IEA chief Fatih Birol on how to navigate the age of electricity
IEA chief Fatih Birol on the age of electricity, Europe’s strategic energy mistakes, the Hormuz shock – and why trust is now an energy asset....
by José Caballero Published October 9, 2026 in Geopolitics • 8 min read
On 24 August 2026, prosecutors in Taiwan charged nine people, including employees of Nvidia and Supermicro, with illegally exporting 74 advanced AI servers containing Nvidia B300 processors to China via Indonesia, Japan, and Hong Kong. Prosecutors allege that some of the accused set up a Japanese shell company and used false paperwork and fake websites to conceal the shipments. The case remains before the courts.
A few years ago, an alleged scheme built around three countries and a shell company to move computer hardware would have been unusual. In 2026, it has become significantly less so. The effort spent to move 74 servers across three borders shows how much an export license is now worth and how much companies and individuals are willing to spend to avoid detection.
This case is one clear example of a change that has been building since the United States first restricted Nvidia’s exports of advanced chips to China in 2022. These changes are a key part of the modern profile of 30 economies surveyed in the annual Hinrich–IMD Sustainable Trade Index The case illustrates how geopolitical tensions, commercial incentives, and supply-chain governance can complicate the foundations for long-term sustainable trade. The 2026 edition of the Index will be published on 13 October.
Washington tightened those restrictions in each of the years since, loosened them in one category of chips in 2025, then reimposed limits on the newest generation. Each change in the rules altered which processors could be sold, to which buyers, and under what review. The overall pattern has not changed. Access to the chips that train and run advanced AI systems now depends less on price and more on where a buyer’s government stands in this rivalry and the buyer’s own standing with the companies and regulators that control the supply. The case in Taiwan shows what happens when the value of avoiding that trust requirement is high enough to risk criminal charges.
Export controls on technology are not new. During the Cold War-era, Western Governments used export restrictions to limit the Soviet bloc’s access to military and dual-use technology. For decades, controls focused on a relatively narrow set of products with an obvious military use.
That changed with computing hardware. In 2019, the United States placed Huawei on its Entity List, restricting its access to American chips and chip design software. The shift accelerated in 2022 when Washington expanded controls beyond finished chips to include the equipment and software needed to design and manufacture them. For the first time, restrictions targeted an entire category of civilian technology and the ecosystem required to produce it.
The rules tightened further in 2023. Nvidia responded each time by designing China-specific chips, including the A800, the H800, and the H20, designed to comply with the latest restrictions. Each was eventually caught by new controls, culminating in a 2025 licensing requirement for sales of the H20 to China.
The United States also asked its allies to adopt similar controls. The Netherlands restricted ASML, the only maker of the most advanced chipmaking equipment, from servicing Chinese customers, and Japan placed similar limits on its own equipment makers. What began as one country’s national security policy became a coordinated, multinational system capable of reshaping access to critical technologies.
The economic logic holds that goods flow to wherever they are valued most, and that price is the mechanism that sends them there. A government can distort this with a tariff or a quota, but the underlying assumption of most trade policy for the last several decades was that the distortion was the exception and the price mechanism was the rule. The system now governing advanced chips works differently. A buyer with unlimited funds cannot simply pay more to obtain a restricted processor. The chip is not for sale to that buyer at any price, because the constraint is not about cost. It is about which side of a boundary the buyer is on, and that boundary is drawn by geopolitics rather than markets.
in January 2025, the US attempted to formalize this logic into an explicit structure. It divided nearly every country in the world into tiers for the purpose of buying advanced AI chips. A small group of close allies faced few restrictions. A second, much larger group of countries faced caps on how much computing power they could import without individual government approval. A third group, including China, faced near-total restrictions. Governments that had spent decades negotiating tariff schedules and market access commitments were now sorted into a small number of trust categories, and a country’s placement in these categories mattered more to its technology companies than any tariff rate did.
This tiered system was rescinded in May 2025, but the exact rules in place at any given time now depend on decisions made in Washington that can change with little notice. What has proven durable is the underlying method. Instead of setting a general rule and letting companies decide who to sell to within it, governments increasingly decide in advance which specific buyers and which specific countries are allowed to participate at all. The United Arab Emirates government negotiated directly with the United States government for the right of its companies to buy advanced chips, and the terms of that access were tied to specific commitments about where the chips could be used and who could not gain access to them through the UAE. Saudi Arabia negotiated a similar arrangement. These are not trade agreements in the traditional sense. They are closer to admission decisions, made case by case, based on how much a government trusts another government to enforce the terms it agrees to.
