For years, economists and investors have debated whether artificial intelligence is a massive financial bubble waiting to burst. However, Dhaval Joshi, a seasoned strategist and former head of strategy at BCA Research, suggests that asking if AI is a bubble is fundamentally the wrong question. According to Joshi, we aren’t dealing with one giant balloon destined for a single catastrophic explosion, but rather a rolling sequence of smaller bubbles that inflate and pop in rapid succession across different sectors.
This phenomenon explains why the broader market hasn’t collapsed despite significant volatility in specific niches. Instead of a total meltdown, investors are cycling through various bets on who will actually capture the value of AI. For instance, software stocks initially soared on the promise of productivity gains, only to crash as traders realized that AI agents might actually destroy the traditional subscription models these companies rely on. Similarly, silver prices spiked because it is an excellent conductor for energy hungry data centers, but plummeted once the market decided such a surge wasn’t justified by actual demand. Now, Joshi argues that semiconductor stocks may be facing a similar fate as astronomical profit margins eventually collide with reality.
While some argue this is simply standard price discovery where markets test theories and correct themselves, Joshi believes the sheer speed and scale of these swings point toward something more manic. To him, when fortunes are made and lost within weeks due to narrative contagion rather than fundamentals, it qualifies as a bubble regardless of how small the sector is. This fragmented approach helps explain why tech giants continue to spend billions on infrastructure even as critics warn of overextension; the money isn’t disappearing so much as it is migrating from one perceived winner to the next.
Despite this cyclical resilience, there are limits to how long this game of musical chairs can last. While leaders like Jamie Dimon and Sam Altman have acknowledged elements of bubbliness in the current climate, many analysts expect the heavy spending phase to peak around late 2026 or early 2027. The ultimate danger remains external factors like rising interest rates or sharp bond yield spikes which could cause investors to flee risky assets entirely. Until then, according to Joshi, we are simply watching a series of rapid inflations and deflations as the world tries to guess where the true value of AI resides.