Philippa Kaur

LAST NOVEMBER something strange happened in Mountain View. A thick fog enveloped the headquarters of Alphabet, the parent company of Google. Not the meteorological sort—this stretch of Silicon Valley is reliably sunny. It was a fog of confusion. Its cause was ChatGPT, an artificially intelligent conversationalist created by OpenAI, a startup backed by Microsoft. The effect was, by all accounts, panic. ChatGPT was giving uncannily humanlike answers to questions put to it by users. And answering questions is the bread and butter of Google’s lucrative search business. Were OpenAI and Microsoft, which in February launched an enhanced version of its Bing search engine, about to eat Google’s lunch?Eight months on the mist has mostly cleared. On July 25th the company reported another set of solid quarterly results. Revenues rose by 7% year on year, to $75bn. It continues to create piles of cash: in the 12 months to June it raked in $75bn of operating profit. Bing has taken no discernible bite out of Google’s share of global monthly search queries, which remains above 90%. Most important, Google has put to rest any notion that it has fallen behind technologically. In May Sundar Pichai, chief executive of both Google and its corporate parent, unveiled more than a dozen AI-powered products at I/O, an annual event for software developers. These included AI tools for Gmail, Google Maps and Google Cloud. Investors found it reassuring—not least after a rushed launch in February of Bard, Google’s chatbot, during which the AI helper made a factual error. Since then the firm has launched AI products and features left and right. On July 12th it brought out NotebookLM, an AI-assisted note-taking tool trained on a user’s documents. On the same day Nature, a scientific journal, published a paper by Google researchers describing an AI model that matched human doctors’ responses to questions about the right treatment for patients. A day later it expanded a now less error-prone Bard, proficient in more than 40 human languages and over 20 computer ones, to the EU. Work on an AI model to eclipse ChatGPT, codenamed Gemini, is proceeding apace. Having nearly fallen below $1trn in November, Alphabet’s market value is back up to $1.7trn. Crisis over? In the short run, probably. Like all heart-in-mouth moments, though, the chatbot panic invites broader questions: about the current state of one of the world’s biggest firms, its future and—as Google turns 25 in September—about the demands of different stages of corporate life. The view from the top Alphabet is, without a doubt, one of the greatest business successes of all time. Six of its products—Google search, the Android mobile operating system, the Chrome browser, Google Play Store for apps, Workspace productivity tools and YouTube—boast more than 2bn monthly users each. Add those with hundreds of millions of users, such as Google Maps or Google Translate, and, by one reckoning, humans collectively spend 22bn hours a day on Alphabet’s platforms.The ability to command so much attention is worth a lot of money to the people who want a slice of it, namely advertisers. Since going public in 2004 Google’s revenue, 80% of which comes from online ads, has grown at an average annual rate of 28%. In that period it has generated a total of $460bn in cash after operating expenses, virtually all of it from advertising. Its share price has risen 50-fold, making it the world’s fourth-most-valuable company. Given these eye-popping numbers, it may seem churlish to ask why Alphabet isn’t doing better. In fact, the question is warranted, and is being asked by Mr Pichai, his underlings and investors alike. The company finds itself at a delicate juncture—not only, or even primarily, because of AI. The core digital-ads business is maturing, with sales growth no longer consistently in double digits and increasingly tied to economic cycles. At the same time, finding new sources of material growth is difficult for a company that brings in $300bn in annual revenues. This quest is further complicated by investors calling for greater cost efficiency and capital discipline, which in turn requires a shake-up of its free-wheeling corporate culture. Consider the cash cow. Throughout the 2010s digital advertising seemed invulnerable to the business cycle. In good times advertisers spent like there was no tomorrow. In worse ones they diverted some of their non-digital marketing budgets online, where Google and other giants like Facebook (now Meta) offered to target adverts more precisely than a TV commercial or a page in a glossy magazine could. Now, as the online share of total ad spending touches two-thirds, businesses have smaller non-digital ad budgets to eat into. Insider Intelligence, a data firm, expects global sales of digital ads to increase