imperfect, the U.S. has compromised with authoritarian regimes before, as seen in the post-World War II division of Europe with the USSR. This outcome would likely leave Southeast Asia within China’s sphere of influence, Taiwan under Chinese control, and the U.S. maintaining a military presence in Korea and Japan. The specifics of the compromise would determine whether it ultimately favors one side more than the other.

Outcome 2, China Deterred: This outcome is the least disruptive to the world order. The United States successfully signals its permanent economic and military presence in the region, deterring China from further aggression. This buys the U.S. time to rebuild its military capacity and wait for China to potentially enter a secular decline (don’t bet on this one, friends). Regional stability is preserved, at least temporarily.

Outcome 3, US or Chinese victory: If war occurs, the victor will shape the region's future.

• U.S. Triumphant: A decisive U.S. victory would restore the status quo ante bellum, reinforcing the existing security architecture that has underpinned peace and stability in East Asia for decades. It would reaffirm the United States’ commitment to its allies and its role as a global leader.

• Chinese Triumphant: A Chinese victory would dismantle the current regional order, establishing Beijing as the unchallenged hegemon in East Asia. This outcome would represent a fundamental realignment of global power dynamics, signaling the decline of U.S. influence and ushering in an era of authoritarian dominance. It would be a grim future for the liberal world order.

Conclusion

The United States should prioritize deterrence as the best strategy to preserve the status quo and avoid catastrophic conflict. However, if military defeat appears likely, pursuing a grand compromise may be a better alternative than engaging in unwinnable war. The grand compromise outcome might seem un-American, but it might be the only available, depending on how things unfold. Even a military victory would likely be pyrrhic, resulting in heavy losses and would almost certainly set the stage for future conflicts.

This is a two player game afterall, so for China too, direct confrontation carries significant risks. Failure in a military campaign could be devastating. It will all come down to how strong China thinks it is (perception of strength vs. reality does often have a delta, especially for authoritarians). While Beijing might prefer a grand compromise if it guarantees meaningful gains, the lure of decisively reshaping the regional order could still push China toward escalation if leaders perceive a high probability of success.

The United States must act on two levels:

• Immediate Actions: Invest in more deterrence based military capabilities, so as to create an asymmetry with China, strengthen alliances with Japan and other pivotal players, and secure critical supply chains through onshoring.

• Long-Term Strategy: Accurately assess the balance of power, anticipate shifts in regional dynamics, and align policies to manage risks and opportunities effectively.

The ramifications of screwing this up are massive. And that’s what part 2 is all about. In Part 2, I will delve into a speculative "history of the future," taking the perspective of a historian in 2040, reflecting on a world where China emerged triumphant. This imagined retrospective will explore how China's victory over Taiwan set the stage for its rise as the unchallenged regional hegemon and the ripple effects that reshaped the global order.

Introduction

China's recent decision to ban the export of critical minerals, including gallium and germanium, to the United States was a long time coming. The Chinese had been tightening the rare earth metal noose for quite some time, especially those rare earth metals that are used as semiconductor substrates (called III-V materials, after their valence electron structure), which is what we’re going to talk about today. Everyone sort of expected it in the defense electronics industry, at least. Germanium is less of a problem. It’s only 50% in China. Prices will go up, but we shouldn’t panic.

Gallium, however, is a real issue. While there is significant Gallium inside the United States, we haven’t mined it since 1987, Germany not since 2016. China controls upwards of 90% of world Gallium mining. There is significant uncertainty on its downstream impact, especially on defense production. I’ve asked around to some of the players in the industry and have gotten vague answers. The stockpile levels are opaque (purposefully so).

This move is widely viewed as a direct response to U.S. sanctions aimed at curbing China's technological advancements, particularly in the military sector. To be honest, we should recognize that fair’s fair. The Chinese kicked us back in the shins in a spot where we’re vulnerable.

First let’s dig into the “why”. Why are these materials so important? And then we can dig into what I expect the ramifications to be, specifically for Gallium, since Germanium is less of a risk factor.

Who Cares?

