Every gold rush ends the same way. Here’s why this one is no different.
In 1999, a delivery company called Webvan was worth more on paper than most supermarket chains in America.
It had never turned a profit. It didn’t need to – investors were betting on where the internet was going, not on what the company actually earned. Eighteen months later, Webvan was bankrupt, and the word ‘dot-com’ became shorthand for the most spectacular wealth destruction in modern market history.
We tell ourselves we’d never fall for that twice. And yet here we are, watching companies with barely any profit trade at valuations that assume decades of flawless growth, while the people running those companies quietly admit, in interviews and boardrooms, that the excitement might be running ahead of the reality.
Sound familiar?
Right now, the world’s biggest tech companies are pouring hundreds of billions into AI infrastructure, chasing a future that hasn’t arrived yet. The share prices say the future is guaranteed. The balance sheets tell a messier story.
So here’s the question worth asking before your pension, your portfolio, or your business gets any more tangled up in the AI trade: what happens when the story stops being enough?
Here’s the thesis, stated plainly: the AI boom has all the structural fingerprints of a classic financial bubble, and the gap between what’s being spent and what’s being earned is now too wide to ignore.
That doesn’t mean artificial intelligence is fake, useless, or going away. The internet was real too – and it still crashed a trillion dollars of market value before it grew into the thing we use today. Being right about the technology and being right about the price are two completely different things.
Below is the evidence: the spending, the circular deals, the warning signs the world’s most respected investors are flagging in public, and – because a fair argument earns its conclusion – the strongest case for why this time might genuinely be different.
The All-In-One Marketing Platform That Just Works – FREE!
1. The Spending Doesn’t Match the Earning
Start with the simplest test of any boom: is the money coming in anywhere close to the money going out?
By early 2026, OpenAI alone had committed roughly $150 billion to infrastructure while projecting somewhere near $15 billion in revenue for the year – a ten-to-one gap between ambition and income. Zoom out to the four biggest hyperscalers – Amazon, Alphabet, Microsoft and Meta = and reporting from the Financial Times put their combined AI infrastructure spend on course for around $725 billion, a splurge projected to push their combined free cash flow down to roughly $4 billion for a single quarter. For companies that size, that’s close to a decade low.
None of this proves AI won’t eventually pay for itself. But it does mean the industry is currently being financed on faith, not on earnings – and faith is a lot more fragile than a balance sheet.
2. The Money Is Going in Circles
Here’s the part that should make anyone nervous, however much they believe in the technology.
A growing share of AI’s headline ‘growth’ comes from what analysts now call circular financing. Nvidia invests in OpenAI. OpenAI spends that money on Nvidia chips and on cloud capacity from partners like Oracle and CoreWeave. Those partners, in turn, spend billions buying more Nvidia chips to deliver the capacity they’ve promised. Everyone in the loop books the same pool of money as fresh revenue.
By 2026, analysts estimated more than $800 billion tied up in these interlocking arrangements. Defenders call it a sensible way to finance an expensive technological build-out where suppliers and customers both need certainty. Critics call it what the dot-com era called it the last time vendors financed their own customers: manufactured demand. Either way, the practical effect is the same – a handful of companies are appearing on both sides of the same transaction, and the industry’s growth numbers are propped up by deals that never touch an outside customer at all.
3. Four Warning Signs Wall Street Can’t Ignore
You don’t have to take a blog’s word for it. Some of the most respected names in finance have been saying the quiet part out loud.
Economist Ruchir Sharma uses a checklist he calls the ‘four O’s’ to diagnose bubbles: overinvestment, overvaluation, over-ownership, and over-leverage. By his own account, AI is flashing red on all four. Spending is surging at a pace comparable to the dot-com build-out. Valuations of the major AI players are stretched against long-term earnings and cash flow. Everyday investors are holding a record share of their wealth in equities, and most of those bets are AI-related. And after years of famously cash-rich balance sheets, Big Tech has started issuing serious debt to keep funding the arms race.
Bridgewater founder Ray Dalio has put a number on it, describing the AI bubble as sitting at roughly 80% of the euphoria that preceded the 1929 crash and the 2000 dot-com bust. Even OpenAI’s own CEO has publicly admitted investors might be ‘overexcited about AI.’ When the person running the company at the centre of the boom says that out loud, it’s worth listening.
4. This Isn’t Spread Out – It’s Concentrated
A bubble in one obscure corner of the market is a curiosity. A bubble sitting at the centre of everyone’s pension fund is a different kind of problem.
AI and AI-adjacent stocks now make up a larger share of major indices than dot-com stocks did at the peak of that bubble in 2000. That concentration doesn’t guarantee a crash. But it does mean that if the AI trade stumbles, it won’t stay contained to a niche sector – it drags the index, the pension fund, and the retirement account down with it, whether or not you ever bought a single AI stock on purpose.
5. The Cracks Are Already Showing
This isn’t a purely theoretical risk sitting somewhere in the future. Bits of it are already playing out in real time.
In June 2026, weak guidance from chipmaker Broadcom triggered a sharp sell-off across the exact stocks that had powered the AI rally – Nvidia and the major cloud providers turned into the day’s biggest losers, as traders openly questioned whether earnings growth could keep justifying current valuations without imminent interest rate cuts. It wasn’t the first wobble, and analysts don’t expect it to be the last.
