Philippa Kaur

BYJU’S WAS piling on users even before covid-19 closed classrooms around the world. India’s most valuable private startup was co-founded in 2011 by Byju Raveendran, a celebrity maths tutor whose classes have drawn crowds big enough to fill stadiums. By 2019 tens of millions of Indian children had signed up to use the firm’s flagship product, an app that serves up online lessons intended to supplement regular schooling. That year Byju’s began sponsoring India’s national cricket team.Since then India’s schools have spent more time shut than open—and the fortunes of Byju’s have only improved. The number of children whose parents pay for them to have full use of its app has more than doubled, to 7m. Late last year investors valued the firm at over $20bn, a three-fold increase since pre-covid days. In January Bloomberg reported that Byju’s may soon unveil plans to go public in New York, by merging with a blank-cheque company. The news agency had previously rumoured that such a deal could raise around $4bn, valuing the firm at a cool $48bn.Byju’s is the biggest of a clutch of young companies benefiting from breakneck growth in online learning. Venture capitalists (VCs) plonked around $21bn into education technology companies in 2021, according to Holon IQ, a research firm (see chart). That was three times the amount raised in 2019 and 40 times more than a decade ago. Seventeen ed-tech startups became “unicorns” (private companies valued at more than $1bn), three times as many as had passed that milestone during any previous year. Half a dozen of them went public. They included Coursera, a marketplace for online courses with a stock market value of nearly $3bn, and Duolingo, an app for language learners which is worth around $4bn. Holon IQ has predicted that global ed-tech revenues could almost double from $227bn that year to around $400bn in 2025, a fifth higher than its pre-pandemic forecast.Until recently ed-tech firms had rarely made investors sit up. Schools and universities control much of the $6trn spent globally on education each year. They tend to be cash-strapped and conservative. In 2019 only about 3% of all education spending went on software or online teaching. Tory Patterson of Owl Ventures, who began investing in ed-tech firms in 2009, admits that speaking up for the sector has sometimes won him “blank stares”.No more. The closure of school buildings and college campuses forced educators to try out new kit (especially in India and America, where disruptions to learning have been particularly drawn out). Governments have given children stacks of tablet computers and sped up efforts to improve broadband in schools. They have also given teachers extra cash to spend on tools they think will help pupils “catch up”. Lawmakers in America have earmarked an extra $200bn or so for schools since the pandemic started. That sum is equal to about one-quarter of what is spent on these institutions in a typical year.For years many of the zippiest ed-tech firms have chosen not to sell to schools and universities but to go direct to learners. This category of companies has also benefited during the pandemic. Parents in Asia have long been keen to pay for tutoring and other services (such as Byju’s app) that might give their offspring an edge. Now families in Europe and America are also getting keen. Supervising remote learning has made parents everywhere more engaged in their children’s education, more aware of how they are performing in comparison to classmates and in some cases more critical of what they are being taught. Companies that offer after-school lessons—such as Outschool, an American unicorn, and GoStudent, an Austrian one—are growing fast as a result.Another type of outfit getting a boost from the pandemic are those that offer learning to adults. Workers furloughed during lockdowns commonly took online courses that they thought would improve their prospects. Remote working has made more roles plausible to more jobseekers, giving them more reason to reskill. At the same time, a flurry of job-switching in Britain and America has made big employers nervous. They are becoming more convinced that spending on staff training can help them hang on to workers and cut the cost of plugging holes. This is benefiting companies such as Coursera, which says selling subscriptions to corporate customers is its fastest-growing business. Up-and-coming firms include Guild, which helps blue-collar workers at giants such as Walmart and Disney gain new qualifications, and Better Up, an American company that helps professionals find coaching.Ed-tech’s pandemic report card is not without blemishes, however. In China, its single biggest market, the Communist Party declared last July that businesses could not typically make a profit from providing after-school tutoring to children in primary and middle schools. The regime has worried for years that huge demand for private education is widening inequalities and impoverishing the middle class. Even charitable tutoring could no longer take place during holidays and at weekends. Within days the share prices of New Oriental, TAL Education and Gaotu, the industry’s three listed Chinese giants, had fallen by two-thirds, wiping out $18bn in stockmarket value. Since February 2021 their collective worth has shrivelled from more than $100bn to less than $10bn. China’s most celebrated ed-tech unicorns, Yuanfudao and Zuoyebang, could be worth a fraction of their pre-crackdown valuations of $15.5bn and $10bn, respectively.The Chinese experience has rattled investors, says Thomas Singlehurst of Citigroup, a bank. It blocked a possible exit route for Western startups, some of whose VC backers may have hoped to sell them to China’s ed-tech titans. It may also inspire tighter rules in next-door India, another potentially vast market where some parents accuse ed-tech firms of misleading ads and aggressive sales tactics. Last month India’s education minister said the government was considering new regulation, though he gave no details. Since then at least 15 Indian ed-tech companies, including Byju’s, have created a group promising to scribble new codes of conduct.Western ed-tech firms are unlikely to face similar strictures. But they have their own challenges. In November Chegg, an American company that gives online help to undergraduates, warned that lower-than-usual enrolment in American universities was affecting its revenue. Its market capitalisation, which soared to around $14bn in early 2021, is back down to $4bn, lower than it was before the pandemic. Shares in ed-tech companies that listed in America last year are mostly trading below offer price. Several, including Coursera and Duolingo, have yet to turn a profit.Not straight As, then. But the industry’s boosters think it has room to improve. An influx of users and money in the pandemic has given more firms the muscle to expand abroad and to find ways of retaining users for longer, reckons Deborah Quazzo of GSV, a big educational investor. Take Byju’s. It has spent at least $2.8bn on a dozen acquisitions in an apparent attempt to string together services that will allow it to reach learners of all ages, from toddlers to career-changers. The deals are also helping it reach customers far beyond India. In 2021 it began offering online classes in coding and maths to children in America, Brazil, Britain, Indonesia and elsewhere. A big listing might teach ed-tech sceptics and Western rivals alike a lesson. ■This article appeared in the Business section of the print edition under the headline “Learnings growth” More

