#22. The Splinternet
December 19, 2022
The 5G ban: Last week, at the AQR Asset Management Institute Summit hosted by London Business School, the infamous Nuriel Roubini (or, as many call him, “Dr. Doom”) stressed the importance of the Biden Administration’s recent decision to ban the approval of any new telecomm equipment from Chinese state-owned companies, notably Huawei and ZTE. The decision came after a March 2021 designation of five Chinese companies, including both Huawei and ZTE, as posing a threat to national security.
Different 5G infrastructure = different internets: Roubini’s point was there is now no way to have anything fewer than two internets on this planet. The reason for disallowing Chinese equipment to build our internet is fear of espionage by CCP. At a minimum, the two internets will be Chinese and non-Chinese. A prescient 2020 note from BAML called it the “splinternet” (I’m commandeering it now ™).
The Chinese splinternet has, in effect, existed for over a decade. The Great Firewall was completed in 2008. Already today, none of the external world’s internet or infrastructure flows through China without illegal VPNs. The Biden Administration’s ruling means that not only is western internet infrastructure cut off from China, but now so too is China’s internet infrastructure cut off from the US.
One of 5G’s most famous use cases is the operation of autonomous vehicles. If China and the US have completely different 5G internets, then neither Chinese autonomous vehicles nor their components will work in the US, and vice versa. Each host country either (1) won’t allow the equipment to be bought for the same reason the infrastructure has been banned (espionage) or (2) will allow it to be bought but will severely limit it into inadequacy.
Different internets = different manufacturing: If American components won’t work in China and Chinese components won’t work in America, each country will need to manufacture their own components for autonomous vehicles. But that’s just where it starts. Autonomous vehicles are the poster child for 5G and the internet of things (IoT), but they’re neither the only use case nor the first.
IoT exists in nearly every home in America today. In our connected coffee machines, smart TVs, our smart speakers, my laundry machine is even WiFi-enabled. Tons of cheap things you buy on Amazon have wifi-enabled microchips in them. Not because they’re spying on you, but because life is just easier when they’re connected to the internet. Soon (very soon), anything that can have a microchip in it will have a microchip in it. “But surely some things are too small” says somebody. No, wrong. Microchips are tiny. The smallest are literally microscopic. And simple radio chips are inexpensive.
If the US remains rightfully dogmatic of protecting its internet infrastructure, it doesn’t end with 5G towers and tools. It extends to everyday smart devices that pose an equal, or possibly greater, risk due to their ubiquity. And because of the ultimate convenience of an internet-connected world, that inevitably extends to everything. So we will not just have to manufacture our own autonomous cars and components. We will have to manufacture our own everything. Different internets mean different everything.
The biggest point of friction for western countries won’t come from who to appoint to build 5G infrastructure and future technologies, but how to move a manufacturing base away from the splinternet (i.e., China) back into our internet (i.e., US allies). Building 5G and internet infrastructure across the world is akin to building the digital roads and walls but, in an internet-enabled world, physical goods will need to follow these roads and walls as well.
Regular microchip plug: Conversation around reshoring microchip manufacturing is timely and very relevant here. TSMC, the biggest microchip manufacturer in the world, upping it’s investment in a new Arizona manufacturing plant to $40 billion is fantastic.
Today, Taiwan manufactures 60% of all microchips and 90% of advanced microchips (think smartphones, laptops, etc.). While China, who eyes Taiwan hungrily, manufactures 5%, the US produces 10%. Japan creates 17%. And India is also making a big push (see my note on India’s push for chip supremacy here).
I think about this tweet a lot:
Different internet, infrastructure, & manufacturing = different world orders: Last step into the rabbit hole. The question isn’t how the world will produce goods after exiling China because China has no interest in being exiled. Publicly, China’s Belt and Road initiative was created develop transcontinental connectivity and cooperation between China and the external world. Less publicly, China is actively taking control of key partnerships throughout the world, particularly in Africa, and taking control of real assets. In Africa, China provided Djibouti a high interest loan to rebuild a crucial port that connects Asia to the US. A few weeks ago, Djibouti suspended all debt payments as it inevitably failed to keep up with interest payments, which now positions China to effectively take control. This has also happened to a power station in Zambia and a preeminent port in Sri Lanka. And it’s not just in developing countries. In Europe, against the recommendation of nearly every German ministry, Chancellor Olaf Sholz approved the sale of 24.9% of the Port of Hamburg, the EU’s third largest port, to China.
The US has said this stops at our shores, specifically with regards to 5G infrastructure. European countries like the UK, France, Sweden, Belgium, and Denmark are standing up as well in the Huawei ban.
So this isn’t solely about 5G, autonomous cars, or microchips. And it’s not about exiling China. The US will have to shore up our allies and woo others to build the future with us instead of with China. The Biden Administration has seemingly begun this process with the emphasis it has put on the US-Africa Leaders Summit in DC last week where Biden said the US is “all in” on Africa’s future.
Source: Bank of America Merrill Lynch 2020 note
If China can no longer be the world’s factory, the US and its allies will need to quickly build a new manufacturing base, which will require significant buy in from and incentives for our biggest companies (all high-end iPhone Pros are made in one Foxconn factory in China). I don’t expect the CHIPS Act to be the last of its nature in the public sector and I don’t expect recent private sector interest in American Dynamism and defense/manufacturing tech to slow down anytime either (see my note on defense tech here). There’s a new world to build.
