AI Bubble vs Dot-Com Bubble: Why the Comparison Fails

The AI bubble vs dot-com bubble comparison fails because today's AI leaders already book billions in real revenue, while most 2000 dot-com failures never earned a profit at all. Nvidia, Alphabet and Microsoft sell physical hardware and cloud capacity rather than a website and a promise. The remaining risk sits in infrastructure spending, not in empty ticker symbols.

AI Bubble vs Dot-Com Bubble: What Actually Differs

The AI bubble vs dot-com bubble comparison fails on earnings: Nvidia, Alphabet and Microsoft already sell hardware and cloud capacity at scale, while most 2000 dot-com listings had no profit at all. The 2026 debate is about whether heavy infrastructure spending pays back, not whether a ticker symbol has revenue behind it.

The 2000 crash is often illustrated with Pets.com, an online pet-supply retailer that went public in February 2000, raised roughly 80 million dollars and began liquidating by November of the same year. The company sold pet food below cost and let shipping absorb the margin. Nvidia, by contrast, reported 96 billion dollars in quarterly revenue in July 2026, with 89 billion of that attributed to data-center sales in the video's account.

Lucas Montano makes the case in his 2026 video that the two cycles are structurally different, and his argument rests on who is borrowing and who is earning. He also concedes he is not giving financial advice and that a correction remains possible. That concession matters because the strongest version of his position is narrower than the viral version.

Why Regulators Still Warn About a Correction

The Bank of England's Financial Policy Committee flagged in late September 2026 that the risk of a sharper correction persists, especially if earnings expectations are shocked by the pace of AI development or adoption. That is a warning about valuation and concentration, not a prediction that AI companies resemble Pets.com.

The same committee record points to growing indebtedness among AI firms, coupled with opacity and what it calls circular arrangements. Morgan Stanley data cited in the video puts AI-linked debt issued through early September 2026 at roughly 450 billion dollars, more than double the whole of the previous year. Treat that figure as reported by the speaker rather than independently reconstructed here.

Read together, the evidence supports two claims at once. AI capital spending is enormous and partly debt-funded, and the largest companies in the cycle are highly profitable. A correction in the first does not require a collapse in the second, which is why the 2000 analogy overstates the downside shape.

The Numbers Behind the Comparison

Earnings growth, not price alone, drives the current valuation picture. According to earnings insights dated October 9, 2026 and discussed in the video, the S&P 500 traded near 19 times forward earnings, slightly below its five-year average of 19.8 and in line with the ten-year average.

The table below separates the two cycles on the dimensions where evidence exists.

Where the Infrastructure-Bubble Analogy Fits Better

Railways, electrical grids and fiber networks are better analogues than Pets.com because they all involved real assets, real demand and real overbuilding. Each cycle delivered lasting infrastructure and still produced painful corrections, which is exactly the pattern the AI buildout could follow.

The four largest hyperscalers, Microsoft, Alphabet, Amazon and Meta, told investors in 2026 results that infrastructure spending for the year would land around 720 billion dollars, nearly double the prior year according to the video. Not all of that money funds AI, since Amazon also spends on robotics and satellites and Alphabet develops its own chips and data centers.

That distinction changes the question. The risk is not that the revenue is fictional but that the payback period on data centers may run longer than the current spending pace assumes.

Circular Deals and the Opacity Problem

Circular arrangements are the most underreported risk in the current cycle, because money can move between vendors, investors and customers in ways that flatter reported demand. The Bank of England's committee record names opacity alongside rising AI-firm debt, which makes the true leverage of the sector harder to assess.

A simple illustration helps. If a chipmaker invests in a customer that then buys its chips, the revenue is real, but the demand signal is weaker than an independent purchase would be. Auditors and regulators can only evaluate that pattern if the transactions are disclosed clearly.

What This Means If You Are Building With AI

Builders should plan for cheaper or pricier compute rather than for the disappearance of AI vendors. Demand exists even among people who distrust the technology, and the video's closing point is that mass-market adoption has barely started. Pricing and availability of inference can still shift sharply if a correction hits capital spending.

The practical hedge is portability. Keep model calls behind a thin abstraction, track token costs per feature, and avoid architectural commitments that only work with one vendor's newest accelerator. Developers who treat compute as a volatile input make better decisions than those who treat it as a fixed cost.

Learning to Separate Evidence From Hype

The bubble debate rewards people who can read a filing. Portugal-based developer communities and Brazilian creators like Gustavo Dev Doido often push viewers toward hands-on practice instead of market commentary, and that habit transfers directly: check the earnings release, the debt disclosure and the regulator note before repeating a chart.

Courses such as Crazystack Typescript and the Bootcamp do Dev Doido on crazystack.com.br focus on shipping software rather than on predicting markets. That distinction is useful here, because the infrastructure being built will need engineers regardless of where valuations land.

FAQ

• Is the AI bubble really the same as the dot-com bubble? No. Most 2000 dot-com failures had no earnings, while the largest AI companies in 2026 report billions in revenue and high gross margins. The closer analogue is an infrastructure cycle such as railways or fiber, which delivered lasting assets and still produced corrections.

• What did the Bank of England say about AI stocks? Its Financial Policy Committee record from late September 2026 said the risk of a sharper correction persists, particularly if earnings expectations are shocked by the pace of AI development or adoption. It also flagged rising AI-firm debt and circular arrangements.

• How much AI-linked debt was issued in 2026? The video cites a Morgan Stanley estimate of roughly 450 billion dollars issued through early September 2026, more than double the previous year. That figure comes from the speaker's reading of the report rather than an independent audit.

• How big was Nvidia's quarter in July 2026? Nvidia reported 96 billion dollars in revenue for that quarter, with 89 billion attributed to data centers and a gross margin of 75 percent, according to figures presented in the video. Those are company-reported results, not independent measurements.

• What is the biggest risk in the current AI cycle? Capital allocation. Hyperscalers committed around 720 billion dollars to infrastructure in 2026, and the risk is that data centers take longer to pay back than the spending pace assumes. Debt opacity makes that risk harder to size.

Turn a Video Argument Into a Written One

The distinction between an earnings-backed cycle and an empty-promise cycle is the kind of argument that survives better in writing, where a reader can check each figure against its source. If you have made that argument on camera, the transcript is already the draft.

Skalablog takes a YouTube URL, transcribes the video and produces a structured article you can edit before publishing. Paste a link at Skala Blog and start from the transcript you already recorded.

Source video