NOTPETYA IS A nasty name for the world’s vilest computer attack. Embedded in an innocuous piece of tax software, the virus, which the American government said had the Kremlin’s fingerprints all over it, struck Ukraine in June 2017, knocking out federal agencies, transport systems, cash machines—even the radiation monitors at Chernobyl, the husk of a nuclear-power station.It then went rogue, worming its way from the computers of multinational firms with local outposts in Ukraine to their global operations, causing collateral damage to victims ranging from Maersk, a huge shipping company, and Saint-Gobain, a French construction giant, to Mondelez International, owner of Cadbury chocolate. The total hit was put at $10bn, making it the costliest such attack ever. One of the most expensive blows fell on Merck, a New Jersey-based drugmaker with a market value close to $200bn, which lost 40,000 computers in the blink of an eye and was forced to halt manufacturing of its human-papillomavirus vaccine.Merck sought to cover its cyber-losses with a $1.4bn property-insurance claim. However, its insurers refused to pay, invoking a clause in the contract called war exclusion. This precludes coverage in the event of warlike action by governments or their agents. The matter ended up in a New Jersey court. Years later, as Russian troops and cyber-warriors are once again threatening Ukraine, a judgment in the case offers a timely reason to explore how much companies have learned since then about dealing with potentially catastrophic cyber-warfare. The short answer is: not enough.The Merck judgment, made public last month, is potentially a landmark one. It tackles a question of great importance in the context of modern-day belligerence: is cyber-warfare war? Merck’s insurers, including firms like Chubb, argued that there was ample evidence that NotPetya was an instrument of the Russian government and part of ongoing hostilities against Ukraine. In other words, it was an act of warlike behaviour covered by the war exclusion. The court, however, sidestepped the question of who was responsible for the assault. Instead, it said that insurers did nothing to change the language of their contracts to suggest that the war exclusion included cyber-attacks. It said it was reasonable for Merck to think that the exclusion applied only to “traditional” warfare, ie, tanks and troops, not worms, bugs and hackers.It is not the final verdict. A similar war-exclusion case involving Mondelez and its insurers continues in an Illinois court. But though it marked a victory for Merck, it may be a Pyrrhic one for companies at large. That is because many insurers are now seeking to strengthen the language in policies the better to shield themselves from payouts related to state-sponsored cyber-mischief. If a NotPetya-like virus were to come from Russia’s warmongering in Ukraine and burrow itself into the world’s supply chains, insurers are keen to ensure they limit their exposure to it. The consequences of that for corporate victims could be severe.The evidence suggests companies have a lot to fear. Last year a report by HP, a technology firm, said that state-sponsored attacks had doubled between 2017 and 2020, and that businesses were the most common targets. Increasingly, the state hackers’ weapon of choice is malware inserted into the software or hardware of suppliers, which is particularly hard for companies up the value chain to detect. Unlike other cyber-criminals, who attack and move on, states have strategic patience, lots of resources and are above the law within their own borders. They cover their tracks well, too, so it can be particularly hard to attribute blame for an attack.In the face of that, the insurance industry’s caution is understandable. It is already facing a surge in ransomware claims from companies during the covid-19 pandemic, which is driving up the price of cyber-insurance. The NotPetya attack revealed the risk of “silent cyber”, or unspecified cyber-risk hidden within insurance contracts. These could pose a systemic risk to the industry in the event of a large-scale, correlated attack. Partly in response to such threats, Lloyd’s Market Association, an advisory group, recently issued four model clauses for excluding war coverage from cyber-insurance policies. They enable insurance companies to customise their exclusions more easily and give companies more clarity on which risks are covered and which aren’t. But they appear to protect the insurers more than the insured.It is still an evolving market. The Merck war-exclusion judgment relied on case law rendered before cyber was even a word. The cyber-insurance industry, though growing fast, is still small and immature. Eventually, the actuarial techniques for gauging cyber-risk will improve, and the insurance industry will get better at requiring clients to introduce the cyber-equivalent of fire alarms and sprinkler systems to minimise danger. For now, though, the risk of considerable confusion persists if something close to a cyber-war were to break out.Self-isolationSo what should companies do? A well-known checklist of safety measures to implement includes things like two-factor authentication and swift software updates, which help keep hackers at bay. In light of the danger of infection along the supply chain, either from compromised hardware or software, firms should painstakingly assess their contingent exposures: factories or offices in far-flung locations, outsourced IT, cloud computing and even cyber-security itself.Corporate boards need to have a stronger grasp of the threat levels. As one former cyber-spook says, they need not just gender and racial diversity but technological diversity, too, in order to grill the company’s techies on cyber-defences. Furthermore, they need to recognise cyber-war as one of the growing number of geopolitical risks that firms face. Ensuring that any of a firm’s contact points with Ukraine and Russia are not a vulnerability for the rest of its operations is the first of many steps they should take. ■For more expert analysis of the biggest stories in economics, business and markets, sign up to Money Talks, our weekly newsletter.This article appeared in the Business section of the print edition under the headline “Cyber-rattling” More