by 10% or less annually in the next few years, down from a rate of 20% or so in the past decade (see chart 1). A slowdown last year offered a glimpse of the future, spooking investors. Nor can Google easily capture a bigger slice of the slower-growing pie. Trustbusters already believe its share is too high and have sued Google in America for abusing its search monopoly. Google’s arrangement with Apple, whereby it pays a reported $15bn a year to be the default search engine on the 2bn or so iDevices, has also come under scrutiny. Although search remains immensely lucrative, with operating margins of nearly 50%, according to Bernstein, a broker, how people look for things on the internet is changing. Most product searches these days start not on Google but on Amazon, the e-commerce giant. According to Google’s own executives, 40% of teenagers and young adults seek recommendations for things like restaurants or hotels on TikTok, a short-video app, or Instagram, a similar app from Meta. Google may entice some of these “search-nevers”, as Mark Shmulik of Bernstein calls them, to its platform, as YouTube is doing already with a TikTok lookalike called Shorts. Yet videos are unlikely to monetise as nicely as the search box. Then there are the chatbots and other “generative” AIs that, having been trained on a web’s worth of texts, images and sounds, can serve up simulacra of human-generated content. Mr Pichai’s insistence that Alphabet is an “AI-native” company rings true. Most observers believe that deep pockets and ample talent will allow Google to solve the technology’s teething problems, such as the bots’ tendency to “hallucinate” (make stuff up) or the high cost of serving up responses (which egg-headed Googlers are busy tackling). That still leaves open the question of how much money bot-assisted products, for all their expected ingenuity, will actually make. Set aside search, and Google’s knack for creating extraordinary products is matched by its inability to monetise them. There is no reason to think that its AIs More

MOTORING’S SOUNDTRACK use to be generated by the petrol engine and car radio. Near-silent battery power and snazzy infotainment systems have provided an aural void—and a high-tech way to fill it. Carmakers are giving their electric vehicles’ occupants, and anyone within earshot, an alternative set of sounds to enliven the journey.Listen to this story. Enjoy more audio and podcasts on More

IN THE EARLY 1980s AT&T Corporation, then America’s telecoms monopoly, was the darling of Wall Street. As big tech of the day, it was the mightiest company in the S&P 500, accounting for 5.5% of the blue-chip index’s total market value. Today its largest descendants, AT&T and Verizon, can only dream of their parent’s former glory. The two companies make up less than 0.7% of the index—and falling. Their combined market capitalisation of $250bn is roughly half what it was at the start of 2020; the S&P 500 is up by more than two-fifths since then (see chart). Factors beyond the undynamic duo’s control, such as rising interest rates or the recent discovery of a network of old lead-sheathed telephone cables, which injected uncertainty over their potential liability for the toxic assets, are partly to blame. Yet much of AT&T’s and Verizon’s malaise is of their own making. The two telecoms incumbents are finding just how difficult it is to be a mature firm in a saturated market, especially if you also have a mountain of debt to manage. Their core business of selling mobile and broadband subscriptions is stagnating. On July 25th Verizon said that these revenues grew by less than 1% in the second quarter, year on year. The next day AT&T announced that its equivalent figure rose by just 2.4%. What has long been a cosy industry is turning more competitive. A merger in 2020 between T-Mobile and Sprint created a 5G powerhouse that is offering equally fast mobile connections at lower prices. And on July 26th DISH Network, a satellite-TV firm, unveiled a partnership with one of today’s tech titans, Amazon, to provide mobile services for $25 a month to members of Amazon’s Prime loyalty scheme.Previous big bets have backfired. AT&T’s disastrous $200bn foray into media, including the purchase of Time Warner and DirecTV, has been unwound. But its effects on morale and on the balance-sheet, weighed down by net debt of more than $130bn, continue to be felt. Verizon has been less spendthrift, though it still splurged $53bn in 2021 on 5G spectrum—an investment which was deemed necessary at the time to compete with T-Mobile but which has yet to produce a return, as early hype over 5G has dissipated. Its effort to build a wholesale business, by allowing cable providers such as Comcast and Charter to piggyback on its networks, created new competitors, which are offering bundles