Every bubble starts as a reasonable success story. Only later does it mutate into a fragile structure that cannot bear its own weight. Goldman Sachs says the pieces for a classic stock bubble are not in place. Fair. The better question is what would snap into place to turn this rally into one. Valuation by itself is not enough. Bubbles need an accelerant. In markets, that accelerant is often cheap funding married to a dominant narrative that suppresses doubt. When the story feels irresistible and the cost of being wrong feels low, the system loads risk silently.
A rally crosses into bubble territory when three conditions align. First, easy and scalable leverage. Second, a singular growth story that compresses debate and rewards index hugging. Third, reflexive mechanics that transform flows into further gains. In the 1920s it was broker credit and industrial electrification. In the late 1990s, margin debt and the internet narrative. In housing, adjustable-rate credit and securitization. AI today supplies the story. The remaining ingredients are leverage and reflexivity. Goldman is right that not all the puzzle pieces are in place. But they never are until late in the game. Stability breeds risk taking. The phase change is nonlinear. Like a bridge that seems solid until resonance meets load.
Valuations warn you about impact, not timing. Shiller CAPE was elevated in the mid-1990s and stocks doubled again before the break. The Nifty Fifty stayed expensive until inflation and rates rose. Today’s leaders in AI are profitable. Cash flows are real, as several institutions have emphasized. Relative to some private assets and segments of real estate, mega-cap tech does not look like the most extreme corner of the market. That is why calling a top on valuation alone has been painful for decades. A veteran at Goldman has pointed out how costly it is to bet against a swelling tech trend. The base rate says expensive can get more expensive if funding is loose and the narrative is dominant. High multiples increase fragility to shocks, but they do not cause the shock.
Every historical bubble is a leverage story. Sometimes the leverage is obvious, like margin debt. Sometimes it is hidden in structures and funding chains. Collateral rules relax. Credit standards slip. Equity behaves like call options when financing is cheap. Option-like behavior then calls in even more option sellers, reaching for carry. With policy rates higher, it is easy to assume leverage is contained. That assumption is the fragility. Leverage migrates. It hides in basis trades, total return swaps, structured notes, vendor financing, and even corporate buybacks funded with debt. When real funding costs fall or volatility drops and stays low, risk systems greenlight bigger positions. Funding stability, even at higher nominal rates, can act like cheap credit if spreads compress. Bubbles are a function of perceived safety, not sticker price on money.
Reflexivity turns a strong trend into a flywheel. Passive index funds buy more of what goes up. Concentration rises, then flows chase concentration. Option hedging amplifies the move. When dealers are short calls to retail or institutions, they buy stock as it rises to stay hedged, mechanically pushing prices higher. The rise of very short-dated options compresses realized volatility and creates the illusion of control. It works until a large gap forces hedging in the other direction. This is a known feedback loop. It is not new, but its scale is. Combine a narrow leadership group with passive’s mechanical buying and option-induced convexity, and you have a system that looks stable right up until a liquidity air pocket appears. In game theory terms, this is a coordination problem. Everyone stays in as long as everyone else stays in. Exit doors are narrow.
It is comforting that today’s leaders make money. It should be. Profitability is a cushion. But a cushion is not a shield. Earnings strength has concentrated in a few platforms and their supply chains. AI has real demand, but it also has real bottlenecks. Data center buildouts require power, land, and specialized chips. Permits take time. Utilities face constraints. Supply elasticity is not infinite. A shock to any link — energy costs, chip yields, policy scrutiny, export controls — can ripple fast through a concentrated ecosystem. Think monoculture in agriculture. It produces more per acre, until a single blight wipes out a harvest. In markets, concentration lowers surface volatility while increasing tail risk. Profits are not the problem. Fragility to a specific set of inputs is.
The market is not a classroom. It is a job. Career risk shapes flows. If you short leaders and they rise, you look wrong and may be out of work. If you underweight the index and it runs, you underperform. The safest place for many is in the consensus. That is the Keynes beauty contest at work. It is why valuation arguments lose to price in the short to medium term. This is not moral failure. It is system design. Even Goldman’s own stock has carved out a technical base and pushed higher, mirroring the tailwinds a strong market brings to deal flow and trading. Banks thrive on activity and rising asset prices. When the system rewards participation and punishes skepticism, bubbles become more likely. Not because people forget history, but because incentives overpower memory.
If there is one thing that would flip this rally into a full bubble, it is a durable easing in effective funding conditions while the AI narrative remains dominant. That could come from falling real yields, a compression in credit spreads, or a long stretch of low volatility that invites leverage. The signal to watch is not headlines or even index levels. It is the migration of risk into weak hands. Look for a surge in unprofitable IPOs getting done at scale. Watch for covenant-light credit pushing out maturities at low spreads. Track the share of daily volume dominated by very short-dated options. These are not tips. They are base-rate markers. When financing becomes frictionless and skepticism is punished, the rally becomes reflexive. Then small shocks get amplified, not absorbed.
Goldman is right to say the puzzle is incomplete. That is how it looks before every late stage. The paradox is that solid profits and real use cases extend the runway. Time is what leverage needs. The market becomes fragile not when people are euphoric in public, but when dissent is costly and risk controls are calibrated to recent calm. Minsky called it stability breeding instability. Taleb would call it the absence of redundancy. In nature and engineering you pay for resilience with slack. Markets hate slack. So the line between a durable trend and a bubble is crossed when funding gets easy, concentration tightens, and reflexive flows take over. That is the trigger. You will not see it on a banner headline. You will feel it in how easy it becomes to lever a good story and how hard it becomes to stand aside.