This is a significant change in how access to a valuable technology is allocated. A company cannot win a sale by lowering its price if its buyer’s government has not been granted enough trust to purchase at all, and a country cannot improve its access through better terms of trade. It can only improve its access by convincing another government that it will enforce the same restrictions that government enforces on itself.
For a company that makes or depends on advanced chips, the first practical step is to treat export control status as a property of the customer and the country, not just of the product.
For a company that makes or depends on advanced chips, the first practical step is to treat export control status as a property of the customer and the country, not just of the product. A chip that is legal to sell today can become illegal to sell tomorrow if the buyer’s country moves into stricter tiers established by the United States, and a company that only checks the rules at the point of sale can eventually find itself holding inventory or contracts that were legal when signed and are not legal anymore. This is a harder problem for companies that depend on a small number of suppliers for critical equipment, since most of the alternatives are based in the same countries and covered by the same rules, and the only real defense is planning that assumes the rules will change without warning.
For smaller firms further down the supply chain, individual employees cannot reasonably investigate every customer’s ultimate destination and true ownership, and the Taiwan case shows what happens when that responsibility falls on them alone rather than on a proper compliance system. Industry associations and larger firms in the same supply chain have an interest in building shared verification systems and shared databases of restricted parties, since a failure by one firm affects the reputation of every firm in the same chain, and this kind of infrastructure is far cheaper to build together than alone.
For countries, a government that is not a close ally of either major supplier of this technology has an incentive to build the same kind of institutional record that recently earned other governments expanded access, including a credible domestic system for verifying end use and preventing re-export to restricted destinations.
Governments increasingly decide in advance who is allowed to participate in markets, based on institutional comparability rather than price and quality alone.
The individuals charged in Taiwan were not unusually reckless, and the scheme they allegedly built was not especially sophisticated. On the face of the charge, it required a shell company, some falsified paperwork, and a willingness to route servers through three countries instead of one. What drove this alleged scheme in the first place was not the value of the chips themselves since the same processors sell for a fraction of the cost in a market where they are legal. It was the value of access, in a system where access is granted or withheld based on trust between governments rather than on price. That value is concrete; it just isn’t paid to Nvidia anymore. Instead, it is captured as a premium on access, reflecting the compliance risk of moving restricted chips. That premium is precisely what makes the effort to circumvent the restrictions worthwhile, and it is what Taiwan’s prosecutors are now trying to expose.
This is not a story only about chips. It is the clearest example of an ongoing pattern across trade. Governments increasingly decide in advance who is allowed to participate in markets, based on institutional comparability rather than price and quality alone. The chip case makes this visible because the stakes are high and the technology is concentrated in a few hands. The same forces are at work across almost every other aspect of global trade. The Sustainable Trade Index exists to help policymakers and business leaders see how geopolitical alignment, commercial incentives, and governance gaps collide before they ever reach a courtroom.
Senior Economist at the IMD World Competitiveness Center
José Caballero leads the IMD World Competitiveness Center’s research team in the development and implementation of new models of assessing competitiveness. His research interests focus on the sources of the competitiveness of countries and, more specifically, on the competitiveness of enterprises. He is also an expert on the political economy of Latin America.
August 20, 2026 • by David Bach in Geopolitics
IEA chief Fatih Birol on the age of electricity, Europe’s strategic energy mistakes, the Hormuz shock – and why trust is now an energy asset....
July 24, 2026 in Geopolitics
How can leaders stay competitive amid geopolitical uncertainty? David Bach explores resilience, AI, energy security, trust and European reform...
July 15, 2026 • by Lindsay Sarah Krasnoff in Geopolitics
Once the preserve of men in pinstripes, international sports stars have redefined the art of diplomacy. Lindsay Sarah Krasnoff explains the new rules of the game and how business leaders can benefit...
July 13, 2026 • by Jerry Davis in Geopolitics
Only Elon Musk can fire Elon Musk. This does not bode well for investors, pensioners, or democracy, warns Jerry Davis....
Explore first person business intelligence from top minds curated for a global executive audience