End of the Bull Run? What a Dot-Com Bubble 2.0 Could Mean for Gold

Markets rarely ring a bell at the top. They usually change character first.

That is what makes the latest wobble in artificial-intelligence shares more important than a one-day sell-off. A Bloomberg call questioning whether the bull run has finally reached its limit has landed at a moment when several signals are beginning to rhyme: stretched technology valuations, rising AI infrastructure debt, concentrated index leadership and a gold market that is trying to decide whether higher yields or financial stress matter more.

The immediate trigger came from the semiconductor complex. Reuters reported that the Nasdaq fell 1.16% on July 7 as AI-linked chipmakers sold off, with Micron and Sandisk down sharply after Samsung’s strong earnings still failed to satisfy investor expectations. That matters because the AI rally has become less a broad equity advance and more a narrow confidence trade built around the idea that data-centre spending, chips, cloud capacity and model adoption can all compound at extraordinary rates for years.

The bullish case is not silly. AI is not Pets.com. The largest companies funding the buildout have real revenue, real customers and, in many cases, fortress balance sheets. Recent academic work has also warned against treating every surge in a general-purpose technology as pure speculation. A 2026 paper on technology adoption and bubble detection argues that standard bubble tests can mistake a genuine productivity shock for speculative excess when a transformative technology is diffusing through the economy.

But the fragile part of the story is increasingly visible. Another 2026 study, “Boom, Bubble, or Buildout?”, concludes that AI is best understood as a real technological revolution with localized bubble dynamics. That phrase is useful. The internet was also real in 2000. Fibre networks, e-commerce, search and cloud computing did not disappear. What broke was the market’s willingness to pay any price for distant profits.

That is the risk now. The AI trade is moving from software margins to infrastructure economics. Amazon’s reported $25 billion bond sale is part of a broader shift in which hyperscalers are raising debt and equity to fund hundreds of billions of dollars of AI capital expenditure. If investors begin to question payback periods, the story can move quickly from growth optimism to balance-sheet scrutiny.

A dot-com bubble 2.0 would not need to copy 2000 exactly. This cycle is more institutional, more cash-rich and more strategically important. But it has its own pressure points: chip demand that must keep surprising to the upside, power and data-centre constraints, circular financing between model companies, cloud providers and infrastructure owners, and a private-credit market that has become a quiet funding layer beneath the public-market excitement.

For gold, the first reaction to an AI unwind would not necessarily be straightforward. In a sharp liquidation, investors often sell what they can, not only what they want to sell. Gold can be dragged lower temporarily if funds raise cash, the dollar rallies and Treasury yields rise. That is partly why bullion has been choppy despite the risk headlines. Reuters reported spot gold around $4,128 an ounce on July 8, supported by geopolitical tension and Fed uncertainty, but still constrained by a stronger dollar and higher rate expectations.

The second-order effect is more interesting. If a technology-led equity correction becomes a broader confidence shock, gold’s role changes. It stops being merely an inflation hedge and becomes a balance-sheet hedge: a liquid asset without corporate earnings risk, without refinancing risk and without dependence on a future AI monetisation curve.

That is what happened after the original dot-com peak, although with a lag. Gold bottomed near the end of its long bear market around 1999–2001, just as technology optimism was peaking and then collapsing. From there, bullion entered a powerful multi-year advance, helped by lower real rates, dollar weakness, geopolitical stress and a broader reassessment of paper assets. The lesson is not that gold automatically spikes the moment tech falls. The lesson is that a broken equity narrative can create the conditions for a longer gold cycle if policymakers respond with easier money and investors look for stores of value outside the equity complex.

Today’s setup is complicated by gold already trading at historically elevated levels. That makes the metal less of a hidden bargain than it was after the first dot-com crash. Yet it may still benefit from a different kind of scarcity. AI equities are claims on future growth. Bonds are claims on future repayment. Gold is a claim on neither. In a market that has spent years capitalising the future, that simplicity can become valuable.

The critical question is whether this is the end of the bull run or merely another rotation within it. If AI earnings broaden, capex produces visible returns and debt markets stay open, the correction may prove healthy. If leadership keeps narrowing, chip disappointments multiply and infrastructure borrowing keeps rising faster than cash generation, the market may begin to treat AI less like a miracle and more like a duration asset with a very long payback period.

For gold investors, the signal to watch is not only the Nasdaq. It is the combination of real yields, the dollar, credit spreads and central-bank demand. A dot-com bubble 2.0 would become truly bullish for bullion if it forces the Fed back toward easing, weakens confidence in long-duration growth stocks and pushes central banks and investors to diversify away from financial assets priced for perfection.

The bull run may not be over. But the burden of proof has shifted. AI now has to deliver not just excitement, but cash flows large enough to justify the capital being poured into it. Gold, by contrast, only has to remain what it has always been: the asset investors rediscover when the future becomes harder to price.