Much of human life is animated by totemic abstractions: markets, laws, social taboos, and so on. Those who control them and who know their inner workings have a privileged form of power. The Internet is one of the most pervasive of these abstractions. But it’s also a collection of wires connecting computers, and all of those computers exist in physical space under the control of particular governments.
The state, too, is an abstraction. It is a set of norms and agreements that govern behaviors—increasingly, behaviors which are tracked and influenced by the powerful companies built upon those wires. States and tech companies wield power over each other, as well as against their rivals. This leads to strange confrontations: most recently, India sparked a global wave of technological nationalism by banning TikTok, an app mainly renowned for letting teenagers share short dance videos. Soon after, the U.S. followed suit with its own prelude to a ban.
India and the U.S.’ actions against TikTok reveal deep truths about modern international relations. Countries have always competed economically, spied on one another, and engaged in tit-for-tat retaliation over real and perceived slights. But in a globalized world where life and politics are increasingly mediated by software, retaliatory protectionism doesn’t mean autonomy. It means choosing which superpower to be a vassal to. The alternatives to Chinese apps are generally not home-grown competitors, but American ones. Saying no to Beijing means saying yes to Silicon Valley. In turn, saying yes to Silicon Valley means saying yes to Washington.
China, the U.S., and India each experience different dynamics between state power and the private tech sector. In China, the party-state holds the reins and strategic firms are private only in a nominal sense. On the other hand, the U.S. government seems to be increasingly in conflict with its tech giants. Google and Facebook need foreign policies, and theirs are not necessarily that of the American state and its departments, which are barely unified on that front anyway. But despite this, U.S. attacks on foreign tech giants de facto benefit its domestic players. India represents a third model: its private telecom giant took the initiative to forge deep links with the state. In each of these cases, a similar strategy emerges: markets as a testing ground, network effects and state protections—direct or indirect—as invaluable to scaling up.
Americans have the peculiar privilege of setting the world’s defaults. Mandarin is the world’s most common native language, with 920 million native speakers. But the most common second language—spoken in addition to the native mother tongue—is English, at 900 million speakers. This makes it more than three times more popular than the next most common second language, Hindi. It’s a linguistic equivalent to the power of the dollar, a value which is universally recognized. Countries in developing markets tend to measure costs internally in dollars, even if their external reporting is in reals or naira. Similarly, they’ll use English at conferences and in advertising, while speaking something else at the water cooler.
As a consequence, what looks like globalization to Americans reads as Americanization to everyone else. The metric system is perhaps the only significant aspect of globalization that doesn’t treat American preferences as a universal default.
In the case of software, most of the world does searches on Google’s search engine, socializes on Facebook, Instagram, and WhatsApp, and uses Microsoft’s operating systems on desktops and Google’s Android OS on mobile. E-commerce is a bit more diffuse—it’s easier to add server racks than fulfillment centers, but even so, American companies dominate. The number one e-commerce company in Japan is Amazon Japan; the number two, Rakuten, uses English as its primary language internally. In the UK, Germany, France, Italy, and Spain, Amazon similarly ranks at the top. And, over time, this dominance has gotten even more extreme. Ten years ago, five search engines had gained the lead in at least one country: Yandex in Russia, Seznam in the Czech Republic, Naver in Korea, Baidu in China, and Google almost everywhere else. Now, Google, Baidu, and Yandex are left.
In India, the immediate result of the TikTok ban was a spike in popularity for local competing apps, one of which was adding 100,000 new users an hour. But these apps can’t muster the same funding TikTok had—some of them have raised small rounds, but TikTok’s growth was fueled by endless advertising on competing social networks. And now, those competing social networks are much more direct competitors: days after the TikTok ban, Facebook expanded its competing service to India.
A month after India’s ban, the U.S. followed suit, giving TikTok 45 days to sell to an American company or be shut down. The U.S. calculus is a bit different from India’s: there are multiple American companies that can operate and monetize a business at TikTok’s scale, and technology companies are a core part of U.S. global economic competitiveness. India has less to lose from a dramatic shutdown, but in the U.S. case it’s more important for outside investors to know that they’ll get some of their money back. A further complication is that many of ByteDance’s outside investors are American venture capitalists and hedge funds.