VLADIMIR PUTIN, Russia’s president, has portrayed his aggression on the Ukrainian border as pushing back against Western advances. For some time he has been doing much the same online. He has long referred to the internet as a “CIA project”, and his deep belief that the enemy within and the enemy without are effectively one and the same means that if Alexei Navalny, Mr Putin’s foremost internal foe, uses YouTube—his video of the president’s seaside palace was viewed more than 120m times—then YouTube and its parent, Alphabet, are enemies, too.Faced with such “aggression” he wants a Russian internet that is secure against external threat and internal opposition. He is trying to bring that about on a variety on fronts: through companies, the courts and technology itself.In early December VK, one of Russia’s online conglomerates, was taken over by two subsidiaries of Gazprom, the state-owned gas giant. In the same month a court in Moscow fined Alphabet, Google’s parent company, a record $98m for repeated failure to to delete content the state deems illegal. And Mr Putin’s regime began using hardware it has required internet service providers (ISPs) to install to block Tor, a tool widely used in Russia to mask online activity. All three actions were part of the country’s effort to assure itself of online independence by building what some scholars of geopolitics, borrowing from Silicon Valley, have begun calling a “stack”. His efforts could serve as an inspiration, and a model, for tyrants elsewhere.In technology, the stack is the sum of all the technologies and services on which a particular application relies, from silicon to operating system to network. In politics it means much the same, at the level of the state. The national stack is a sovereign digital space made up not only of software and hardware (increasingly in the form of computing clouds) but also infrastructure for payments, establishing online identities and controlling the flow of information.Benjamin Bratton, a political philosopher at the University of California, San Diego, sees the stack as a set of new dimensions for the state, piled up one on top of the other, each of them analogous to the territory defined by its physical borders. The default stack is largely American, because that is where the internet grew up. But other places are trying to differentiate their stacks, some seeing opportunity, some staving off perceived threats. The EU, with ambitions to become the world’s super-regulator for all things digital, is putting together a what it hopes will be a more open stack, less tied into proprietary technologies and monopolistic applications. India, Japan and Taiwan are all working on their own distinct digital edifices.Most germane to an autocrat like Mr Putin is what has gone on in China. China built its internet with censorship in mind: the Great Firewall, a deep-rooted collection of sophisticated digital checkpoints allows traffic to be filtered with comparative ease; the size of the Chinese market means that indigenous companies, which are open to various forms of control, can successfully fulfil all its users’ needs; and the state has the resources for a lot of both censorship and surveillance.Mr Putin and other autocrats covet such power. But they cannot get it. It is not just that they lack China’s combination of rigid state control, economic size, technological savoir-faire and stability of regime. They also failed to start 25 years ago. So they need ways to achieve what goals they can piecemeal, by retrofitting new controls, incentives and structures to an internet that has matured unsupervised and open to its Western begetters.Russia’s efforts, which began as purely reactive attempts to lessen perceived harm, are becoming more systematic. Three stand out: the creation of domestic technology, controlling the information that flows across it and, perhaps most important, building the foundational services that underpin the entire edifice.Take the technology first—microprocessors, servers, software and the like. Although Russia has some notable firms in these areas—Baikal and Mikron in semiconductors, ABBYY and Kaspersky in software—for the most part companies and government agencies prefer Western wares. Russian companies’ share of the semiconductor market was less than 1% of the global total in 2020 according to EMIS, a data provider. In servers and business software the situation is much the same.The government has made moves to restart a chipmaking plant in Zelenograd near Moscow, the site of a failed Soviet attempt to create a Silicon Valley. But it will not operate at the cutting edge. So although an increasing number of chips are being designed in Russia, they are almost all made by Samsung and TSMC, a South Korean and a Taiwanese contract manufacturer. This could make the designs vulnerable to sanctions. An added problem is that they are often not up to snuff. Some experts have doubts about the capabilities of Russia’s home-grown Elbrus processors designed by a firm called the Moscow Centre of SPRAC Technologies.For crucial applications such as mobile-phone networks Russia remains highly reliant on Western suppliers, such as Cisco, Ericsson and Nokia. Because this is seen as leaving Russia open to attacks from abroad, the industry ministry, supported by Rostec, a state-owned arms and technology giant, is pushing for next-generation “5G” networks to be built with Russian-made equipment only. The country’s telecoms industry does not seem up to the task. And there are internecine impediments. Russia’s security elites, the siloviki, do not want to give up the wavelength bands best suited for 5G. But the only firm that could deliver cheap gear that works on alternative frequencies is Huawei, an allegedly state-linked Chinese electronics conglomerate which the siloviki distrust just as much as security establishments in the West do.It is at the hardware level that Russia’s stack is most vulnerable. Sanctions which might be raised if Russia were to invade Ukraine would probably see the country as a whole treated as Huawei now is by America. Any chipmaker around the world that uses technology developed in America to design or make chips for Huawei needs an export licence from the Commerce Department in Washington—which is usually not forthcoming. If the same rules are applied to Russian firms, anyone selling to them without a licence could risk becoming the target of sanctions themselves. That would see the flow of chips into Russia slow to a trickle.When it comes to software Russia’s government is using its procurement power to amp up demand. Government institutions, from schools to ministries, have been encouraged to dump their American software, including Microsoft’s Office package and Oracle’s databases. It