of internet and TV at a steep discount.That leaves the two incumbents with few options. One is to protect margins. Both Verizon and AT&T are touting their premium plans. On July 24th Verizon raised the price of its wireless home broadband by $10, to $35 a month. Both companies are also cutting costs, including by shutting down retail outlets, which helped each trim operating expenses by 2.5% in the first half of 2023, year on year. A more radical move would be to follow some European peers, such as TIM in Italy, and spin off their fixed networks. This would raise capital, lower fixed costs and allow management to focus on faster-growing segments such as wireless broadband. Such a deal would, though, be at odds with the industry’s trend towards convergence, whereby cable companies are becoming more like telecom providers, and vice versa, in a battle for consumers. Offering both home and mobile connections, especially in a bundle, makes consumers stickier, reduces churn and increases long-term profitability. That at least is the thinking. To investors, it seems increasingly wishful. ■To stay on top of the biggest stories in business and technology, sign up to the Bottom Line, our weekly subscriber-only newsletter. More

Power is a fact of corporate life. It also affects behaviour. Research suggests power makes people less likely to take the advice of others, even if those others are experts in their fields. It makes them more likely to gratify their physical needs. In a test conducted by Ana Guinote of University College London, powerful people were likelier than less powerful folk to choose tempting food, like chocolate, and ignore worthier snacks like radishes. In conversations, the powerful are bewitched by themselves: they rate their own stories as more inspiring than interlocutors’. Listen to this story. Enjoy more audio and podcasts on More

MARK ZUCKERBERG dubbed 2023 as Meta’s “year of efficiency”, corporate-speak for admitting that his social-media empire was bloated. Since November Meta has cut 21,000 jobs, or about a quarter of its workforce. Bosses of its fellow tech titans have also embraced the efficiency mantra. Alphabet (Google’s corporate parent), Amazon and Microsoft have collectively shed more than 50,000 jobs since October. As big tech reports its earnings this week expect more talk of “re-engineering the cost base”. The bloodletting (in plain English) is not limited to the giants. According to layoffs.fyi, a website that tracks sackings, nearly 900 technology companies around the world have announced total job cuts of more than 220,000 in 2023.The slump has hit younger firms hardest of all. Rising interest rates make upstarts’ promise of rich profits far in the future look less juicy in the here and now. As a result, venture capitalists are stinting. Globally, venture-capital investment in the first half of this year was $144bn, less than half of the $293bn raised by startups in the same period in 2022. Companies that do manage to raise funds are seeing their valuations squeezed. According to Carta, an equity platform for startups, in the first quarter of 2023 almost a fifth of all venture deals were “down rounds”, where companies raise capital at a lower valuation than before. The valuation of Stripe, a fintech star, fell from $95bn to $50bn after its latest funding round in March.That is forcing aspiring Alphabets and Metas to follow their role models in rethinking some of the habits acquired during the years of easy money. Efficiency is the talk of Silicon Valley. Companies accustomed to spending with abandon to win market share are finding themselves in the unfamiliar position of having to trim fat. And there is plenty of fat to trim. A good place to start is payroll. Battle-hardened founders grumble that salaries are the biggest expense for young firms. In July startup job postings on Hacker News, a news site for coders, were down by 40% compared with the same month last year (see chart 1). The average startup is already looking leaner. Numbers from CB Insights, a data provider, show that the median number of employees at young firms has been steadily declining. In 2018 the typical firm that raised a total of between $10m and $25m had around 50 employees. In 2023 a similar one would employ 41. It is a similar story for larger startups, all the way to late-stage firms which have raised more than $500m (see chart 2). In the go-go years firms hired lots of people who did not have that much to do. Not anymore. Most startups, points out Tom Tunguz, a venture capitalist, can run with smaller teams, with a negligible impact on revenues. Tech firms are, naturally, embracing artificial intelligence (AI). An AI “co-pilot” on GitHub, a Microsoft-owned platform for open-source programs, improves coders’ productivity by 30%. And it is not just