You probably think about semiconductors as mostly being on silicon. We make sand think, not rare earth metals! And yet, a minority of chips, maybe 10% of total semiconductor volume, but an important minority, are made of III-V materials. The most common materials are Silicon Germanium (SiGe, which is pronounced like “Siggie”) and Gallium Nitride (GaN), and Gallium Arsenide (GaAs). Now, the problem is that these components are hard to circumvent and perform critical RF functions, due to their unique electromagnetic properties. As a result, the margins on these bad boys are insane. They can cost 10-20x a normal RF chip. You have one in your cell phone and you have one in a ton of medical devices and cell towers. But those are safe. Many are currently fabbed in Taiwan and more civilian chips will end up there due to these sanctions. The real danger we need to understand is on military production, which primarily happens domestically (for the most part).

First of all, China didn’t ban the export of SiGe, GaAs, or GaN. It banned the export of Gallium and Germanium itself (and a few other rare earth materials). These are then used as inputs, combined with Silicon, Arsenide, and Nitride and grown into the crystals that can be used as wafers.

III-V materials are known for their thermal and electric properties. These materials have high electron mobility, allowing electrons to move swiftly, which is crucial for high-speed and high-frequency devices. This means they can handle rapid signal processing (think switching). Their wide bandgap also allows them to operate at higher voltages and temperatures, due to the heat dissipation of their crystal structures. This means they can deal with a ton of power, without starting to literally smoke (thought they sometimes do).

As the chart above makes clear, there are indeed substitute materials. You can design around the shortage, to an extent. For some applications, using a less performant material might work. You might take the chip you wanted to do in GaN and do it in SiGe (Germanium being less of a worry). If you wanted to do something in SiGe and prices went up, you can do it in plain old Silicon. The performance is worse, but it is possible.

And yet, sometimes you can’t escape their use. It’s just not avoidable in military communications, due to their exceptional thermal properties and power handling capabilities. Simply put, if you want to blast a lightbulb, or often when put in an array, hundreds of lightbulbs, worth of RF power into space you’re going to need a GaAs or GaN chip (that’s orders of magnitude more RF power than your cell phone). And unfortunately, many military applications need to do just that power blasting task.

Systems such as advanced radar, electronic warfare equipment, and secure communication devices rely on the superior performance of III-V semiconductors to function effectively. In modern warfare, especially a war against China itself, these sorts of weapons systems are going to be important.

China's Market Dominance: How Did We Screw This Up?

Like many things, the fact that China got such dominance in the III-V chip materials market and the Free-Tradeniks in our policy elite let it happen is beyond me. Let’s chalk it up to incompetence. But it’s a story you’ve heard many many times:

1. “This whole thing is very environmentally bad. Let’s make it impossible to fulfill regulations to mine it here.”

2. “Well, cool, we did it…now we have more domestic industry. Let’s buy it abroad.”

3. “Ah well…the Chinese have poured insane subsidies into putting everyone else but them out of business. Guess we’ll have to trust in the magic of free trade.”

For this section, let’s focus on Gallium. Germanium is another story, and frankly, it’s going to be fine. We mine plenty of it in Alaska and can scale it up.

To be clear, China's grip on gallium production was a purposeful push to dominate an industry that they saw as vulnerable in the West, but very essential. It was strategic and it was smart. The trick was that China already had quite a large Aluminum mining industry, going back to the early days of Communist China, and Gallium was a marginal investment. This is because Gallium is mainly a by-product of aluminum production from bauxite ore. China just had to integrate gallium extraction into this sector. Between 2000 and 2022, China's aluminum output skyrocketed from 4.2 million to 40.2 million tons. Gallium production followed alongside.

As China started to see the first inklings of what would become the trade war between the US and China, back in 2016, China made a play for a dominant position, with foreknowledge of the leverage it provided against US defense production. China used its classic tools of state subsidies and competitive pricing and flooded the market. This aggressive pricing strategy has driven most global producers out of business, leaving China as one of the world's only remaining producers of Gallium and Germanium.

In contrast, the U.S. and Germany just threw their hands up, mostly due to stringent regulations that have hampered domestic gallium production. Environmental concerns and complex permitting processes made it just too challenging to develop new mining operations or expand existing ones. This regulatory environment snuffed out the U.S. gallium industry in the 80s. Germany stopped production in 2016.