is also encouraging the creation of alternatives to foreign services for consumers including TikTok, Wikipedia and YouTube.From Russia, with likesHere the push for indigenisation has a sturdier base on which to build. According to GroupM, the world’s largest media buyer, Yandex, a Russian firm which splits the country’s search market with Alphabet’s Google, and VK, a social-media giant, together earned $1.8bn from advertising last year, more than half of the overall market. VK’s VKontakte and Odnoklassniki trade places with American apps (Facebook, Instagram) and Chinese ones (Likee, TikTok) on the top-ten downloads list (see chart 1).This diverse system is obviously less vulnerable to sanctions—which are nothing like as appealing a source of leverage here as they are elsewhere in the stack. Making Alphabet and Meta stop offering YouTube and WhatsApp, respectively, in Russia would make it much harder for America to launch its own stack-to-stack warfare against the country—as would disabling Russia’s internet at the deeper level of protocols and connectivity. And it would push Russians to use domestic offerings more, which would suit Mr Putin well.As in China, Russia is seeing the rise of “super-apps”, bundles of digital services where being local makes sense. Yandex is not just a search engine. It offers ride-hailing, food delivery, music-streaming, a digital assistant, cloud computing and, one day, self-driving cars. Sber, Russia’s biggest bank, is eyeing a similar “ecosystem” of services, trying to turn the bank into a tech conglomerate. In the first half of 2021 alone it invested $1bn in the effort, on the order of what biggish European banks spend on information technology (IT).Structural changes in the IT industry are making some of this Russification easier. Take the cloud. Its data centres use cheap servers made of off-the-shelf parts and other easily procured commodity kit. Much of its software is open-source. Six of the ten biggest cloud-service providers in Russia are now Russian, according to Dmitry Gavrilov of IDC, a research firm. He says most successful ones are “moving away from proprietary technology” sold by Western firms (with the exception of chips). And as in the West, cloud computing has allowed specialised providers of online software to break through; in Russia this has included amoCRM, Miro and New Cloud Technologies.Import substitution is a slow process and success is by no means guaranteed. However, it can no longer be considered a “joke”, in the words of Andrei Soldatov, editor of Agentura.ru, an online portal, and co-author of “The Red Web”, a book about digital activism in Russia. “The government is making steady progress in dragging people into a domestic digital bubble,” he recently wrote.If technology is the first part of Russia’s stack, the “sovereign internet” is the second. It is code for how a state controls the flow of information online. In 2019 the government amended several laws to gain more control of the domestic data flow. In particular, these require ISPs to install “technical equipment for counteracting threats to stability, security and functional integrity.” This allows Roskomnadzor, Russia’s internet watchdog, to have “middle boxes” slipped into the gap between the public internet and an ISP’s customers. Using “deep packet inspection” (DPI), a technology used at some Western ISPs to clamp down on pornography, these devices are able to throttle or block traffic from specific sources (and have been used in the campaign against Tor). DPI kit sits in rooms with restricted access within the ISPs’ facilities and is controlled directly from a command centre at Roskomnadzor.This is a cheap but imperfect version of China’s Great Firewall, says Roya Ensafi of Censored Planet, a project at the University of Michigan to measure internet censorship. It has improved Roskomnadzor’s ability to block sites and interrupt the virtual private networks many use to camouflage internet usage. It also allows the regulator to block, as it did during protests in 2019, live-video streaming without taking down whole mobile-phone networks.Complementing the firewall are rules that make life tougher for firms. In the past five years Google has fielded 20,000-30,000 content-removal requests annually from the government in Russia, more than in any other country (see chart 2). From this year 13 leading firms—including Apple, TikTok and Twitter—must employ at least some content moderators inside Russia. This gives the authorities bodies to bully should firms prove recalcitrant.The ultimate goal may be to push foreign social media out of Russia, creating a web of local content controllable through courts, corruption and bully boys. But this China-level control would be technically tricky. The DPI boxes are unable to filter out all foreign traffic. It would also be unpopular: Russians are keen on YouTube and WhatsApp. And it would make life more difficult for Russian influence operations, such as those of the Internet Research Agency, to use Western sites to spread propaganda, both domestically and abroad.A view to instill“Russia is less about blocking and more about shaping the information environment,” says Justin Sherman of the Atlantic Council, a think-tank. Strategically placed constraints, both online and offline, should suffice to guide the digital flow without hard barriers. Making foreign services less reliable will shift consumers towards domestic ones. Facing throttling, fines or worse, Western firms are likely to comply with government demands, as they did when leant on to remove apps Mr Navalny’s supporters designed to show voters which opposition candidates were best placed to win elections.Russia’s homegrown stack would still be incomplete without a third tier: the services that form the operating system of a digital state and thus provide it its power. In its provision of both e-government and payment systems, Russia puts some Western countries to shame. Gosuslugi (“state services”) is one of the most visited websites and downloaded apps in Russia. It hosts a shockingly comprehensive list of offerings, from passport application to weapons registration. Even critics of the Kremlin are impressed, not least because Russia’s offline bureaucracy is hopelessly inefficient and corrupt. Sergey Sanovich of Princeton University observes that by leapfrogging into the virtual world, leaders in Moscow showed they could deliver, and got a better grasp of what agencies far from the capital are doing. Privacy concerns, which can be a barrier to online government, were not much of a worry.The desire for control also motivated Russia’s leap in payment systems. In the wake of its annexation of Crimea sanctions required MasterCard and Visa, which used to process most payments in Russia, to ban several banks close to the regime. In response, Mr Putin decreed the creation of a “National