the geeks who benefit. Other employees use AI-based tools, from chatbots like ChatGPT that churn out emails for marketers to clever software that improves sales efficiency. One founder of an early-stage startup with fewer than ten employees estimates that AI has already boosted his company’s productivity by 30-40%. The austere spirit is visible even among one of the few categories of startup that is unaffected by investors’ newfound stinginess: those which develop all the AI tools. Anthropic, a firm founded by defectors from OpenAI, which created ChatGPT, has raised $1.2bn with 160 employees. Adept, a company started by former employees of DeepMind, an AI lab owned by Alphabet, has raised $415m with 37 employees. Compare that with darlings of the previous startup boom. Klarna, a Swedish payments firm that experienced wild growth in the go-go years, had 2,700 employees by the time it raised $1.2bn. Databricks, a database-maker, had a staff of 1,700 at a similar stage. ■To stay on top of the biggest stories in business and technology, sign up to the Bottom Line, our weekly subscriber-only newsletter. More

When Amazon announced the $13.7bn acquisition of Whole Foods Market in 2017, it followed some oddball attempts to strengthen its grocery business, some conceived by Jeff Bezos himself. One was to develop an “ice-cream truck for adults”, driving into neighbourhoods with lights flashing and horns honking, to sell porterhouse steaks, Shigoku oysters, Nintendo games and other goodies. It was quietly shelved. Another was to create a product so unique that only Amazon could supply it. The answer was the “single-cow burger”, a Wagyu beef patty made from the meat of one animal. You can still find them on its website—though they are now permanently out of stock.Amazon’s purchase of Whole Foods signalled it would take a more conventional approach to the supermarket business. That is probably why, when the deal was announced, Amazon’s share price soared and those of its rivals, such as Walmart, fell. But since then Amazon has treated grocery more like a science experiment than an exercise in seduction, with weak results at Whole Foods and in other formats. Its best-known addition to the retail experience is the “just walk out” technology in physical stores, equivalent to its one-click shopping online. Yet cashierless supermarkets sound like something more beloved of geeks than grocers. What may cut down on time-wasting queues also minimises what some people love about shopping: the human interaction at the till, the hunter-gatherer instinct as they jostle at the meat counter, the Columbian exchange between fellow foodies at the spice rack. Amazon is trying to refresh the experience. Last year it recruited Tony Hoggett, a former executive from Tesco, a British supermarket chain, to bring grocery nous to a business hitherto obsessed with overhead cameras, QR codes and data collection. The Brit, who started out as a Tesco “trolley boy” aged 16, has a big job. When Schumpeter visited an Amazon Fresh store in Los Angeles recently, the fresh meat and fish counters were so barren they looked like part of a going-out-of-business sale. He bought one of the three rotisserie chickens on display out of sympathy, because he feared they had been there all day. Under Mr Hoggett, Amazon is trying to make the Fresh stores less soulless. Human cashiers and self-checkouts are back for those who prefer them. Whole Foods’ expertise is being used to rethink store location. It is part of an effort to make grocery shopping as habitual on Amazon as it is at a Walmart. Andy Jassy, the CEO, says it is aiming to build a “mass grocery format” commensurate with Amazon’s size. Yet if anything Walmart looks more likely to invade Amazon’s territory than the other way around.Neither firm thinks of itself as competing head-to-head with the other. But they are, because both have big growth ambitions. For Walmart, that means expanding its e-business beyond grocery into general merchandise, as well as attracting higher-income online customers. Both of these are Amazon’s forte. For Amazon, it means a bigger presence in grocery, both online—where food shopping still accounts for only about 10% of America’s $800bn supermarket business—and offline. In bricks-and-mortar, Walmart’s lead is huge. It has the largest footprint in America, with about 4,700 outlets, compared with 530 Whole Foods, 44 Amazon Fresh and 22 Amazon Go shops. Grocery accounts for most of its sales, whereas for Amazon they are a sliver. Its “everyday low prices” work: a survey by MoffettNathanson, a research firm, found equivalent products at Amazon Fresh were far pricier. Walmart’s speed of delivery matches Amazon’s. What Amazon lacks in stores, it hopes to make up for in membership of its Prime loyalty programme, which is estimated at 