Assessing the Damage

As I stated above, the majority of the commercial processes will be fine. Most of them, around 50%, are fabbed at Win Semi, a fab in Taiwan. So commercial III-V Chips really aren’t the issue. It’s defense, which is the point. The problem that we face is while there are workarounds today, if we basically do away with all regulations, we will need a few years to get our own mining operations in place (there’s plenty of Gallium in the United States, for example in Texas and Montana). If we don’t basically zero out compliance, we’re toast.

There’s a few pieces of good news that should lessen our immediate panic. First of all, we do know that there is a strategic stockpile/reserve in the US, somewhere, somehow. People I’ve spoken to at defense fabs seem to be sanguine about the situation. The other piece of good news is that defense volumes are fairly low. Most wafer runs are in the thousands and tens of thousands. We’re not going to run out of these chips tomorrow (though prices will go up). If things get dicey, the United States does also have Gallium and Germanium recycling operations.

The bad news is that we do need someone outside of China, ideally in the US itself, to get up and running eventually, and most mines, thanks to China’s incredibly obvious strategy, are not currently functional for III-V metals. And the road to getting things set up again could be arduous.

The steps in this process are the following:

• Resource Assessment and Feasibility Studies (1–2 years).

• Permitting and Regulatory Approvals (2–3 years).

• Infrastructure Development and Construction (2–4 years).

• Commissioning and Ramp-Up (1–2 years).

That’s a grim amount of time. There’s some options in Hungary, Kazakhstan, and Germany and the infrastructure there is more reasonable. They only stopped production in the late 2010s. That might be 1-2 years. For the US, our only option is to basically get rid of any feasibility studies, assessment steps, and regulatory approvals. Zero them out. And then, we get to around 3-4 years.

Conclusion

To summarize: These things are important for defense applications, especially electronic warfare and communications. Perhaps China is cutting us off from them to show force, leverage in future negotiations, or because they rightly want to kneecap us the way we kneecap them. It’s a move that we should respect and should have expected.

Additionally, although we're not going to run out tomorrow, we need to get ready and it’s frankly a little surprising to me that we’re not already moving forward on initiatives to start Gallium mining. Maybe there are and it’s not public, but I am skeptical.

We should take this whole thing as a “learning moment”: Free trade is a national security risk when you’re obtaining a critical component from your adversaries. Overregulation killed an industry, one of many, and we now have to make the right decisions, not just for Gallium, for America’s industrial and mining base as a whole.

The CHIPS Act represents a critical step toward securing America's semiconductor industry. But opportunities for improvement remain. While the Act has already shown promise in attracting major investments from chipmakers like TSMC and Samsung, targeted reforms could significantly enhance its effectiveness. This analysis presents a comprehensive framework for optimizing the CHIPS Act through three key mechanisms: strategic use of tariffs, regulatory streamlining, and additional support measures, including targeted Chinese semiconductor tariffs and a national chip reserve.

Introduction

The semiconductor industry has become fundamental to both modern economic power and military capability. While the United States maintains dominance in areas like equipment manufacturing, chip design, and software, it lags in other areas. Most notably, semiconductor fabrication — the actual process of manufacturing chips — has largely moved overseas. This shift wasn't driven by comparative advantage or economic inevitability, but by historical policy decisions that now demand correction. In 1990, the United States represented 37% of global fabrication capacity. Today it represents around 10%.1 The CHIPS Act, which aims to bolster U.S. semiconductor fabrication capabilities, represents a necessary (though imperfect) targeted industrial policy intervention, with a goal of getting the United States back to a lean 20% of global chip production.  The incoming administration has the  chance  to optimize its implementation.

There is an argument, frequently made by industry analysts, that semiconductor fabrication should be left to other nations as part of the global supply chain, with some even suggesting that Asian countries possess some inherent advantage in chip manufacturing. I remember seeing one online commentator point to “Confucian values” as a reason why industrial policy around chips will fail in America. 