Payment Card System”, which was subsequently made mandatory for many transactions. Today it is considered one of the world’s most advanced such systems. Russian banks use it to exchange funds. The “Mir” card which piggybacks on it has a market share of more than 25%, says GlobalData, an analytics firm.Other moves are less visible. A national version of the internet’s domain name system, currently under construction, allows Russia’s network to function if cut off from the rest of the world (and give the authorities a new way to render some sites inaccessible). Some are still at early stages. A biometric identity system much like India’s Aadhar, aims to make it easier for the state to keep track of citizens and collect data about them while offering new services (Muscovites can now pay to take the city’s metro just by showing their face). A national data platform would collect all sorts of information from tax to health records—and could boost Russia’s efforts to catch up in artificial intelligence (AI).These plans must be taken with a dollop of salt. “Russia’s industrial policy seems that of a superpower, but in reality it is an economic minnow,” says Janis Kluge of the German Institute for International and Security Affairs, a think-tank. Even if it had the means, he says, it does not seem willing to spend what it takes. Mr Putin has said that national capabilities in AI will determine who becomes “the ruler of the world”. But Russia is not making those capabilities a particularly high priority.That said, as technology gets cheaper and more openly available, a country like Russia will be able to do ever more with only a modest effort. Stacks are modular; their layers can in principle be swapped out. You do not have to control all of them to get your way. In other words, Russia does not need the latest and smallest semiconductors, say, to build a serviceable edifice on top of what it has; and if it is hard to reach what is available elsewhere, serviceable may be good enough. The country’s bureaucrats have shown that they are able to learn quickly and improvise around technologies they lack.The Kremlin’s progress is being observed by others, including Iran (which requires censorship by software at ISPs), Kazakhstan (which may delegate some of its digital transformation to Sber) and Turkey (which demands the physical presence of foreign firms’ content moderators). They may back Russia diplomatically as it promotes its digital ambitions. Together with China, Russia has stalled UN talks aimed at defining responsible state behaviour in cyberspace, instead insisting on “information sovereignty”—code for doing whatever it pleases. Now it wants a Russian, Rashid Ismailov, to take over as secretary-general of the International Telecommunication Union (ITU), which governs swathes of the telecoms world. Mr Ismailov’s resumé includes stints as a deputy telecoms minister and Huawei executive.Russia wants the ITU to replace the Internet Corporation for Assigned Names and Numbers as the overseer of the internet’s address system. America and its allies will block this. But the idea appeals to countries desiring stack sovereignty, which may be enough to win Mr Ismailov the votes he needs to beat Doreen Bogdan-Martin, an ITU official from America, in October, when the new secretary-general will be chosen.Try another dayIf push comes to shove in Ukraine, the strength of Russia’s stack against sanctions, and perhaps other forms of attack, will be tested. The costs could be high: capabilities would be lost and networks degraded. Russia may become more dependent on Chinese hardware and software, something its own elites fear (though this would hardly be a win for the West).Whatever the upshot of such “stack-to-stack warfare”, as Mr Bratton calls it, the Kremlin’s efforts have shown would-be imitators that there is plenty of mileage in trying to take control of what layers of the internet you can, and of aligning yourselves with fellow travellers. New ways of instantiating the state offer new forms of influence, diplomacy and common cause—as well as of war. For more expert analysis of the biggest stories in economics, business and markets, sign up to Money Talks, our weekly newsletter. More

SHORTAGES AND bottlenecks have been a source of constant frustration for manufacturers around the world for two pandemic-afflicted years. For a handful of companies in the business of keeping factories running and supply chains intact, these frustrations have been a source of cheer—and profits. Japanese makers of industrial equipment, in particular, have seen orders surge as companies turned to automation, first amid the disruption wrought on human workforces by covid-19, then as a result of tight labour markets and rising wage costs.The world’s stock of industrial robots has tripled in the past decade. According to the International Federation of Robotics, a trade group, Japan furnishes 45% of new ones each year. It also produces lots of other automation equipment, from laser sensors to inspection kit. Even after the recent sell-off in tech stocks, Japan’s four standout gear producers—Keyence, Fanuc, SMC, and Lasertec—are collectively worth two and a half times what they were five years ago (see chart). Last year the founder of Keyence, Takizaki Takemitsu, briefly became Japan’s richest man. His $29bn fortune is half as large again as that of Son Masayoshi, a flamboyant tech investor who is corporate Japan’s most globally recognisable face. Mr Takizaki’s firm and its fellow equipment-makers are hardly household names. But the hardware they produce is becoming as mission-critical to many industrial supply chains as semiconductors are.It is no surprise that Japan, a famously robot-loving place, has spawned a strong Automation Inc. Just-in-time manufacturing, pioneered by efficiency-obsessed Japanese companies such as Toyota in carmaking or Panasonic in consumer electronics, has involved replacing humans with machines for decades. This source of competitive advantage became an existential necessity for domestic manufacturers after Japan’s working-age population began to shrink in the 1990s. Today it is becoming one for other rich countries as they enter demographic dotage. Keyence and SMC now derive more than half their revenues from abroad. Fanuc and Lasertec are even more international, with more than 80% of sales coming from overseas.Some of the new foreign demand is the result of the world’s insatiable hunger for computer chips. SMC, which sells pneumatic control devices to chipmakers, has seen its business boom, especially as places including America and Europe strive to bring more semiconductor production home, says Masahiro Ota, who sits on SMC’s board. Lasertec enjoys a near-monopoly on inspection tools for the most advanced semiconductor photomasks—plates through which circuit patterns are etched onto silicon wafers. Its share