What the AI bubble means in 2026

22nd January, 2026

In this blog, Jason Whittaker (author of You Want What We’ve Got) ponders the meaning of AI on the economy and society in 2026.

***

When I began writing You Want What We’ve Got: Big Tech v. Big Journalism, ChatGPT had not even launched. Automated journalism was the preserve of a few select areas, such as financial reporting. By the time that I completed the book I, like many others, was convinced we were witnessing a tech bubble that could rival the dotcom peak of the early 2000s.

While the final numbers from 2025 have yet to be confirmed, investment in AI by American tech companies is estimated at more than $350 billion, up from $104 billion in 2024. These huge investments are starting to extract a heavy price, from well-known corporations such as the data giant Oracle, which has seen its shares slide by more than 40 percent, to relative unknowns such as the cloud-computing company, CoreWeave, which lost a staggering $33 billion in six weeks because of its failure to deliver new data centres.

Will the bubble burst in 2026? My short answer is: no. My slightly longer answer is that it will begin to deflate as investors ponder when – if – they will ever see a return on the vast amounts of money being poured into concrete and chips. While the current AI bubble has many similarities with the dotcom boom and bust a quarter of a century ago, there are some important differences, not least that some investors were paying more than 250 times the revenue of some tech stocks during the late 1990s, while with a few exceptions it is closer to forty times as much today.

It’s also true, however, that the stakes are much higher. In the early 2000s, the US stock market was worth roughly twice as much as Europe’s; today it is three times as much. Much of that value is locked up in a handful of companies, such as NVIDIA whose value bounces around $4.5 trillion, followed closely by Apple, Alphabet, and Microsoft, all bumping about the $4 trillion mark. Big Tech has never had it so good or, in other words, Big Tech has become too big to fail. The ‘magnificent seven’ – NVIDIA, Apple, Alphabet, Microsoft, Amazon, Meta and Tesla – account for a third of US stock market value. The world has never been richer, but nor has so much wealth been dependent on so few companies.

In the very short term, the magnificent seven will flourish but this doesn’t mean that the giants of today will continue to dominate tomorrow. Trump’s trade tariffs and his economic war with China (not to mention the threats of an EU ‘trade bazooka’ as he lays claim to Greenland), have shown Apple to have feet made more of clay than the tungsten and aluminium that encases their iPhones. Meanwhile, the company that kickstarted the generative AI revolution – OpenAI – is flying, Icarus-like, too close to the sun, with some experts predicting it will soon burn through $45 billion a year to train and run the ChatGPTs of the future.

But what does this mean to the rest of us? I began writing about technology, first as a journalist and then as an academic, in the early 1990s when the internet barely registered in public consciousness. Of the four waves of digital revolutions that I’ve witnessed – personal computing, the internet, smartphones and now artificial intelligence – I am increasingly convinced that it is the latter of these that will be the most transformative, for better or for worse. This in part is because it stands on the shoulders of the preceding developments: if the rise of the personal computer in the seventies and eighties democratised computer power, it was still too expensive for a long time to influence most people; the internet was a key driver in globalisation at the turn of the century, while we became used to incredibly powerful personal devices and even greater connection after Steve Jobs debuted the first iPhone in 2007.

Each wave of innovation is usually invoked as a sign of the inevitable determinism of technology, what was mockingly referred to in the nineties as ‘west-coast fascism’: you will assimilate – and have a nice day! That ignores the tendency of humans to innovate in the ways that the inventors of technology don’t expect, whether Roentgen’s accidental discovery of the X-ray or Android being developed as a system to power digital cameras before pivoting to mobile phones.

Bill Gates pointed out that AI represents a fundamentally different way of interacting with digital technology. Rather than having to learn the computer’s logic, whether a specific programming language or a user interface, innovations such as natural language processing and multimodal interaction (using voice, gestures and touch) mean that computers are better at understanding our contexts and intent. We speak, and they – apparently – understand.

And yet, to paraphrase George Orwell’s Animal Farm, if all human intentions are equal, some are more equal than others. Those eyewatering numbers at the beginning of this article, the billions of dollars poured into artificial intelligence that will soon roll over into trillions, need to be justified somehow. In AI circles, rather than a bubble many often talk about a race – a race between the tech companies and between the USA and China to dominate the next huge industry of the twenty-first century. Recently, the World Economic Forum published its global predictions for the impact of AI on the world of work: four scenarios, in which only one – what they call the co-pilot economy – provides gradual shift towards automation that augments human abilities rather than displacing them entirely. At present, Big Tech is chasing what the WEF calls ‘supercharged progress’, exponential breakthroughs that transform everything while social safety nets and governmental frameworks struggle to keep up. Accelerate into the future is everything, and everyone else be damned.

Because of the very obvious threat it poses to jobs and livelihoods, many people fear the future that AI promises – not least when the USA appears to be throwing caution to the wind on so many fronts. But whether AI accelerates or stagnates in the coming year, just like the internet and mobile technologies it is not going away. The WEF proposes one approach to artificial intelligence that means you do not have to be at the mercy of Big Tech, which is worth noting in conclusion: start small, run controlled experiments, test what new technologies can actually do rather than what the companies claim, learn from failure, and build only on what works.

– Jason Whittaker

Jason Whittaker is Professor of Communications at the University of Lincoln and previously worked as a journalist for IDG Media. His books include Tech Giants, Artificial Intelligence and the Future of Journalism (2019) and Divine Images (Reaktion, 2021).

You can buy You Want What We’ve Got: Big Tech v. Big Journalism through our online shop.