All of this seems somewhat trivial. TikTok videos are not exactly the axis around which world history revolves. But TikTok is not to be underestimated: one security researcher who reverse-engineered the app called it “a data collection service that is thinly-veiled as a social network,” adding that “[i]f there is an API to get information on you, your contacts, or your device…well, they’re using it.” There are two good explanations for TikTok’s incredibly aggressive data-collection tactics, which are not mutually exclusive. One is that the app shows users content based on machine learning algorithms—though new users get a lot more of an instant boost than on other platforms, providing a powerful hook—and every data point provides a slightly better signal for what to show users. The other is that the app is a tool of the Chinese state, designed to spy on as many people as possible. TikTok assures the American public that they’re not controlled by their Chinese parent company, ByteDance; in reality, China’s national intelligence law gives the state broad powers to demand data from Chinese companies, even on non-Chinese citizens.
To the extent that other countries care about their sovereignty, the only difference between these outcomes is which problem is more immediate. From India’s perspective, a spyware network controlled by a country that is fighting a low-level border dispute, with fatalities on both sides, is problematic enough. But an app that’s designed to control attention as aggressively as possible, and to monetize that attention through advertising, is a deeper longer-term threat. TikTok’s parent company, ByteDance, has been credibly accused of censoring anti-China content in Indonesia, and India’s politicians may worry that the same thing has happened in India.
A foreign company that’s not an arm of the state will still make allowances for the preferred narratives of states whose patronage they rely on. American technology companies are similarly sensitive to the regulatory and PR risk of allowing some kinds of content on their platforms. China’s government recognizes the value of media soft power, and gently nudges overseas media companies away from damaging narratives. So far, the results impact American life on the margin. For example, a video game about China invading the U.S. turned into a game about North Korea invading the U.S. The time travel movie Looper was partly funded by a Chinese investor, and includes several scenes in which time travelers from the future mention how prosperous China has gotten. This is hardly wholesale censorship. It’s just a gradual, barely-noticeable shift in the Overton Window, and it moves in only one direction.
For countries that don’t house global social media powerhouses, an app with a global content feed makes the media environment less focused on local issues and more focused on internationally attention-grabbing viral outrage. India, despite its enormous population, is decidedly lacking in digital influence on this front. This means that its near term decisions consist mainly of deciding which foreign organization will determine what news its citizens read on social media. TikTok’s content moderation has led to controversy in the past, with videos about Hong Kong and the Uyghurs getting removed from the app’s main feed.
Somewhat paradoxically, India has achieved more digital self-determination in a more challenging field: network equipment. China is the world’s largest exporter, but has only two globally-recognized brand names: TikTok and Huawei. Huawei is the dominant provider of 5G equipment, and there is ample evidence that Huawei is controlled by the Chinese state. Huawei is the world’s largest provider of 5G equipment. There is also evidence that their equipment has security flaws, though the extent to which these were deliberate weaknesses is unknown. As with TikTok, there’s a plausible, comforting explanation: the company is just another heavily-subsidized national champion, and the Chinese state is willing to earn a poor return on its investment in exchange for the prestige of being the top seller of a technologically-impressive product. The darker version of the Huawei story—that it’s installing espionage equipment in the heart of global telecom systems—is both hard to prove and hard to refute. There are precedents for this behavior. A Swiss encryption hardware company turned out to be a CIA front, after decades of selling strategically broken hardware to the U.S.’s international foes.
In the case of 5G, India has a national champion of its own: the vast oil refining, petrochemicals, retail, and now telecom conglomerate Reliance. Reliance is an enormous company, accounting for 2.6% of India’s GDP, and it has long exercised a high degree of control over India’s economy. In 2016, Reliance launched a new mobile phone service, Jio, which now has nearly 400m subscribers, and which has raised $20bn from outside investors. In 2018, Jio acquired a U.S.-based telecom company that made 5G hardware, and earlier this year, Jio promised a “made-in-India” 5G solution, ready for use in 2021 and ready for export thereafter.