price has ballooned four-fold since the start of 2020, making it one of the best-performing blue-chip stocks in Asia. Keyence’s precision sensors are likewise crucial for the detection of flaws in semiconductor surfaces.The companies’ devices are, of course, also handy in other sectors. Fanuc, which makes large factory-floor robotic arms, has long been a fixture of car assembly lines. Mike Cicco, who runs Fanuc’s American operations, notes that the development of electric cars requires a range of new capabilities on the part of carmakers—and that in turn necessitates new types of robot. Fanuc expects to supply Ford’s factory in Cologne, in Germany, with 500 robots this year as the plant becomes the Ford Cologne Electrification Centre.Being indispensable has proved to be lucrative. All four stars of Japan’s automation-industrial complex boast operating-profit margins of over 20%. That of Keyence, the most profitable of the lot, exceeds 50%. The firm has reported record net profits in each of the past three quarters. Like chip firms such as Nvidia, Keyence does not manufacture products but rather designs them and assists customers in deploying them in their factories. Lasertec, too, does little of its own manufacturing. This capital-light approach helps sustain profits. Keyence spends just 3% of its net sales on research and development (R&D). Similarly, SMC spends around 4%. Fanuc does make almost all its products independently and invests more in production capacity and R&D. But it uses that capital efficiently, not least, as befits a robot-maker, by deploying plenty of its own robots to build robots for customers. Its biggest “lights out” factory can run for more than a month with no pricey human operators around.Japan’s automation firms also owe some of their success to corporate culture. SMC maintains a network of 6,000 salespeople who double as systems engineers with in-depth knowledge of customers’ equipment. Keyence uses no middlemen to sell its products, relying entirely on its own sales force. As with SMC, many are engineers, who spend a lot of time on customers’ factory floors identifying niggles and tweaks that might otherwise go unnoticed. They are rewarded handsomely for their efforts. Nikkei, a Japanese publisher, reports that average salaries at Keyence exceeded $150,000 in the last fiscal year.The automation stars, like Japan Inc as a whole, tend to be less generous with shareholders. Most sit on piles of cash; Keyence held over $10bn in current assets in the last financial year. The reserved character of the companies and their tightfistedness is so well-established that some investors say any sudden shifts in that attitude may be a sign of big and possibly unwelcome changes at the firms.Investors have to rely on such rune-reading because it is not always clear what is going on inside the companies, at least by contemporary Western standards of open shareholder relations. SMC’s “traditional Japanese approach to corporate governance”, as Baillie Gifford, a tech-focused British asset manager, delicately put it in 2020, offers only limited engagement with shareholders. One asset manager with a stake in Keyence reports never speaking directly with its management.As the companies become ever more international, they will face pressure to be more candid—and less frugal, both with payouts to shareholders and with investments. Fanuc increased its dividend sharply in 2015 under pressure from Third Point, an American activist hedge fund. As Japan becomes less averse to gadfly investors, Automation Inc should expect more such calls. To maintain their innovative edge, meanwhile, the firms may need to spend considerably more on R&D. Amid tech-inflected geopolitical tensions with the West, China wants to reduce its reliance on foreign suppliers of all manner of advanced technology, including robotics. If successful, the Chinese strategy would at once deprive the Japanese firms of a big market and create new global rivals. Becoming indispensable is one thing. Staying so is quite another. ■For more expert analysis of the biggest stories in economics, business and markets, sign up to Money Talks, our weekly newsletter.This article appeared in the Business section of the print edition under the headline “Automation Inc” More

THROUGHOUT 2021 corporate profits in America seemed immune to infection, inflation and snarled-up global supply chains. With most of the country’s biggest firms having reported their latest quarterly results, revenues and earnings in the last three months of the year seem poised to set another record. The technology industry’s annual profits rose by a quarter, compared with 2020. Among property firms in the S&P 500 index they swelled by a third. Net-profit margins have edged down from their recent peak in the second quarter of 2021 and analysts forecast lower earnings and sales in the first three months of 2022. It may prove to be a brief sniffle caused by unusually strong seasonal factors. But America Inc no longer looks totally impervious.■For more expert analysis of the biggest stories in economics, business and markets, sign up to Money Talks, our weekly newsletter.This article appeared in the Business section of the print edition under the headline “How long can America Inc’s profits keep rising?” More

THE RICH world has a sleep deficit. The average American adult snoozes almost two hours less than their great grandparents did. More than a third of Americans get less than seven hours of kip a night. The resulting fatigue has been linked to Alzheimer’s disease, hypertension and other ailments. It may cost America’s economy as much as $400bn a year, according to one study. Other wealthy countries are similarly sleepless. Consumption of alcohol and caffeine are partly to blame, as is exposure to phone and computer screens. Ironically, people are turning to some of those same devices for help.Tiny sensors are now more easily embedded into wearable gadgets to observe users overnight. Consumer-electronics giants such as Google, Samsung and Huawei offer sleep-related technology in their gadgets. Although Apple seems to be winding down Beddit, a Finnish maker of bed sensors it acquired in 2017 for an undisclosed amount, it has incorporated sleep functionalities into its new smart watches.Specialist “sleep-tech” startups offer fancier wares. Oura Health, also from Finland and valued at nearly $1bn, sells a $300 titanium ring that weighs a few grams and has built-in heart-rate, oxygen and activity monitors; Kim Kardashian is a fan. Kokoon, a British firm, makes a wireless headset whose tiny earbuds play relaxing sounds while sensors infer the sleep stage from blood-oxygen levels. Eight Sleep, an American one, charges $2,000 for its app-synched mattress that heats up and cools as the sleeper’s body temperature changes through the night.The combination of more sleeplessness and