Andy Constan argues the current AI-led market has bubble characteristics: a new technology shock, escalation through FOMO and policy easing, and a peaking phase driven by stretched expectations around semis and AI capex.
Watch on YouTube ›Get the market thesis, key claims, assets, contradictions, and follow-up questions from any financial video — then unlock a version personalized to your portfolio, watchlist, and favorite speakers.
This is a long-form interview on Excess Returns with Andy Constan about bubbles, how they form, and how the current AI/semiconductor complex compares with prior episodes. Constan says bubbles are extremely hard to identify in real time and impossible to time precisely, but bubble regimes have repeatable ingredients: a new catalyst, escalation events, and then a peaking phase. He walks through five historical cases he views as relevant templates: the early-1980s to 1987 equity regime, the internet bubble beginning with Netscape in 1995, the 2005-2008 housing/credit bubble, the post-GFC/ZIRP bond bubble that climaxed around COVID, and today’s AI boom. His core framework is that bubbles usually begin with something genuinely new: a technological breakthrough, major regulatory shift, extreme monetary easing, or a major exogenous shock. …
Near term, the AI/semiconductor trade still has momentum, but the key risk is whether the market can digest the next round of issuance and capex financing without wobbling. If supply of equity and debt grows faster than risk appetite, this becomes a tactical headwind.
Over the next few months, the base case is continued leadership as long as earnings revisions and capex commitments keep surprising higher. The setup weakens if funding costs rise, buybacks stay suppressed, or investors stop extrapolating the current growth curve.
Structurally, Constan sees AI as a real productivity technology that can reprice capital markets, but not without creating a distribution fight over who captures the gains. The long-run regime question is less whether AI works than whether its financing and social consequences trigger policy resistance or a broader reallocation of returns.
You cannot reliably identify a bubble top in real time; the best you can do is recognize the regime.
He says it is easy in retrospect but impossible to determine when a bubble will pop.
The current market looks like a bubble regime, with AI as the new catalyst and peaking behavior already visible.
He explicitly says the environment today looks similar to prior bubble environments and that we are in the peaking phase.
Bubble regimes tend to begin with a genuinely new thing: technology, regulatory change, monetary easing, or an exogenous shock.
He lays out the root-condition framework with multiple historical examples.
How do you define a bubble and how can investors see it in real time?
Andy says bubbles are incredibly easy to define in retrospect but nearly impossible in foresight. He emphasizes that the conversation should be about what is different in a bubble regime, not about predicting when a bubble will pop, which he calls the holy grail of investing that nobody can actually achieve.
How do you separate a market that is expensive and overowned from one seeing bubble-like characteristics driven by a real technological breakthrough?
Andy first challenges the premise, arguing that current market prices are neither expensive nor cheap — every investor has exactly what they want to own at equilibrium. He says there is no such thing as overbought or oversold. He then explains that bubble regimes have identifiable root causes and characteristics, pointing to four similar environments in his career (1981-87, the internet bubble, 2005-2008 housing, and post-GFC government bonds), and notes that today's environment looks similar. He outlines that these regimes typically start with something new — in today's case, AI technology.
Did you have any takeaways from the Japan bubble of the 80s, given you were investing through it?
The speaker worked for Solomon Brothers, which had a massive presence in all those markets. As a personal investor on the sell side he was highly limited, but he remembers buying Japanese NIK put warrants in 1989 as the bubble peaked. He notes the Japanese bubble had many of the same aspects: a housing bubble, equity bubble, easy money driven bubble.
Unlock the full claims, asset map, scores, related transcripts, follow-up questions, and AI chat — shaped around your portfolio, watchlist, favorite speakers, and risks.
