Why Grabbing Greenland could kill 6G and burst the AI bubble

You’ve probably noticed that there’s a bit of a penchant for land grabbing at the moment.  Russia wants Ukraine, Israel wants Palestine, China wants Taiwan, and now, America wants Greenland.  The problem is that as more of it goes on, the behaviour becomes normalised and the international barrier to grabbing is lessened.  America’s Christmas foray into Venezuela has further regularised the concept of interference, which  means there’s probably a mandarin in Beijing who’s already cancelled their summer holiday and written “Taiwanese unification” on their wall calendar.

The problem with grabbing Taiwan is that it’s not just land-grab.  It comes with some interesting unintended consequences, which may profoundly alter the technology balance between China and the US, bursting the AI bubble and upsetting telecoms evolution along the way.

The unintended consequence with China annexing Taiwan is TSMC.  Hardly anyone outside the high tech industry has heard of it, but it rather pragmatically stands for the Taiwan Semiconductor Manufacturing Company.  Over the last forty years, they have been quietly dedicating themselves to making more complex silicon chips than anybody else in the world.  They make them for companies like Apple, Nvidia, Broadcom, AMD and Qualcomm, running most smartphones and tablets.  They also produce the chips powering the current AI revolution, as well as the ones which Amazon, Google and Meta design for their data centres.  If a product or application is bleeding edge, the chips in it are almost certainly made by TSMC.  And they’re only made in Taiwan.

The rest of the world started to notice this during Covid, when supply chains in the semiconductor industry caused problems.  As a result, the US, Europe and China poured hundreds of billions of dollars into building new wafer fabs (short for “fabrication” – the factories that make chips).  They soon discovered that whilst you might be able to buy the equipment to make the chips, the expertise to run them requires immense engineering experience, which takes a lot longer to acquire. While TSMC was investing in a highly trained and efficient workforce, the rest of the world wasn’t, which is why Taiwan is the only place which can manufacture these chips in volume.   As a result, TSMC is probably around five years ahead of its nearest competitor in building the most advanced chips we use today.

Politics has also entered the arena.  Not only do other countries want to make these chips within their own borders, but the US is actively trying to stop China gaining any technical advantage.  These advanced chips shrink transistors down so that you can fit over 300 million of them in a square millimetre of silicon.  Only one company – ASML in Holland, makes the extreme ultra-violet lithography machines that are capable of producing structures on a silicon chip that are just a few nanometres wide.  These machines are incredibly complex, costing around $400 million each.  In an attempt to stop China developing a lead in semiconductor technology, the US has persuaded the Dutch government to ban the sale of these machines to China.  Should China try to grab Taiwan, it’s likely that they will try to move some of TSMCs units to semiconductor plants in mainland China. 

Which is where the problem comes in.  Within the industry there a general belief that if there is an invasion of Taiwan, these machines will be destroyed before China gets to the factory gate.  That may be apocryphal, but there are almost certainly military strategists in the US and Taiwan who are aiming a few missiles at the TSMC wafer fabs to make sure China doesn’t get their hands on them.  That  military strategy ignores the unintended consequence, which is that without TSMC, the world loses its bleeding edge chips.

From a consumer viewpoint, a lot of recent phones would disappear.  Apple would probably be unable to build anything more recent than an iPhone13.  From a wider industry perspective, 5G would barely be worth supporting, as the chips in any new phones would only support minimal 5G features.  Not many people would notice, but it would put a major spoke in the plans for 6G.  The telecoms industry’s business plan relies on a regular transition to the next standard, but it is already struggling selling the benefits of 6G, with analysts declaring 2026 as the year of 6G slop.  If the lack of new silicon adds a three or four year delay, the economics of infrastructure renewal may become untenable.  At least they will in the West.  The US policies of trying to restrict China’s access to technology over the past few years has proven the old adage that “necessity is the mother of invention”.  Huawei has looked at alternative ways of squeezing performance out of less dense chips, with results which appear to compete with TSMC’s more advanced ones.  If the West loses access to TSMC, it allows China to take a significant technology lead in telecoms, using its home-grown technology.

The same is true in AI.  China’s DeepSeek AI model stunned the world with its approach to large language models (LLMs) running on simpler hardware.  The rest of the world relies on Nvidia’s chips, which are built by TSMC.  Without them, the current AI business models are probably unsustainable, bursting the AI bubble.

If the AI bubble bursts, there’s a knock-on effect.  The projected growth in AI has led to billions of dollars being invested in fission and fusion start-ups to generate the additional energy required to run the data centres.  When the AI bubble bursts, that money, and most of the startups depending on it, will be gone.

Nobody can predict if, or when this will happen, but as world leaders continue to trumpet their territorial acquisition hopes and policies, the likelihood of it happening increases significantly.  If it appears imminent we will probably see a repeat of what happened at Peenemünde at the end of the second world war, when Britain, the US and Russia tried to extract all of the German rocket scientists.  In this case, the US, China, and probably South Korea will be grabbing anyone working at TSMC to bolster their semiconductor workforce.

Anything that affects TSMC’s position as the global leader in chip manufacturing will have far reaching ramifications on the next generation of technology, and probably propel China to a two to three year lead over the US.  Second place would probably go to Samsung in South Korea.  But that’s OK, as I’m sure nobody wants to annex that. 

It’s likely that the national leaders are unaware of these consequences – they’ll be keeping themselves busy colouring in their maps.  For technology and venture capital, it could be a very different matter.

(As is often the case, this is history repeating itself.  For a commentary on similar circumstances in the 1960s, I highly recommend Tom Lehrer’s “Send the Marines” and “Werner von Braun”.)