This argument falls apart under historical and economic scrutiny. Historically, the decision to outsource fabrication was as much about integration of East Asian allies into American supply chains as it was about any economic reality. From an economics perspective, much of the raw material for fabrication is already extant in the United States. American companies already manufacture most of the world's semiconductor equipment — the complex machinery required for chip fabrication. U.S. firms like Intel, Micron, and Global Foundries operate some of the world's most advanced fabrication facilities, consistently demonstrating American excellence in this sector. The notion that complex manufacturing requiring advanced engineering somehow cannot succeed in the United States contradicts both historical evidence and current reality. U.S. manufacturing remains the most productive on the planet. 

More importantly, the strategic stakes are too high to accept this flawed logic. China is aggressively (unprofitably) subsidizing domestic semiconductor production, aiming to dominate the sector. That alone creates an untenable strategic vulnerability. Meanwhile, Taiwan, which currently houses much of the world's fabrication capacity, faces growing geopolitical pressure from China.  With the risk of Taiwan falling under Chinese control, the strategic vulnerability becomes existential.

The CHIPS Act serves dual purposes: near-term national security, the need for which is outlined above, as well as long-term development of a resilient domestic semiconductor fabrication industry. While the United States dominates semiconductor design — with American firms like Broadcom, NVIDIA, Intel, and Texas Instruments controlling about 60-70% of the market — we've become dangerously, and perhaps inefficiently, dependent on overseas fabrication. Firms like Broadcom and NVIDIA make their chips in overseas fabs, primarily in Taiwan. While the United States dominates semiconductor design - with fabless American firms controlling about 60-70% of the market - we've become dangerously, and perhaps inefficiently, dependent on overseas fabrication. By taking back fabrication to America is not only strategically intelligent,  it is sound economic policy, too, as it increases  verticalization and reduces foreign dependency. 

The primary challenge is financial. Fabrication facilities require massive capital investment, often in the tens of billions of dollars. These costs have historically deterred domestic construction. The CHIPS Act aims to bridge this gap, incentivizing both foreign and domestic firms to build in the United States.

Recent statements suggesting complete repeal of the Act have been misattributed to President-elect Trump. I am confident he has no interest in such a counterproductive action. But to the extent he has criticized the Act, then he and the new GOP majority in Congress should seize the opportunity to fix that which they deem broken by adopting three  narrowly-tailored and linked reforms: (1) strategic application of tariffs based on successful precedents from the previous administration, (2) a strategic chip reserve, (3) streamlining regulatory requirements and thoughtful changes to the financial structure of subsidies (particularly “Notice of Funding Opportunity”, or NOFO clauses). These core improvements, supplemented by targeted yet massive tariffs on Chinese semiconductors and a strategic chip reserve, could significantly enhance the Act's effectiveness in achieving its stated goals.

The CHIPS Act and its Challenges

Industrial policy succeeds best when it pushes against something targeted.The CHIPS Act does exactly that, or at least it aims to. The fundamental challenge is that building semiconductor fabrication plants in the United States costs approximately 30% more than it does in Asia. The Act aims to bridge this gap, making domestic production economically viable. As we will see, though, the major problems with the act are overregulation and overreliance on subsidies.

I am going to unload a lot of criticism here but we should stay directionally positive. The CHIPS Act has indeed already shown promising results. TSMC and Samsung are building major fabrication facilities in the United States, with early reports suggesting the Arizona fab is achieving yields 4% higher than comparable facilities in Taiwan — though this data comes from early, limited production runs.2

To give a brief rundown of the bill, The Creating Helpful Incentives to Produce Semiconductors (CHIPS) and Science Act, despite its awkward acronym, was enacted in August 2022. The Act allocated $280 billion over the next decade, with $52.7 billion dedicated to semiconductor manufacturing, research and development (R&D), and workforce development. An additional $24 billion provides tax credits to stimulate chip production, while $3 billion more supports wireless technology development.