better technology has led to a boom in the sleep-assistance industry. Global Market Insights, a research firm, reckons that worldwide revenues from sales of such gizmos reached $12.5bn in 2020 and could be more than triple that in five years. Matteo Franceschetti, boss of Eight Sleep, thinks the addressable market for his company is “literally everyone in the world”. After all, everybody sleeps.True. But not everybody sleeps poorly (or can afford to splurge $2,000 on his firm’s self-styled “Lamborghini of mattresses”). And the technology, though it is improving, remains far from perfect. Sleeping with a watch strapped to your wrist is irritating, and the battery may die overnight. Your correspondent struggled to wear the Kokoon headset over the satin scarf protecting her hair, and the “brown noise” designed to drown out snoring sounded more like the jarring static of an old television set.There are problems with sleep-tech’s business models, too. People can get bored of wearables, and frustrated when the touted improvements fail to materialise. According to a survey last year by Rock Health Advisory, a consultancy, almost 40% of sleep-wearables users abandoned their devices, mostly because they did not have the desired soporific effect. Kokoon, Oura and Eight Sleep have all recently introduced membership models to try and keep people updating their devices. Subscriptions give the companies a more stable revenue stream than one-off device sales, as well as providing data that can then be used to improve their products. But it can also be interpreted as an implicit acknowledgement that the devices are not an instant cure. (Oura says that it now offers various other insights into ring-wearers’ health that are not directly related to sleep.)Many scientists worry that, as with many emerging consumer-health technologies, sleep-tech often lacks the gold-standard randomised controlled clinical studies where it is tested on many patients and against placebos. Ingo Fietze runs a sleep centre at Charité Berlin, a big university hospital, and studies novel gadgets and mattresses at a private lab he set up on the side. He says that when he asked Samsung, a South Korean device-maker, and Huawei, a Chinese one, to share the methods behind their watches’ metrics, he did not hear back. In any case, says Mr Fietze, no existing wearables, which track sleep using various proxy measures, can match a clinical polysomnogram (PSG), which takes data directly from the brain using electrodes. Samsung did not respond to a request for comment. Huawei says its device measures sleep duration with accuracy comparable to a PSG.Sleep-tech may, scientists concede, help mild insomniacs and sensitive sleepers decide whether they need clinical interventions. Monitoring blood-oxygen in real time, as some wearables do, can help identify disorders including sleep apnea, a condition whose sufferers stop breathing while they sleep and which afflicts perhaps 1bn people around the world. But ultimately, Mr Fietze believes, “no gadget can make your sleep better.” If consumers in need of more shuteye reach a similar conclusion, sleep-tech investors’ dreams of riches may turn into a profitless nightmare. ■For more expert analysis of the biggest stories in economics, business and markets, sign up to Money Talks, our weekly newsletter.This article appeared in the Business section of the print edition under the headline “Slumber party” More

CORPORATE CULTURE is often like mist—indubitably there but hard to pin down. Occasionally it solidifies into something ugly. Take the following figures from an external investigation commissioned by Rio Tinto, a global mining giant, into its workplace culture. Almost half of Rio’s employees report having experienced bullying in the past five years. Almost a third of its female workers have endured sexual harassment; 21 women reported an actual or attempted sexual assault. Two in five Australian Aboriginal and Torres Strait Islanders employed by the company have experienced racism.The report is an admirably open attempt to face up to a toxic culture. As well as survey data, it contains testimony from interviews and focus groups. It was published on Rio’s website earlier this month, along with an apology from Jakob Stausholm, the firm’s boss. Such unusual transparency seems to be building trust: half of the firm’s employees said they were extremely or very confident that Rio would make meaningful progress in stamping out sexual harassment and racism. The report is a product of specific circumstances. Rio’s reputation was badly tarnished in 2020, when it destroyed Juukan Gorge, a mining site in Western Australia whose ancient rock shelters were sacred to indigenous people. That cost Mr Stausholm’s predecessor his job, kick-starting efforts to change the way the firm was run. Rio’s culture is traceable, at least in part, to the idiosyncrasies of the mining industry. Its workforce is 80% male, and the worst behaviour occurred on remote sites where employees fly in or drive in for stays of several days, or live full-time in company housing. Machismo and isolation make for poor bedfellows. But it would be a mistake to regard Rio’s soul-searching as a curio from the world of alpha males and excavators. For both its findings and the fact of its existence hold wider lessons. First, it shows how a corporate culture can rot. The worst abuses may have been more prevalent in the firm’s remote reaches but they were present at its heart, too. The highest rate of sexual harassment was found in the firm’s iron-ore division, but next came Rio’s strategy, sustainability and development group. Widespread suspicion of the company’s internal reporting mechanisms and a fear of speaking out are evident. “The minute you raise an issue about a senior leader, you’re done,” said one employee. “I don’t want to rock the boat so hard that I fall out of it,” echoed another. Interviewees accused Rio of rewarding bullies, and of pushing high performers up the corporate ladder irrespective of how they behaved. Among other things, the firm says it will set up a specialist unit designed to respond to complaints of harmful behaviour, and to provide support to people who blow the whistle. Whatever the right answer, the report raises questions that executives in all organisations confront: what to do with talented jerks, and how to make sure people voice concerns if something is going badly wrong. Second, it may be a harbinger of wider demand for data on corporate culture. For all that managers bang on about people being an organisation’s greatest asset, precious little information is available to outsiders about how employees are treated and encouraged to behave. This may be because of the mist problem: it is hard to measure culture. It may be because investors haven’t much cared. That may be changing. Labour shortages have focused attention on how well firms retain workers. Research from Donald Sull at the Massachusetts Institute of Technology and his co-authors finds that a toxic culture is ten times more important than pay in predicting industry-adjusted staff turnover. Movements like #MeToo and Black Lives Matter have pushed issues of gender and racial equity up the corporate agenda. Allegations of sexual misconduct have battered the reputations of Axel Springer, a media giant, and Activision Blizzard, a video-game publisher just acquired by Microsoft. Late last year investors in the software giant adopted a shareholder proposal requiring it to report on its own sexual-harassment policies. Regulators are making noises about more disclosure on human capital; Gary Gensler, chairman of America’s Securities and Exchange Commission, wants proposals in this area.Rio Tinto’s problems are extreme. But they are not unique. And in opening up about its corporate culture, it is, in one way at least, ahead of its time. More

