Terrified that China might use cutting-edge GPUs for military or surveillance purposes, the US government imposed sweeping export bans on high-end AI chips and fabrication tools. While the intent was to keep China technologically one step behind, the result was anything but, with China rapidly emerging as a chip-design powerhouse. (Image: ChatGPT)
By any rational measure, the history of technological protectionism reads like a tragicomic epic of self-sabotage – a recurring spectacle where governments, in the name of “national security” or “trade fairness” accidentally hand their competitors an innovation stimulus package.
The logic is always the same: If we don’t let them have it, they can’t build it. The outcome is always identical: They build it anyway – and better, faster, cheaper.
On 10 October, Trump posted on Truth Social: “[…] United States of America will impose tariffs of 100% on all goods and any tariffs that China is currently paying. In addition, the US will impose export controls on critical software starting 1 November 2025, to protect American technological leadership and national security.”
Most commentators reacted to the first sentence. But it is the second which lands with a dull, ugly thud.
Let’s revisit three shining examples from recent history: the US ban on strong encryption, the satellite export restrictions of the early 2000s and the current AI-chip embargo on China. Each was born from paranoia or trade politics, dressed in patriotism, and ended up with unintended consequences.
In the early 1990s, US policymakers discovered a new enemy: mathematics.
Specifically, strong encryption. The same algorithms that made online banking, e-commerce, and private communication possible were suddenly reclassified as munitions. Astoundingly, exporting robust encryption software was legally equivalent to selling a Stinger missile. A graduate student sharing an open-source encryption paper abroad could, in theory, be charged with arms trafficking.
The logic was simple enough: if foreigners can’t buy American encryption, they’ll be forced to use weaker stuff, and the NSA can keep listening in. Unfortunately, foreigners took this as a challenge, not a restriction. European and Israeli firms built their own strong encryption. Open-source communities gleefully published cryptographic libraries that anyone could download.
The US, meanwhile, handcuffed its own software companies. Netscape, Lotus and Microsoft had to release crippled “international versions” of their products – like selling safes that come with the combination printed on the door.
Eventually, the absurdity collapsed under its own weight. Courts ruled that source code was protected speech, export controls were relaxed, and the global encryption genie was long out of the bottle. But by then, America had taught the world a valuable lesson: if you ban maths, the rest of the planet will start studying calculus.
Fast-forward a decade. The US again sought to “protect” its technological lead – this time in space. After a Chinese rocket carrying a US-built satellite (Intelsat 708) exploded in 1996, Congress panicked and reclassified commercial satellites under the International Traffic in Arms Regulations (Itar).
Suddenly, communications satellites were officially “weapons”. Any component with a US microchip was subject to labyrinthine export restrictions. American companies were forbidden from selling satellite parts to China – and often to anyone who might, heaven forbid, launch with China.
Law of unintended consequences
The result? US firms were grounded whileEuropean competitors popped champagne. French and Italian manufacturers proudly marketed “Itar-free” satellites, assembled without a single American screw. International buyers rushed to them to avoid US red tape.
A 2014 report to Congress estimated that American satellite manufacturers lost billions in contracts and global market share – all in the name of “national security”. China, meanwhile, accelerated its domestic satellite and rocket programmes, leading directly to the emergence of today’s Long March heavy-lift fleet and BeiDou satellite navigation system – both now Chinese strategic assets.
In short: the US locked the door to keep China out, but forgot that China knew how to build another house.
Which brings us to the present day – and to Nvidia, the crown jewel of America’s AI boom.
Terrified that China might use cutting-edge GPUs (graphics processing units) for military or surveillance purposes, the US government imposed sweeping export bans on high-end chips and fabrication tools. The intent was clear: keep China technologically one step behind, ensure “strategic dominance”, and satisfy the domestic political craving for “toughness”.
Read more: Nvidia ushers in the ‘AI factory’ age with a whole new language of computing
Read more: China’s AI ecosystem has ‘outbrained’ that of the US
The results should have been predictable.
Chinese firms – previously happy to buy Nvidia’s world-leading chips – are now pouring billions into designing their own. Huawei, SMIC and a host of start-ups have sprinted into the semiconductor arms race. A ban meant to cripple China’s AI progress has become an R&D rocket booster.
Even Nvidia’s CEO, Jensen Huang, called the policy “a failure”.
He noted that those controls had shrunk Nvidia’s market share in China from around 95% to about 50%, and said the curbs gave Chinese firms “the spirit, the energy and the government support to accelerate their development”. These comments were made in March. Today its share of the high-end chip market in China is exactly zero.
The company lost access to one of its largest markets; the Chinese state gained a new existential incentive to innovate. More humiliating, China has now actually banned the import of advanced Nvidia chips. The same pattern unfolds: protectionism as a motivational speech for your rival.
At its heart, technological protectionism is an economic misunderstanding disguised as patriotism. It assumes innovation is a finite commodity – that if one nation restricts another, it can preserve its lead indefinitely. But innovation behaves like information: it leaks, spreads, recombines and resurfaces in unexpected places.
When you block access to a technology, you don’t eliminate the demand – you simply divert it into new channels. China’s rapid emergence as a chip-design powerhouse mirrors the European encryption wave of the 1990s and the satellite boom of the early 2000s. Restriction breeds reinvention.
It might hobble China for a short while, but in fact it will be the US that has shot itself in the foot. DM
Steven Boykey Sidley is a professor of practice at JBS, University of Johannesburg and a partner at Bridge Capital and a columnist-at-large at Daily Maverick. His new book "It's Mine: How the Crypto Industry is Redefining Ownership" is published by Maverick451 in SA and Legend Times Group in UK/EU, available now.