China's aggressive moves in the semiconductor industry largely precipitated the CHIPS Act. The Chinese government is investing heavily to dominate the market for mature-node chips (25nm and above), targeting 50% market share by 2030. This strategy extends beyond mere self-sufficiency — it positions China to control global supply chains for everything from consumer electronics to military hardware. Combined with the risk of China seizing Taiwan's advanced chip fabrication capabilities, the threat  has become too serious to ignore. You can read more about this in another article I wrote.3

The Act relies heavily on subsidies, and subsidy-like instruments, for driving investment and results. While subsidies are a necessary component of any serious industrial policy they cannot succeed alone. The Act needs additional mechanisms which can complement subsidies: tariffs, to protect these investments and ensure long-term commitment from firms, and a strategic chip reserve, to give firms a guaranteed income stream. 

Additionally, the Act actually often counteracts the subsidies it relies on by forcing firms to comply with an extensive array of “everything bagel” regulatory requirements, from environmental to reporting to labor (and DEI for good measure). This slows down the distribution of funds and undercuts the subsidies through compliance costs. 

The most painful regulation is the “NOFO” clause. Under this provision, “[r]ecipients receiving more than $150 million in CHIPS Direct Funding will be required to share with the U.S. government a portion of any cash flows or returns that exceed the applicant’s projections (above an agreed-upon threshold specified in the award).” Basically, the government can ask firms to return any revenue stream above a certain level if the government’s investment in that firm ends up being “too successful.” 

A corollary of the regulatory burden is the slowness of fund distribution. Bureaucratic delays are killing the Act's momentum. Instead of rapidly deploying capital to accelerate domestic chip production, the program has become mired in administrative complexity. The fact that Intel, a key domestic producer, hasn't received any funding to date exemplifies this paralysis. This is particularly dangerous given the pace of global competition — Europe is investing $47 billion (with Germany adding another $20 billion) in semiconductor initiatives, while China is committing more than $300 billion to dominate the industry.

Let’s dig into the solutions: tariffs, chip reserves, and regulatory streamlining. 

Tariffs as an Instrument

Tariffs are often dismissed by free-market purists as distortions that raise consumer prices and provoke retaliatory measures. This skepticism is understandable. However, when applied selectively to strategic industries, tariffs can ensure long-term commitment from foreign firms and protect the massive initial investments that domestic firms make.4 If you want to make the CHIPS Act subsidies successful, coupling them with tariffs is needed. We can see this with two examples tariffs from the first Trump administration, plus a historical example from the 1980’s in the semiconductor industry. 

While not universally successful (the general tariffs across manufacturing and agriculture produced mixed results)5 Trump’s strategic tariffs from the first term did achieve some measure of success for narrow industries, particularly those on washing machines and steel. The washing machine tariffs are an example of getting foreign firms to behave well. And the steel tariffs are an example of protecting net new domestic manufacturing investments. Both of these are things we would want coupled with subsidies. 

Washing machines provide an informative case study about how to get foreign firms to invest the way you want and how to protect them while they do so. These targeted and strategic tariffs, implemented in January 2018, ranged from 20% to 50%. The tariffs triggered significant “tariff-jumping investment,” with both Samsung and LG establishing major U.S. manufacturing plants to avoid import duties. Samsung invested $380 million in South Carolina and LG invested $360 million in Tennessee, each creating around 1,000 jobs. Contrary to concerns, washing machine prices quickly stabilized as these new plants began production (as seen below).6

The steel tariffs offer another instructive example of protecting domestic investment. After Trump imposed 25% tariffs in 2018, U.S. Steel and Nucor committed over $5 billion to modernizing their domestic facilities. Without protection from cheap imports, these massive capital investments in more efficient blast furnaces and electric arc facilities would have been too risky. Just as with washing machines, critics warned of devastating price effects, but the tariffs succeeded in their core goal - protecting the substantial investments needed to upgrade American steel production capacity.

The semiconductor sector's own history with tariffs, particularly the U.S.-Japan DRAM conflict of the 1980s, offers crucial lessons for today's competition with China. Japan had emerged as a dominant force in DRAM (Dynamic Random Access Memory), slashing U.S. market share from 70% to 20% by 1986 through better manufacturing and aggressive pricing. When Japan failed to honor initial agreements to stop chip dumping, the U.S. imposed 300% tariffs. While some dismiss this intervention because Intel soon pivoted to microprocessors, the tariffs achieved their goal — they saved America's DRAM industry. Micron not only survived but thrived, eventually capturing 20% of Japan's domestic market.

