Not “AI is fake” — “AI is currently priced as though the future already arrived,” which is a different and much more falsifiable claim.
“Peak X” framing has a real intellectual lineage worth knowing before using it casually. M. King Hubbert’s 1956 peak-oil model — the point at which extraction from a finite resource stops accelerating and begins to decline — gave the pattern its name and its rigor: a peak is not a judgment that the resource, or technology, is worthless, only that its rate of expansion has a ceiling the current trajectory is approaching or has passed.
Applied to technology and capital rather than oil, Carlota Perez’s Technological Revolutions and Financial Capital (2002) offers the more directly relevant model: every major technological revolution goes through an “installation period” where speculative financial capital massively over-funds the new technology’s roll-out, produces a bubble, and then — after the crash — a “golden age” where the surviving, now-cheap infrastructure gets put to genuinely productive use. Railways, electrification, and the dot-com era all followed this arc.
Whether AI investment specifically is in that installation-period bubble phase is now itself a live research question, not just online debate — a 2026 multi-method study by Wang and Chen concludes AI shows both genuine fundamentals and localized bubble dynamics simultaneously, rather than being purely one or the other. “Peak AI,” read this way, is a claim about pricing getting ahead of monetization in specific parts of the AI stack, not a claim that the technology is a mirage.
Source: Hubbert, M.K. (1956), “Nuclear Energy and the Fossil Fuels,” Shell Development Company; Perez, C. (2002), Technological Revolutions and Financial Capital: The Dynamics of Bubbles and Golden Ages, Edward Elgar Publishing; Wang, Q. & Chen, Z. (2026), “Boom, Bubble, or Buildout? A Multi-Method Evaluation of Whether Artificial Intelligence Is in an Ongoing Financial Bubble,” arXiv:2606.01575.