Underneath the market jitters sits a harder problem: adoption. One widely cited MIT study found that 95% of generative AI pilots inside companies had so far failed to produce a measurable return. That’s not a technology failing to work – it’s a technology struggling to translate into the kind of enterprise revenue that would justify the trillions being spent building the infrastructure to run it. Add rising power costs and grid constraints around the data centres themselves, and the picture gets more precarious, not less.
Powering 500,000+ entrepreneurs Get your free account now!
| The One Number That Matters Most
OpenAI – the company that kicked off this entire boom – is reportedly on track to lose around $14 billion in 2026, nearly triple its 2025 losses, while it has committed over $1 trillion in infrastructure spending through 2035. That’s not a company being cautious with other people’s money. That’s a company betting the entire industry’s credibility on demand that hasn’t shown up yet. |
Dot-Com Bubble vs. AI Boom: Side By Side
Skeptics keep reaching for the same comparison. Here’s how the two eras actually line up, side by side.
| Dot-Com Bubble (2000) | AI Boom (2026) |
| Companies spent on unproven growth, funded largely by debt and stock issuance. | Companies are spending on unproven growth, increasingly funded by rising debt alongside cash flow. |
| Vendor financing inflated apparent demand between telecom firms. | Circular financing between chipmakers, cloud providers and AI labs inflates apparent demand. |
| Tech stocks reached a historic share of major indices. | AI-related stocks now hold an even higher concentration of major indices than in 2000. |
| The underlying technology (the internet) was genuinely transformative and survived the crash. | The underlying technology (AI) is widely seen as genuinely transformative – the debate is about price, not potential. |
6. The Strongest Case Against ‘Bubble’
A fair argument doesn’t hide from its best counterpoints, so here’s the case the bulls are making – and it’s not a weak one.
Unlike most speculative manias, today’s AI leaders are genuinely profitable. Microsoft, Alphabet, Meta and Amazon have real earnings, real cash flow, and real customers outside the AI trade entirely. Analysts at Fidelity note that, historically, these companies have funded the bulk of their AI capital spending from earnings rather than debt – a meaningfully different starting position than the loss-making dot-com darlings of 1999. Goldman Sachs and J.P. Morgan have both argued the growth is fundamentally justified by real productivity gains, not just hype.
The honest answer sits between the two camps: AI could change the world and still be overpriced today. Those aren’t contradictory statements – they’re exactly what happened with the internet. The technology won. A huge number of the companies betting on it in 2000 didn’t.
Before you close this tab, here are the questions people ask most when this topic comes up.
Frequently Asked Questions
Is the AI bubble definitely going to pop?
No one can say that with certainty, and anyone who tells you otherwise is guessing. What’s measurable is the gap between spending and earnings, the scale of circular financing, and rising leverage – all classic pre-correction signals, not proof of an imminent crash.
What would actually trigger a pop?
Most analysts point to interest rates. Higher rates make the cheap capital funding AI’s build-out harder to access and put downward pressure on the growth-stock valuations the whole trade depends on. A sharp slowdown in enterprise AI adoption would have a similar effect.
If AI is a bubble, does that mean the technology is worthless?
Not at all. The dot-com bubble popped and the internet still went on to reshape the entire global economy. A bubble is about price catching up with reality, not about whether the underlying technology works.
Should I sell my tech stocks right now?
This isn’t financial advice, and no blog post should be the deciding factor in your portfolio. What’s worth doing is understanding your own exposure – directly through individual stocks, or indirectly through index funds and pension funds that are more concentrated in AI than most people realise.
How is this different from normal market ups and downs?
Ordinary volatility is normal and expected. A bubble is specifically when valuations detach from the earnings that are supposed to justify them, propped up by leverage, concentration, and – in this case – deals where the same handful of companies are financing each other’s revenue.
Nobody rings a bell at the top of a bubble. The people who called the dot-com crash sounded paranoid for years before they were proven right – right up until they weren’t.
What we do know is this: the spending is unprecedented, the earnings aren’t close to catching up, a huge slice of the industry’s ‘growth’ is companies financing each other in a closed loop, and some of the most credible investors in the world are calling this exactly what it looks like. That’s not a prediction. That’s just what the evidence says today.
Whatever happens to the AI trade, one thing won’t change: relying on a single sector, a single narrative, or someone else’s balance sheet for your financial future was never a solid plan – bubble or not.
Bookmark the numbers in this post. Whichever way the AI story goes over the next 12 months, you’ll want to remember you read them here first.
Build Income That Doesn’t Depend on the Next Big Tech Earnings Call
Chasing the AI trade is a bet on someone else’s business. Building your own digital products and income streams is a bet on you – and it doesn’t unravel because a chipmaker missed a revenue target. Systeme.io gives you the sales pages, email automation and checkout tools to start that, free.
Start your free Systeme.io account and build something that’s actually yours
Let Us Help You Grow
Fast, Easy & Effective Classified Ads | The Essentials in Leading Digital Marketing Training | The Best Budget SIM-Only Deals in the UK | Website Hosting Built for Success | Your Classified Ad Promoted to 1000’s+ Advertising Pages Each Month! | Run your business – leave the email marketing to us | The All-In-One Marketing Platform That Just Works! | Online Booking System For All Service-Based Industries