Having lost ground to its main rival, OpenAI has lately been picking up the pace. Last week it told investors that revenue from its enterprise sales doubled in the third quarter compared with the preceding three months, helped by price discounts intended to undercut Anthropic. For the pair, however, enterprise spending remains concentrated. Revenue from agents currently adheres to an “extreme power law”, says one Anthropic investor. The biggest customers contribute an unusually high share of sales, and the heaviest users within those companies—who are nearly always software engineers—are behind much of the spending. That poses a risk if these users switch to cheaper models, as many have begun to do. But it also suggests that there is still plenty of scope to extend adoption. This year Anthropic and OpenAI have both rolled out general- purpose enterprise agents that non-coders can use to perform a wide variety of tasks. And in addition to working with the consulting giants, the model- makers have set up their own independent advisory businesses, with private- equity backing, to help customers deploy their technology. A person familiar with Anthropic’s Ode, which is supported by Blackstone and Hellman & Friedman, says that more than half of the clients it has helped to redesign their processes have boosted their profits by at least $10m. Last month Anthropic established a separate consulting venture with Hg, another private-equity firm, focused on Europe. Aaron Levie, chief executive of Box, a cloud-storage company that works with both AI labs, says that having “the right kind of product experiences with the right sales efforts” could prove even more important than having the best models. Even so, the models will still matter—and here Anthropic holds a clear lead. Investors in both firms reckon performance on “long- horizon” tasks that take several hours or even days—such as building an app from scratch—will become increasingly important for the frontier labs as businesses make greater use of less capable models for simpler work to keep costs down. On MirrorCode, one long-horizon benchmark, OpenAI’s leading model, released last month, is still outperformed by the top offering from Anthropic as far back as July. That may reflect Anthropic’s more specialised research programme. Whereas OpenAI has sought to remain at the frontier across a wide range of
capabilities, including image and voice, Anthropic has focused on the ability of its models to write code and interact with other software. That, in turn, highlights a difference in the strategies of the two. Although OpenAI has sought to regain ground in enterprise applications, the consumer segment remains a priority. ChatGPT has 1.2bn weekly users worldwide, almost 100m of whom pay upwards of $8 a month for a subscription. Even those who don’t pay provide OpenAI with revenue via the advertising business it launched in February, which by August had an annualised revenue of $1bn. By contrast, Anthropic’s IPO prospectus barely mentions consumers. It has sworn off advertising, worrying it could muddy Claude’s ability to give helpful answers and incentivise the company to build addictive products. That leaves OpenAI with a bigger addressable market. But its approach has drawbacks, too. In the consumer segment it must compete with Google, whose Gemini chatbot has gained ground over the past year, as well as personal agents such as Instinct and Meta’s Muse, which have beguiled users with their ability to take on tasks from arranging a dinner with friends to securing sold-out concert tickets. OpenAI has deprioritised some efforts, shutting down its video-generation tool, Sora, in March. But it continues to pursue a strikingly broad set of opportunities. In the coming year it is expected to unveil a device—a small doughnut-shaped speaker—that consumers can use to interact with its AI systems. It is also developing a humanoid robot. The danger is that it spreads itself too thinly. Unlike Anthropic, which generated its first quarterly operating profit in the second three months of this year, OpenAI remains deep in the red. It will need to follow its rival into the public markets as soon as it can. ■ This article was downloaded by zlibrary from https://www.economist.com/business/2026/11/15/anthropics-giant-ipo-puts-the-heat- on-openai
Nvidia’s roaring results can’t hide its troubles Jensen Huang’s departure will cause lasting damage Nov 19th 2026 In certain respects it was like any Nvidia earnings call over the past few years. The chipmaker reported yet another quarter of blistering growth, notching up revenue of $116bn for the three months to October, more than double the figure from a year before. Operating profit soared at an even faster rate. And yet the call on November 17th was unlike any in Nvidia’s history—for it was the first without Jensen Huang, who until a few days before had been the only person to run the company since it was founded three decades ago. Mr Huang’s resignation as chief executive late on November 11th, following revelations of affairs with a suspected Chinese agent and at least one subordinate, shook the company and its investors. Analysts dialling in were barred from inquiring about the circumstances surrounding his departure,
and Harvey Jones, Nvidia’s newly appointed chair, said little about the process to choose a new boss (other than that the board had no intention of allowing Mr Huang to stage the comeback he is seeking). Instead, Colette Kress, Nvidia’s finance chief and interim boss, stuck largely to the numbers. That will not be enough to dispel worries among investors that Nvidia’s outlook is less rosy without its superstar founder in charge. Although the firm’s share price perked up briefly after the call, it soon retreated, and is down by more than a tenth since Mr Huang’s departure. He played an outsize role in every facet of Nvidia—from its technical wizardry and carefully orchestrated supply chain to its charm offensive in Washington. How diminished will the company be without him? Even before Mr Huang’s exit, the threats to Nvidia were mounting. Since the release of ChatGPT in late 2022 it has generated $300bn in free cashflow thanks to its AI chips, and seen its market value increase ten times over. But its share price has lagged behind the semiconductor industry since the start of this year (see chart 1). Rivals have been closing in. In simulations by Epoch AI, a research group, the newest chip designed by AMD outperforms Nvidia’s best when measured by the number of simultaneous requests they can handle (see chart 2). Google, a big Nvidia customer, has also developed high-performing AI processors, and is peddling these to others.
