”Börsen prisar in AI-kaos – historien ger en annan bild”

The experience of past tech revolutions suggests savvy incumbents might muddle through and even thrive.

For a sense of how the stock market has been struggling to understand the disruptive threats posed by AI, consider the bout of selling set off by Florida-based Algorhythm earlier this year.

A karaoke machine maker that pivoted into AI in 2024, Algorhythm’s $2mn market cap made it the least valuable company on Nasdaq in January, according to Brendan Hopkins, who handles its investor relations.

But when its chief executive, Gary Atkinson, declared in a February press release that its AI software could eradicate the wasteful trips that account for 30 per cent of the global trucking industry’s journeys, US transport stocks plunged.

The calculation was only meant as an illustration and the idea that the world’s truckers would rush to hand their most valuable data to a tiny AI start-up was preposterous, Atkinson says now. But as a case of extreme market jitters, he adds, it was instructive: “Everyone’s looking over their shoulder: who’s about to get disrupted by AI?”

Fear that the technology behind ChatGPT is about to cause upheaval in a wide range of industrial sectors has spread quickly this year — even if investors have been reacting mainly to theoretical threats, including some as far-fetched as that posed by Algorhythm. The impact has been felt most acutely in the software world, where worries have been mounting for some time as AI transforms how code is written.

But 2026 is shaping up to be the year in which those concerns spill out more broadly. One catalyst was the news that AI company Anthropic had repurposed its software-coding agent to take on a wide range of white-collar work, which wiped $830bn off software stocks in the course of a single week.

Suddenly, if the AI boosters are to be believed, AI might crop up almost anywhere and take on work that had previously required a human. Some trigger-happy investors have responded by selling at the drop of a press release.

Another such bout of turbulence came after Altruist, a private company that sells services to investment advisers, announced in February it had launched an AI-powered tax planning tool.

Altruist chief executive Jason Wenk was on a break in Mexico and had just finished an early morning run on the beach when his phone lit up with texts. “At first, I thought this must be some type of cruel joke,” he says.

But, as he watched stock plummet in wealth management companies such as Charles Schwab, the reality sank in: “I was, like, holy cow. The market really is reacting quite aggressively.”

The wrenching reaction was symptomatic of wider concerns that AI may be about to automate entire services and classes of work that provide steady profits for companies and healthy incomes for their workers. “It’s people realising that, collectively, a lot of the things that you used to have to pay a lot of money for, you will no longer have to pay a lot of money for,” says Wenk.

Just how widely and in what form the impact might be felt are questions many business executives, as much as investors, are struggling to understand. Generative AI, a general-purpose technology like PCs or the internet, can theoretically be applied to a wide range of problems.

Most obviously, it could replace human brainpower in fields like legal research. But the effects are likely to be felt much further afield, in any industry where the ability to collect and manipulate large amounts of data could bring a competitive advantage.

Amazon founder Jeff Bezos, for instance, is raising a large investment fund to buy manufacturing companies, in the hope that he can retool them with technology from his AI company Project Prometheus and steal a march over other manufacturers that are slower to adapt.

In the tech world, all the talk of disruption has revived memories of the 1990s, when start-ups riding the first wave of the internet were seen as an imminent threat to many established businesses and the fear of being “dotcommed” was pervasive.

Generative AI, by its nature, might be even more disruptive. But the experience of the earlier digital revolution could still be instructive.

Many of the nimble first movers of that era, untethered by the baggage of existing ways of working, ran out of money before the wider market adopted their innovations, while a tiny number went on to become the biggest corporations in history. Many of the epitaphs written for the supposed dinosaurs of the day proved premature.

For all the frantic market moves of the past couple of months, the winners and losers of this era of tech disruption might be just as hard to predict.

Rebecca Lynn, who in the late 1990s was an executive at NextCard, the first online credit card company, describes the emergence of the internet as “a new distribution channel” that enabled start-ups to break into markets they were previously shut out of.

But AI “has given us a way to solve problems that we have wanted to solve for a long time, and the tech wasn’t there”, adds Lynn, who is now managing director of Silicon Valley venture capital firm Canvas Prime.

OpenAI CEO Sam Altman, for instance, sent a shockwave around the advertising world in 2024 when he said that “95 per cent of what marketers use agencies, strategists and creative professionals for today will easily, nearly instantly, and almost at no cost be handled by AI”.

Just because a new technology could theoretically take on many of the jobs currently done by people, however, doesn’t mean that it will. The real world has a stubborn way of resisting change, whether because vested interests and cultural habits make market structures slow to adapt, because regulation stands in the way, or because the systems needed to support broader change across an economy take time to develop.

The dotcom bubble quickly popped when most of the start-ups ran out of cash and into the reality that there wasn’t yet much of a market for their services. “There clearly was less disruption than was anticipated and it took longer for the disruption to take place,” says Bill Janeway, a vice-chair of investment firm Warburg Pincus during the dotcom boom.

It was easy then to imagine how a new capability, like ecommerce, might rise up overnight to destroy existing retailers, he points out, but online commerce hit a wall when it came to being able to physically handle high volumes of transactions.

It took the development of many new, complementary technologies and services before ecommerce could live up to its initial promise. “I am very sceptical about the generic [AI] solution to everything,” Janeway adds.