These historical examples suggest two ways to strengthen the CHIPS Act:

First, Trump can use lighter tariffs on all foreign produced chips (in the 10-20% range) as a protective measure to augment subsidies. These can protect domestic chip production as it ramps up and give proper incentive for foreign firms to invest, creating the same "tariff-jumping investment" effect we saw with washing machines.

The DRAM story shows how tariffs can be a cudgel. Trump can ensure TSMC and Samsung complete their investments through the potential threat of punitive tariffs if they do not. A credible commitment to this policy will prevent companies from abandoning their investments midway through. 

Strategic Chip Reserves: Guaranteeing Demand and Security

A strategic chip reserve was considered during Trump's first administration but never implemented. While this proposal might seem unconventional, it serves both immediate national security needs and longer-term industrial strategy goals.

The national security case is straightforward. By maintaining stockpiles of critical semiconductor components, the government could ensure continuity of defense systems and critical infrastructure during supply disruptions. Recent shortages have demonstrated how vulnerable our military and industrial base are to semiconductor supply chain interruptions. A strategic reserve would provide a buffer against both short-term disruptions and potential geopolitical crises, particularly any involving Taiwan. 

The industrial strategy benefits are equally compelling. By tying access to reserve contracts to specific domestic production targets, the government could provide semiconductor manufacturers with guaranteed demand — effectively creating a stable revenue floor for new facilities (one could even imagine that a revenue stream might be better for some firms than a subsidy). This approach has historical precedent in other strategic industries. The Strategic Petroleum Reserve, for example, helped sustain domestic oil production during market downturns. Similarly, defense procurement contracts often include minimum purchase guarantees to sustain critical manufacturing capabilities.

The reserve's composition would need careful consideration. Military and industrial applications require a broad spectrum of semiconductors, from advanced logic chips to mature-node analog devices. Defense systems often rely on specialized chips that may be several generations behind the cutting edge, while industrial control systems frequently use robust, proven designs rather than the latest nodes. The reserve would likely need to maintain stocks of both advanced logic chips (7nm and below) for modern weapons systems and mature node chips (28nm and above) for industrial and legacy military applications.

Implementation could follow successful models from other strategic reserves:

• Long-term contracts with domestic producers to ensure steady supply. We could even have the reserves held by the suppliers, rather than the military itself. 

• Regular rotation of inventory to prevent obsolescence. Chips depreciate. 

• Geographic distribution of storage facilities to reduce vulnerability. 

• Coordination with industry to align stockpile composition with critical needs.

• Regular assessment and updating of reserve requirements as technology evolves.

While specific quantities and exact composition would require detailed analysis by semiconductor experts and defense planners, the strategic value and the value as a complement to the CHIPS Act is clear. 

The CHIPS Act Needs Streamlining

The CHIPS Act's implementation has become unnecessarily complex, undermining its strategic purpose. Effective industrial policy requires rapid, directed action — getting capital to firms quickly to achieve specific strategic goals. To an extent, you just need to shovel money at people — but while the Act's subsidies try to drive domestic semiconductor production, excessive regulation and government risk-aversion have created bottlenecks in disbursement and reduced the subsidies' effectiveness. The regulatory burden functions as an implicit tax, offsetting the very cost advantages the subsidies aim to create. Here are just a few examples, but the list could go on. 

The first major obstacle comes from excessive financial and technical disclosure requirements. These force firms to reveal sensitive operational information that many are unwilling to share. SK Hynix has explicitly refused to apply for CHIPS funding due to these disclosure requirements, and Taiwan's national security laws may prevent TSMC from fully complying. The requirements create particular concern around intellectual property protection, with firms worried about revealing proprietary manufacturing processes and future product development plans.

Environmental compliance represents another significant barrier to rapid implementation. Environmental questionnaires and reviews typically delay groundbreaking by 6-12 months, while NEPA compliance alone adds approximately 5-10% to total project costs. The problem is compounded by multiple layers of environmental review, with state-level environmental regulations often duplicating federal requirements. 

