The Bubble That Keeps Moving
The numbers come first. Silver nearly tripled in price on data-center power demand, then collapsed. Software-as-a-service stocks surged on AI productivity hopes, then cratered when investors realized AI agents threatened the subscription model itself. Semiconductor valuations soared on assumptions of limitless pricing power — and are now unwinding, with Joshi warning they have 'further to go.'
That is the pattern Dhaval Joshi, until recently chief strategist for Counterpoint at London's BCA Research, calls a 'rolling sequence of bubbles.' Writing on LinkedIn, Joshi reframed the entire AI-bubble debate: rather than one giant inflation building toward a single implosion, capital is cycling through rapid-fire manias — each sector catching fire, overshooting every plausible fundamental, and then deflating before the narrative migrates elsewhere.
Margins, Not Just Multiples
In an interview with Fortune, Joshi said he slightly disagreed with his former BCA colleague Peter Berezin, who has described the market as an earnings bubble. Joshi calls it a 'profit margin bubble' instead. The price-to-earnings ratio may look defensible, he argued, but the real question is: 'How is the E high? Because you've got very high margins, but can you maintain those margins?'
The semiconductor case illustrates the thesis. Chipmakers rode a narrative of near-monopoly pricing power, but Joshi contends they lack durable 'moats' around their profits. His prediction: 'Astronomical margins will crash back to earth when demand and supply equilibrate, as they ultimately must.'
The silver episode is perhaps the starkest data point. The metal spiked because it is seen as the best electrical conductor for power-hungry data centers. 'On reassessment however,' Joshi wrote, 'this could not justify a near trebling of the silver price when there are other good conductors.' The misallocation, he noted, is not confined to equity markets.
Hyperscaler Cash Flow Enters the Frame
The latest earnings season sharpened the debate. Google went free cash flow negative for the first time in its history as capital expenditure consumed operating cash. Reuters calculated in late July that Microsoft, Alphabet, Amazon, Meta, and Oracle were on pace for capex to overtake free cash flow by 2027. The open question is not whether overspending is occurring but whether it is rational.
Joshi is not alone in his concern. Jamie Dimon has repeatedly flagged elevated valuations. Bank of America Research's Global Fund Manager survey named 'AI equity bubble' as the top tail risk. OpenAI CEO Sam Altman, Goldman Sachs CEO David Solomon, and Amazon founder Jeff Bezos have each acknowledged bubbly conditions — yet the bubble that was supposed to pop in 2025 kept inflating.
The cyclical reinflation has, so far, prevented a correlated selloff across asset classes. But Joshi's framework offers no guarantee that the rolling pattern holds indefinitely. 'In real time,' he told Fortune, 'we are making educated guesses about which rapid inflations are at risk of rapid deflation.'
The Market Has Already Voted — Repeatedly
What Joshi's framework reveals is a market that is doing price discovery at an abnormal velocity and magnitude. Free enterprise depends on capital finding its most productive use; bubbles of this speed and amplitude suggest narrative contagion is doing the allocating, not fundamentals. When silver triples and then collapses in months, when SaaS multiples swing violently on shifting AI assumptions, the signal-to-noise ratio for long-term investors deteriorates sharply.
The question every capital allocator should be asking is not 'is AI real?' — the technology clearly is — but 'which sector is absorbing the next wave of misallocated enthusiasm, and how quickly will the correction arrive?' Joshi's rolling-bubble model does not predict a systemic crash. It does predict that the bill for each mania arrives faster than most portfolios are positioned to absorb.
