Is This Dot-Com All Over Again?

The Parallels

• Infrastructure first. Fiber then, datacenters and GPUs now. What they share is not the technology but the sequence: build capacity against an expectation, then wait for demand to arrive.
• Extreme concentration of capital. In 2026Q1, global VC into AI startups reached $242B — 80% of all global venture funding (Crunchbase). For contrast, the same figure was $45B in 2025Q3: more than a fivefold rise in three quarters. When one sector absorbs four fifths of the money in the market, diversification has stopped existing in practice.
• Circular financing. Chipmakers arrange funding for customers who spend it on chips. Whether Nvidia's $500B financing lined up with Wall Street partners is such an arrangement is itself being argued in public.
• Exposure nobody chose. Norway's sovereign fund reported a record $184B gain from the AI rally while stating plainly that it cannot hedge the exposure. Index investing means an enormous pool of capital is long AI without ever having decided to be.

The Differences

This cycle's leaders are mature companies with cash flow, not concept-stage startups. Microsoft's FY2026 shape is the illustration: Windows client is only about 5% of total revenue, growth has moved wholly to cloud and AI, and profitability there remains well above the old business even under heavy capex — the stock rose about 15% in a day after earnings, adding close to $500B in market value. The dot-com cohort had nothing like this underneath it.

That is not an argument that this isn't a bubble. It only says that if this one breaks, it breaks differently from 2000: not companies vanishing en masse, but profitable companies writing down a class of assets that did not earn their cost. Those are very different risks, and one historical analogy should not be stretched over both.

Where China Stands

China went through its own AI valuation bubble with the "AI Four Dragons" (2015–2018) and the painful digestion that followed; this round is big-tech-led and comparatively restrained. The numeric gap is an order-of-magnitude story: per the same table of the same Stanford HAI report, 2025 US AI private investment was $285.9B against China's $12.4B; in 2024 it was $109.1B against $9.3B. Across those two years the gap widened rather than closed.

Concentration, though, is a shared problem. ChangXin Memory Technologies listed on the STAR Market in July 2026, opening up 471.59% and reaching an intraday market cap of RMB 3.66 trillion — passing Intel and Tencent to become the most valuable A-share company. A memory-chip maker reaching that position is itself a signal about where money is going.

How to Use This Page

The three charts below — Nvidia's market cap, global AI venture funding, and Sequoia's revenue-versus-capex gap — plot verified points only. The signal list filters the feed live under valuation, markets and funding. The parallels and the differences are both laid out here; the site does not pick a side for you.