The Davis Bacon Act wage requirements create a third layer of regulatory burden. These requirements oblige firms with  federal contracts over $2,000 to track and document compliance across their entire contractor network, adding significant paperwork and reporting costs. The Department of Labor's reporting requirements further complicate implementation, requiring dedicated compliance staff and creating additional overhead.

Perhaps most problematic are the NOFO (Notice of Funding Opportunity) provisions. The Act's "upside sharing" requirement forces companies receiving over $150 million to share an undefined portion of profits that exceed their projections. This intentionally vague profit-sharing threshold creates significant uncertainty for firms planning major investments. The provision effectively penalizes successful execution and innovation, while creating a perverse incentive for companies to overstate their initial projections to avoid triggering the clause. While we might want to restrict dividends for CHIPS Act grantees, this mechanism transforms what should be straightforward subsidies into a complex profit-sharing arrangement that discourages the very success the Act aims to achieve. 

The regulations above, as well as many more, simply put, need to be cut out from the Act. They are counterproductive and injurious. Rather than facilitating rapid domestic semiconductor expansion, they create delays, add costs, and discourage participation from key industry players.

Big Nasty China Tariffs

The last suggestion I have is slightly tangential to the CHIPS Act, but is a necessary addition. China's growing presence in the U.S. semiconductor market poses a strategic rather than purely economic threat. As China aggressively expands its chip production capabilities, particularly in mature-node semiconductors, it threatens to create dangerous dependencies in America's hardware supply chain. 

The solution requires decisive action: implementing extraordinarily high tariffs — up to 300% — on Chinese semiconductor imports, up from around 50%, where they are today. This isn't about protecting domestic industry from fair competition; it's about preventing a strategic adversary from establishing control over critical technology components. China's strategy is clear: Flood the market with mature-node chips, establish market dominance, and create U.S. dependence on Chinese suppliers. They're particularly targeting the high-node, less advanced chip market — components that, while less sophisticated, remain essential for industrial and military applications.

While the CHIPS Act's incentives and the proposed strategic chip reserve could help offset any market disruptions from such aggressive tariffs, that consideration is secondary. The primary goal must be to create a complete break from Chinese semiconductor dependence before it becomes entrenched. 

Conclusion

The CHIPS Act represents a crucial first step in rebuilding America's semiconductor fabrication capacity, but its current implementation threatens to undermine its own goals. The solutions proposed here — strategic tariffs, regulatory streamlining, and additional support measures — would transform the Act from a well-intentioned but bureaucratically hampered program into an effective engine of industrial revival.

Perhaps most critically, these reforms would work together synergistically. Tariffs protect investments enabled by streamlined subsidies, while guaranteed demand from the strategic reserve provides long-term stability. Meanwhile, aggressive tariffs on Chinese semiconductors would prevent strategic vulnerabilities from developing while these domestic capabilities are built.

The global race for semiconductor independence is accelerating. Europe is investing heavily, and China is pouring hundreds of billions into domestic production. America cannot afford to let bureaucratic complexity and half-measures undermine this critical industry. With these targeted reforms, the CHIPS Act can fulfill its intended purpose: securing America's technological independence and industrial strength for decades to come.

3

Noahpinion

America's semiconductor policy is missing a key piece

I’ve written a fair amount about the Biden Administration’s policies aimed at maintaining American dominance semiconductor industry — the CHIPS Act and the export controls on China. But what I’ve written only scratches the surface; this is an incredibly technically complex industry with…

Read more

2 years ago · 199 likes · 37 comments · Steven Glinert

4

 

Noahpinion

Why targeted tariffs are more effective than broad tariffs

Trump is about to be back in the White House, so tariffs are back on the menu. Actually they never left — Biden slapped huge tariffs on a variety of Chinese products, including electric vehicles, chips, and other stuff. But Trump is contemplating tariffs that are…

Read more

a year ago · 248 likes · 39 comments · Noah Smith

noah has a good point about how currency fluctuations make general tariffs less effective than strategic tariffs on a certain set of industries. 

Subscribe now