170m in America, compared with about 22m for Walmart+. Eventually, it hopes that online grocery shopping, combined with three different formats—Whole Foods for posh nosh, Fresh for general grub and Go for grab-and-go—will enable customers to buy everything they need via a single app. Amazon has two other advantages: a whopping marketplace platform for third-party sellers, which adds to the range of products available on its website, and an advertising business with a hefty $38bn of revenues last year, which supplements its supermarket business. Yet because shoppers like to see, feel and smell their groceries before buying them, the scarcity of stores is a problem. Dean Rosenblum of Bernstein, a broker, calculates that Amazon Fresh is accessible to just over a third of Americans. In contrast, 90% of them live within ten miles (16km) of a Walmart. If Amazon opened 50 new Fresh stores a year, in a decade’s time it would reach only the size of Whole Foods’ current tally. And that would be a “near criminally irresponsible use of Amazon capital”, Mr Rosenblum says. That view is spreading. Terry Smith, a British fund manager, recently dumped his Amazon stock, arguing that its move into grocery retail risked misallocating capital.Moreover, Walmart appears to be making more headway with online selling and advertising than Amazon is with its real-world stores. Walmart’s online sales were estimated at $82bn last fiscal year, more than four times Amazon’s physical-store sales. It appears to be borrowing Amazon’s model of attracting third-party sellers to its site in order to increase the assortment of products, raise logistics revenue and boost advertising. Last year Walmart’s ad sales grew by 30%, to $2.7bn. That is still a fraction of Amazon’s total. But there is no reason why Walmart should not catch up, thinks Simeon Gutman of Morgan Stanley, an investment bank.One-click M&AAmazon could leap up the league table by buying a large supermarket chain. Given the antitrust pressure on big tech, though, this is probably off the table. If a build-rather-than-buy approach is its only option, it will have to do a much better job of explaining how it will make it profitable. As it continues to waste time experimenting, Walmart is ably copying its best moves. ■ More

IN 2019, AS China’s trade war with America was heating up, the People’s Daily predicted that China’s monopoly on rare earths, minerals crucial to the production of most modern hardware, would become a tool to counter American pressure. “Don’t say we didn’t warn you,” the Communist Party mouthpiece thundered. For years the bluster was just that. Between 2009 and 2020 the number of Chinese export controls on the books ballooned nine-fold, according to the OECD, a club of mostly rich countries. Yet these restrictions were haphazard, informal and aimed at narrow targets—random warning shots rather than a strategic offensive.As America ratchets up its sanctions against China, which among other things make it impossible for Western chip companies to sell Chinese customers cutting-edge semiconductors and the machines to make them, new volleys from Beijing are coming thick and fast. Earlier this month, after China announced its latest export controls, this time on a pair of metals used in chips and other advanced tech, a former commerce ministry official declared that the measures were “just the beginning” of Chinese retaliation. On July 20th Xie Feng, China’s new ambassador to America, said that his country “cannot remain silent” in the escalating war over technology. A response, he hinted, was coming.This time it looks much more deliberate. To counter America’s effort to contain China’s technological ambitions, Xi Jinping, China’s paramount leader, has called on regulators to fight back against Western coercion in what he has called an “international legal struggle”. The result is a flurry of lawmaking that is creating a framework for a more robust Chinese reaction to America’s commercial warfare.The list of recent laws is long. An “unreliable entities” list, created in 2020, punishes any company undermining China’s interests. An export-control law from the same year created a legal basis for an export-licensing regime. In 2021 year an anti-sanctions law enabled retaliation against organisations and individuals who carried out the sanctions of other countries. A sweeping foreign-relations law enacted this year, and prompted by Western sanctions against Russia over its invasion of Ukraine, permits countermeasures against a wide range of economic and national-security threats facing the country. It came into effect on July 1st. The same day an anti-espionage statute came into force, extending the reach of Chinese security agencies. All the while, China has tightened various data and cyber-security rules.The new rules are