The US stock market, as well as many others around the world, suffered notable losses to round off July. There were varied reasons for this, including continued conflict in Iran and rising oil prices, but AI-focused companies like Nvidia led the sell-off. Indeed, Nvidia relinquished its crown as the most valuable public company in the world to Apple. The latter has, rather famously, not been as heavily invested in AI infrastructure as its rival Big Tech companies, even if it is scrambling to roll out Apple Intelligence and Siri AI.
Nevertheless, it’s not quite time to panic yet. Many AI companies are still up many multiples across the months and years, and many early investors are still in profit. Yet, some claim that the recent drawdowns are mounting evidence of an AI bubble.
Previously, many comparisons were made to the Dot Com Bubble at the turn of the century. Back then, tech stocks soared on the prospect of the new economy powered by the World Wide Web. It reached a fever pitch by March 2000, and then the financial world suddenly came crashing down. However, the new web-based economy did emerge, with winners like Amazon and Google. It was just that there was a lot of collateral damage along the way.
All investment is a gamble. Probably not just as direct as wagering at a casino, but a gamble nonetheless. One of the issues with the Dot Com Bubble is that investors were happy to invest in anything with the dot com, i.e., internet-focused companies. You can see parallels with today’s investors – it is called the “AI Trade” in financial circles for a reason.
Dot Com Bubble Looms in Investors’ Minds
However, there are big differences between 2026 and 2000, and one of those differences is based on a term known as CapEx, i.e., Capital Expenditure. Investors right now are worried that AI-focused companies, including those that are usually seen as “cash rich”, such as Meta, Amazon, Google, and even Tesla, are spending far too much money on AI infrastructure, notably chips and data centers.
There are other reasons for concern, including the fact that AI-focused companies seem to be investing in each other in a merry-go-round way, but the main issue with CapEx is the easiest to understand: These companies are, in a nutshell, spending way more on AI infrastructure and development than they are receiving in revenues. OpenAI, still the most recognizable face of the AI boom, made $20B in revenue in 2025 (that’s revenue not profit; it made a significant loss), yet the CapEx spend among all major AI companies is now measured in trillions.
Railway Mania Is Here Again
It has, therefore, become fashionable to compare the AI boom to the railroad boom of the mid-19th century. Back then, loads of early investors lost money – the period was known as Railway Mania – particularly on smaller railway startup companies. But the outcome was similar to the Dot Com Bubble in that many stronger rail and transport companies emerged from the ashes.
Like the railway companies, AI companies now claim to be building the infrastructure of the ‘new’ internet. Also like the railways at the time, that infrastructure is deemed to be essential, but the passage of the railway construction was never easy, and it should be noted that many of the promised tracks were never built as America embraced the automobile.
We don’t know exactly what is ahead of us – almost every financial expert believes a bust will come at some point – though we do not know how deep and how many scars will be left behind. Right now, those behemoth companies like Google and Amazon are telling us something akin to “you have to break eggs to make an omelet.” We just have to wait and see how many eggs, and how big that omelet really is.