IN 2013 THE EU launched an ambitious project. The aim was to double the share of microchips made in Europe to 20% of the global total by 2020. Nearly a decade later it remains stubbornly stuck at 10%. If that were not bad enough, Europe no longer makes any of the most advanced chips of the sort that go into data centres or smartphones (see chart). So, prompted by shortages of semiconductors and their growing importance for all sorts of industries, the bloc is having another go.Judged by numbers alone, the EU’s new Chips Act, unveiled on February 8th, could move the needle. It is meant to generate public and private investment of more than €43bn ($49bn), about as much as a similar package working its way through America’s Congress. More than two-thirds of this money is supposed to take the form of state subsidies for new leading-edge chip fabrication plants, or “mega fabs”—thanks to a more generous interpretation of EU restrictions on state aid. The rest will go to other chipmaking infrastructure.Reality is likely to prove trickier. To understand why, it helps to see the semiconductor industry not just as a collection of huge fabs, of which the most sophisticated can cost more than $20bn a pop, but as a global ecosystem of thousands of companies. Even more than in other high-tech industries, research and development (R&D) usually takes years and costs billions. New chips are designed by specialised firms using complex software made by other companies still. And after chips leave a fab, contract manufacturers assemble, test and package them (ATP, in the lingo).Seen though this ecosystemic lens, the EU’s position is both stronger and weaker than its small share of global chip output might suggest. Start with the strengths. The continent maintains a leading position in semiconductor R&D. One of the industry’s main brain trusts, the Interuniversity Microelectronics Centre (better known as IMEC), is based in Belgium.Europe’s firms also make many of the machines that make fabs tick. ASML, a Dutch firm with a market value of €230bn, is the sole global supplier of the lithographic equipment without which fabs cannot etch the most advanced processors. Only Nvidia, an American chip-designer, and Taiwan Semiconductor Manufacturing Company (TSMC), the world’s biggest contract manufacturer of chips, are worth more. An array of smaller European outfits enjoy dominant positions in the complex chipmaking supply chain. Carl Zeiss SMT makes lenses for ASML’s lithography machines (and is co-owned by it). Siltronic manufactures silicon wafers onto which chips are etched. Aixtron manufactures specialised gear to deposit layers of chemicals onto those wafers to make circuits.Once you widen the aperture to the whole ecosystem, Europe’s biggest chipmakers, Infineon, NXP and STMicroelectronics, also appear less benighted. Yes, half of the continent’s capacity is for chips with structures (“nodes”) measuring 180 nanometres (billionths of a metre) or more, generations behind the technological cutting edge, dominated by TSMC and Samsung of South Korea, whose transistors come in at a few nanometres. But those nano-electronics are most useful for consumer devices, the bulk of which are assembled in Asia. By contrast, the larger European nodes are sufficient for the continent’s many industrial firms that require specialised silicon for things such as cars, machine tools and sensors. “European chipmakers focus on their customer base,” explains Jan-Peter Kleinhans of SNV, a German think-tank.If the Chips Act is a guide, European policymakers worry that these genuine strengths are not enough to offset the EU’s weaknesses. Besides lacking cutting-edge fabs, Europe is short of companies with the know-how to design the smallest chips, such as Nvidia. It is similarly behind in ATP, where most capacity is in China and Taiwan. Once approved by member states and the European Parliament, the EU law is meant to help Europe catch up. Besides the nearly €30bn for mega-fabs, it has pencilled in €11bn for things like a virtual chip-design platform open to all comers and other infrastructure, including pilot production lines for leading-edge chips. But half of that is to come from member states and the private sector. The EU’s contribution of less than €6bn will, as with the bloc’s other programmes, come with many bureaucratic strings attached.A bigger problem is the act’s focus on luring giant chipmakers to build mega-fabs. TSMC and Intel, its American rival, have signalled they would consider Europe only if governments shoulder a big part of the costs (40% in Intel’s case). To enable such deals, the first of which is expected in weeks, the European Commission wants to relax state-aid rules to let member states subsidise such fabs “up to 100% of a proven funding gap” if they are “first-of-a-kind” or would “otherwise not exist in Europe”.If such criteria were meant to avert a subsidy race, they look copious and fuzzy enough for countries to try and game them. Worse, the resulting fabs may end up underused. By the time they are ready in a few years, the chip shortage may have turned into a glut. And if the EU’s efforts to boost Europe’s chip-design firms fail, European fabs would have to rely on foreign chip-designers for custom. Why, asks Mr Kleinhans, would American firms choose to have their chips manufactured in Europe rather than in Asia or at home?Thierry Breton, the EU commissioner in charge of industrial policy, envisions a Europe of “mega fabs” that not only serve the continent’s own demand, but world markets. Europe may be better off propping up its chip ecosystem by investing in things like basic research. Mr Breton doesn’t need to pick Europe’s chipmaking winners. As the EU’s semiconductor stars show, the market can do that just fine.For more expert analysis of the biggest stories in economics, business and markets, sign up to Money Talks, our weekly newsletter. More