India’s Prime Minister, Narendra Modi, has called for “Atmanirbhar Bharat”: self-reliant India. And when Jio announced their 5G plans, Reliance’s CEO proclaimed that he would “dedicate Jio 5G solution to PM Modi’s Atmanirbhar Bharat.” This sounds like over-the-top corporate puffery, and from another country it would be. But Reliance has a storied history of close ties to the Indian government. Like Samsung in Korea, it’s an open question as to whether they’re almost a department of the government or if the government itself is a subsidiary of the firm. Companies like these are essential instruments of national sovereignty. In some fields, like heavy industry, government-owned companies and public-private partnerships have gotten acceptable results, but this is less common in bleeding-edge technology. After decades of effort, China still hasn’t reached parity with Taiwan, the U.S., or Korea in semiconductor fabrication. National champions in this kind of business benefit from the government’s power to veto competitors, but they still tend to be private sector creations. India’s government would have a hard time creating a 5G equipment company by fiat, but would have much more luck protecting a local one that already existed.
The U.S., with its vast technology industry, has both more options and more constraints. American companies can clone TikTok—Facebook is already on its second attempt, and Twitter previously operated Vine, a short video app. American companies can also build 5G hardware. But the U.S. faces more risks because its successful tech companies are global: a focus on dominating the U.S. market risks losing share in the 60% of the global economy that’s neither the U.S. nor China. Moreover, the U.S. economy is also dependent on continuous inflows of foreign capital: eroding property rights to strike back at one tech company will make investors think twice when investing in American corporate bonds, mortgage-backed securities, and other financial products. India benefits from foreign capital flows, but doesn’t rely on them, while they’re an important component of the U.S. economy. This means that American technological nationalism is constrained in the same way that American economic nationalism is: the breadth of U.S. power makes abusing that power prohibitively expensive.
Technology companies all benefit from capital availability and economies of scale, and rising technological nationalism means that fewer and fewer countries will have the scale necessary to support local champions. At the same time, this nationalist trend increases the importance of home-grown solutions. Having a choice between the whims of the Chinese Politburo or those of The New York Times editorial board does not exactly sound like freedom to most heads of state. India is not alone in worrying about this—Japan has subsidized its own 5G champion, which is already making inroads into Western markets. It’s an important time: if relatively large countries with a strong domestic tech ecosystem, like Japan and India, can’t build their way to digital sovereignty, then other nations don’t have a chance. Such a scenario would mean that every country will be actively choosing to align itself with China, or defaulting to alignment with the U.S. Neither option results in an immediate loss of sovereignty, but both cause a gradual shift in the bounds of acceptable discourse—the game of politics still plays out domestically, but the rules of the game are set somewhere else.
The techno-optimist perspective is that code is intrinsically neutral, and that the most technologically advanced solution will be the ultimate winner. The radically pessimistic view is that the fixed cost of each new technological paradigm is so high that only global superpowers can afford it—like nuclear weapons, a few places will produce their own and everyone else will have to operate under a major power’s umbrella. But part of statecraft is the conviction that a country can determine its destiny. China achieved an extreme form of digital sovereignty when it was still a very poor country: the Great Firewall dates back to 1997, when China’s GDP per capita was under $800. States with a less authoritarian bent and smaller ambitions can still get some measure of self-determination, albeit at a cost.
When it comes to predicting who will achieve digital sovereignty, wealth is the wrong thing to bet on. The ability to create private tech actors in strategic industries, and to protect them, matters more. Having a large population and a linguistic barrier to entry helps too, although it hasn’t been the killer app for India. Finally, the country has to be able to project power into a global market of platforms and influence. North Korea probably enjoys a type of digital sovereignty, but not one that will make it a relevant actor in the technological destiny of the human race.
What technological optimists miss is that systems are colored by adversarial behavior. At the local level of sorting algorithms or graphical computations, it’s true that the best technology wins and it tends to be neutral. But telecommunications equipment is immensely complicated, and needs to interface with other complex products. At sufficient levels of complexity, technology is a system beyond just the hardware and software itself. It includes the people who build and maintain it, and the laws under which it operates.
A country that can’t build its own digital infrastructure is unlikely to be in a position to evaluate what’s on offer from other countries. The optimists may have a point: perhaps Huawei is popular because it’s cheap, perhaps the Chinese state subsidized it because it’s a sound business, and perhaps the massive Nortel hack was unrelated. But if one element of statecraft is ambition, another is precaution. After all, countries don’t last forever. A single strategic mistake in who decides one’s technological future could bind you to another power’s whims for generations.