already being used, as opposed to merely brandished. In February Lockheed Martin and a unit of Raytheon, two American armsmakers with non-weapons business in China, were placed on the unreliable-entities list after shipping weapons to Taiwan (which China regards as part of its territory). The companies are blocked from making new investments in China and from trade activity, among other restrictions. In April Micron, an American chipmaker, was hit with an investigation by China’s cyberspace regulator, based on a new cybersecurity law. After Micron failed a security review, regulators banned its chips from critical infrastructure. The laws’ vague wording makes it difficult for Western companies to assess their potential impact on their business in China. The “mother of all sanctions laws”, as Henry Gao of Singapore Management University describes the foreign-relations law, vows to hold accountable anyone acting in a manner deemed “detrimental to China’s national interests…in the course of engaging in international exchanges”. Several foreign law firms in China have been asked by their Western clients to evaluate the risks of being hit by investigations. One lawyer looking into potential Chinese cyber-probes notes that American tech companies producing commodified hardware components, such as Micron’s memory chips, should be on guard for sudden investigations. China’s new laws allowing the government to restrict a broad range of minerals and components, meanwhile, are injecting similar uncertainty into the businesses of their foreign buyers. One affected group, notes David Oxely of Capital Economics, a research firm, is Western manufacturers of green-energy technologies. Battery-makers, in particular, are highly dependent on China across the supply chain (see chart). Last year the commerce ministry proposed a ban on exports of ingot-casting technology used in making solar-panel wafers. If imposed, such a prohibition could hold back the development of indigenous solar-power technology in the West, which would hurt Western manufacturers while increasing foreign demand for finished Chinese solar panels. The restrictions on the two chip metals, gallium and germanium, could pose a strategic headache to America. The rules, which come into force on August 1st, require exporters to apply for licences to sell the metals to foreign customers. China produces 98% of the world’s raw gallium, a key ingredient in advanced military technology. This includes America’s next-generation missile-defence and radar systems. A shock to the supply of gallium could cause long-term problems for the American defence industry, reckons CSIS, a think-tank in Washington. Moreover, a gallium-based compound, gallium nitride, may underpin a new generation of high-performance semiconductors. Keeping it out of foreign hands would stymie Western efforts to develop the technology while furthering Mr Xi’s policy goal for China to control it. China needs to tread carefully. The country reimports many of the finished products that are made abroad using rare earths, notes Peter Arkell of the Global Mining Association of China, a lobby group, so prohibitions could come back to bite Chinese firms. Outright export bans would also prompt the West to build its own relevant production capacity and seek substitutes, observes Ewa Manthey of ING, a Dutch bank. This would in the longer term weaken China’s hand. And labelling as unreliable entities big Western firms with large Chinese operations could jeopardise thousands of Chinese jobs. That may explain why rather than blacklisting all of Raytheon, whose aviation subsidiary, Pratt & Whitney, employs 2,000 people in China, the commerce ministry limited its ban to the American company’s defence unit.So far the relatively pragmatic ministries of commerce and foreign affairs have led the implementation of the various laws. One fear among Western businesses is that more hardline agencies supplant them. If the tech war escalates further, China’s National Security Commission, chaired by Mr Xi himself, may take the lead, fears Mr Gao. If that happens, concerns about potential blowback for Chinese commerce are likely to carry less weight. The consequences are scary to contemplate—and not just for Chinese and American CEOs. ■ More

IT LOOKS LIKE cheese. It smells like cheese. It tastes like cheese (specifically mature cheddar). And it is cheese—at least under the microscope. “Synthetic dairy” is made with the same ingredients as the conventional sort. But instead of getting the main ingredient from a live ruminant, Better Dairy, a three-year-old British cheesemaker, derives some of it from yeast. These microbes are fed sugar, which they then convert into milk proteins in a process which is similar to brewing. Listen to this story. Enjoy more audio and podcasts on More