The dominance of Nvidia’s software layer, known as CUDA, has been eroding, too. It is widely used by model-makers and operates only on Nvidia’s silicon, which has helped keep customers locked in. Yet Anthropic and OpenAI have lately been making greater use of chips from AMD, Google and others, helped by the very models the labs are developing, which make it easier to rewrite chunks of code to run on other chips. Mr Huang’s departure will accentuate Nvidia’s problems. For now, the company has plenty of momentum. Nvidia’s order backlog probably stretches into 2028, and on its earnings call the company raised its revenue forecast for the year ahead. It has secured hundreds of billions of dollars’ worth of output from its suppliers, including for scarce high-bandwidth memory. And its engineers already have technical roadmaps in place for the next few versions of its chips. Even so, Nvidia faces immediate challenges. One source of trouble could come from within the company itself. It is unclear how long it will take the board to appoint a successor. Whoever takes over is almost certain to restructure the top team; Mr Huang had some 60 executives reporting to him directly. As a result, the company risks an extended period of in-fighting and talent flight. A senior employee has already written to the board to warn of the loss of “irreplaceable talent” if Mr Huang is not reinstated.
Geopolitics will be another source of immediate trouble. America has long tried to limit the number of cutting-edge chips that end up in Chinese hands. In light of the spying scandal that has engulfed Mr Huang, export controls look set to tighten further. On November 16th the Commerce Department announced new rules requiring data-centre operators in a number of countries, including Singapore and Malaysia, to obtain a licence before buying advanced AI chips, in order to prevent remote access being sold to Chinese companies. America’s legislature is considering a handful of laws aimed at further controlling chip sales abroad. Timothy Arcuri of UBS, a bank, reckons that these kinds of rules could endanger a tenth of Nvidia’s revenue. And a merry-go-round of investigations into whether the company has ignored or even aided chip smuggling (which it denies) will add to the distraction. Other threats to Nvidia from the departure of Mr Huang will take longer to bite. His personal relationships were vital in helping the company forge its supply chain. Nvidia is particularly dependent on TSMC, a chip manufacturer from Mr Huang’s native Taiwan. The ties he cultivated with it over decades have recently played a decisive role in ensuring Nvidia secures the supply it needs amid a shortage of production capacity for cutting-edge chips. The erstwhile boss also seems to have nabbed better deals than his rivals. SemiAnalysis, a research firm, estimates that Nvidia buys high- bandwidth memory from SK Hynix, a South Korean supplier, at a roughly 20% discount compared with rivals such as AMD. Such coups have boosted Nvidia’s margins, and may be hard for his replacement to replicate. Many of Nvidia’s customers may also come to struggle as a result of his departure. With the cloud-computing giants increasingly turning to custom silicon, Mr Huang had been fostering an alternative ecosystem of “neoclouds” to diversify its customer base. Nvidia has invested in them directly, and offered other forms of financial backing, such as guaranteeing that it will buy their excess capacity if demand falls short. By one analyst’s estimate, neoclouds may make up as much as a quarter of Nvidia’s revenue. Many expect that Ms Kress will be more cautious in bankrolling them. The share prices of five listed neoclouds have fallen by an average of 20% since Mr Huang’s resignation. Developers of open-source AI, an approach enthusiastically backed by Nvidia’s former boss, will also be feeling nervous.
Nvidia’s leather-jacketed boss was undoubtedly a visionary. He was early to spot the potential for Nvidia’s chips to be used in AI, and for years he invested in CUDA despite mounting losses and grumblings from Wall Street. The company he built, and many others that rely on it, will feel his loss. Its rivals will be rubbing their hands with glee. ■ This article was downloaded by zlibrary from https://www.economist.com/business/2026/11/18/nvidias-roaring-results-cant-hide- its-troubles