Similarly, trepidation about how well AI can navigate the open-ended, unstructured environments in which most white-collar work takes place could seriously slow its adoption.

Although AI companies say the coding agents now used widely to produce software prove the technology is ready for wider use, Chris Bradley at McKinsey Global Institute argues that general-purpose agents are more like the driverless taxis operated by Waymo.

These have finally taken to the streets in a small number of cities, but only after many years of training and refinement, and even now they still can’t deal with all the situations that a human driver can master.

Plenty of obstacles remain for AI agents. The continuing problem of hallucinations, or displays of misplaced certainty that accompany an incorrect response, have prompted companies to build guardrails around the technology. 

Altruist’s tax software, for instance, uses AI agents to research a client’s tax situation, before handing over any actual tax calculations to more traditional, deterministic software.

“We’re not letting LLMs do math,” says Wenk. “The end result is that [AI] can’t hallucinate on data.” That doesn’t mean, he adds, that the agents are infallible: “Could it miss some detail in the deep research process? That’s always conceivable.”

That hasn’t stopped the headlong race to turn generative AI into a strategic business weapon, some aspects of which are likely to strongly favour new disrupters over large incumbents.

One is the ability to run a business with a radically smaller headcount than was possible before. Michael Moritz, a former chair of venture capital firm Sequoia, says “it will take far fewer people to start and build companies” in future, handing an advantage to new businesses starting out with a blank sheet of paper (though he believes the often-quoted idea that AI will give rise to one-person companies is “wanton hyperbole”).

For incumbents, the risks of rethinking entire business models and undermining their current sources of revenue act as powerful incentives to stick with existing ways of making money.

Many of those are tied directly to the use of human labour, from the per-hour fees of professional services firms to the per-seat subscriptions of software companies and the “full-time equivalents” (FTEs) that advertising agencies use to calculate their fees. That could put them in the line of fire as new, automated services come to the fore.

Bradley, at McKinsey, says many large company bosses are looking at AI too narrowly, considering only how it could be used to replace existing jobs and how it might change some internal work processes. They should instead be taking a more expansive view, he adds, seeing the technology as the equivalent of “a data centre full of geniuses” that could help reimagine how they deliver value to customers.

Lynn, the venture capitalist, points to the example of legal research firm Casetext, which, after getting early sight of OpenAI’s GPT-4 chatbot, “burnt its entire product roadmap down” to refocus its AI on writing legal briefs. The decision paved the way for its $650mn sale to Thomson Reuters in 2023.

“It’s hard enough for a venture-funded company to do that, but it’s really impossible for a huge company,” she adds. Bigger companies tend to limit themselves to “adding layers” with new technology and “optimising around the edges”, rather than undertaking a root-and-branch overhaul of their businesses.

But previous waves of disruption, like the rise of the internet, have usually left more room for the incumbents than might have seemed likely in the white heat of technological revolution. Many markets don’t change as fast as the disrupters predict, giving incumbents time to adapt — sometimes by buying disruptive upstarts.

Walmart, for instance, was long seen as the chief victim-in-waiting of the rise of Amazon. But Walmart is still very much here, following years of investment to rebuild its business around ecommerce — including $3.3bn spent buying online retailer Jet.com in 2016. After doubling over the past two years, its market value recently climbed above $1tn for the first time.

Altman’s prediction that AI would quickly make advertising agencies irrelevant is dismissed as too simplistic by some industry insiders, even as they agree that it will transform things like the way advertising content is generated and campaigns are planned.

Advertising agencies have for years tried to shift their customers away from people-centric charging models like FTEs that are vulnerable to AI disruption, and towards subscription and software-like revenue models, says Brian Wieser, principal at Madison and Wall, a media and marketing advisory firm. It is their customers who have resisted such changes.

He adds that media buying, which has become the main bread and butter of the agencies, has already been reshaped by years of AI adoption. “This is the world we’ve already been living in — it’s progressively more AI-based,” says Wieser. If they continue to invest and adapt, many ad agencies should be able to keep riding this wave of change, he adds.

It is the companies that fail to adapt that are likely to fall by the wayside, as both rivals and customers wielding AI efficiencies wipe out their previous sources of profit. “There are so many parts of our economy where companies make money because consumers don’t know any better, or they’re lazy, or the infrastructure doesn’t allow them to do better,” says Wenk.

Many of these protected, profitable niches are likely to disappear as AI agents start to scour the digital world relentlessly on behalf of consumers, he says. “AI will not allow people to hide.”

Chris Varelas, co-founder of Silicon Valley investment firm Riverside Capital, says that same force will act as a brutal leveller across the business world. “It’s going to make the smart people smarter,” he says. “People who aren’t able to create a differentiated product are going to be the casualties.”

But if the lessons of previous periods of wrenching change are anything to go by, much of today’s business establishment will muddle through, even as a handful of new AI giants rises to become the most visible winners from the technology.

Economic history has shown that “most industrial change happens by very large new companies doing big things,” says Bradley. “And then the rest of the economy kind of just carries on.” Behind that “just carrying on”, there will be no shortage of upheaval, churn and disruption.

©The Financial Times Limited 2026. All Rights Reserved. FT and Financial Times are trademarks of the Financial Times Ltd. Not to be redistributed, copied or modified in any way.