Sure, let’s give this a bit of a British touch, shall we?
—
In 2025, the esteemed Merriam-Webster dictionary across the pond crowned “slop” as its word of the year. It’s defined as “digital content of low quality that is produced, usually in quantity, by means of artificial intelligence”. This choice highlights a burgeoning concern, reflecting the widespread adoption of AI, especially by corporate bigwigs eager to trim payrolls. Yet, the pitfalls are becoming increasingly apparent. By 2026, AI’s reality check poses a notable economic risk.
One formidable critic, Ed Zitron, argues — in rather colourful terms — that the industry’s economics don’t align. The cost to service a single customer versus the revenues just doesn’t stack up. As he bluntly puts it, it’s akin to “dog poo”. Meanwhile, Cory Doctorow steadfastly insists these AI firms aren’t profitable. In fact, he claims they survive only by burning through astounding sums of someone else’s money.
It’s worth noting that frontier enterprises have historically taken their time to turn a profit. However, profitability usually occurs as expenses diminish. Yet, each version of large language models (LLMs) has only grown costlier, consuming more data and expertise.
These massive datacenters, crucial for training and operating models, are expensive endeavours. They’re often financed through debt, relying on future revenues. Recent scrutiny from Bloomberg indicated $178.5bn in datacenter credit deals in 2025 alone, with novices joining the Wall Street veterans in what can only be described as a “gold rush”.
The chips from Nvidia shining bright in these datacentres have a limited lifespan, potentially outlasting their financial commitments. Besides borrowing, the boom flaunts another bubble-esque trait: intricate financial engineering.
And then there’s the grand narrative. Believers in generative AI argue future revenues will validate the vast investments. They depict a grand transformation. LLMs, it seems, are not just efficient data handlers but are on the brink of “superintelligence”, argues Sam Altman of OpenAI. Mark Zuckerberg suggests they might even replace human friendships.
Yet, they’re certainly replacing some unfortunate employees, as Brian Merchant recounts in “Blood in the Machine”. Many employees, particularly writers and coders, find themselves ousted for AI-produced work of questionable quality.
The potential hazards of rapidly replacing human workers without due diligence have surfaced recently. The British High Court sounded an alarm over AI’s use in legal proceedings, noting cases marred by fictitious legal citations. Similarly, police in Utah found errors in AI-generated transcriptions, once claiming an officer morphed into a frog — amusingly, “The Princess and the Frog” played in the background.
Doctorow argues: “AI’s evolution isn’t the harbinger of ‘impending superintelligence’. Nor will it usher in ‘humanlike intelligence’. Rather, it offers a selection of useful tools which can, on occasion, enhance workers’ lives — but only when used with discretion.”
While these technologies have potential productivity merits, current high valuations and tremendous investments may not be wholly justified. A rethink could disturb financial markets globally. As the BIS recently noted, the “Magnificent Seven” tech firms form 35% of the S&P500, a notable rise from 20% in a mere three years.
A correction in share prices would extend beyond Silicon Valley, impacting retail investors universally. This includes Asian tech exporters and loosely regulated private equity firms that supported the sector’s growth.
In Blighty, the OBR forecasts a potential “global correction” scenario wherein UK and international stock prices might plummet 35% this year. Such a scenario would shave 0.6% off GDP, worsening public finances by £16bn.
Though manageable compared to the 2008 financial debacle, where UK institutions played a starring role, it would still impact an economy struggling to regain its footing. While one might feel a twinge of schadenfreude envisioning tech moguls humbled, we must remember we’re all interwoven within this global tapestry, and none of us would be shielded from the fallout.
