According to Bloomberg Markets, veteran Goldman Sachs strategist and Columbia Business School professor Abby Joseph Cohen has issued a pointed warning that current stock valuations should concern investors — a statement carrying particular weight given that Cohen was a prominent Wall Street voice during the late stages of the 1990s technology boom. Her intervention arrives at a moment when the Federal Reserve and the European Central Bank are charting markedly different courses, a divergence that historically has been among the most reliable preconditions for misallocated capital at scale.

The result is a self-reinforcing dynamic. U.S. asset prices rise not solely because U.S. earnings justify them, but because global capital has limited alternatives. Valuations become untethered from domestic fundamentals, and the correction, when it arrives, is amplified by the reversal of cross-border flows rather than being contained within a single market. For current bubble risk indicators, this mechanism is among the most structurally dangerous in the present environment.

What Cohen's Warning Actually Means

Cohen's credibility on this subject is not incidental. During the dot-com era she was a persistent bull, and her eventual pivot away from that stance in early 2000 — after the Nasdaq had already begun its descent — became a case study in how even experienced analysts can be captured by prevailing sentiment. Her current caution, arriving before an obvious catalyst rather than after one, deserves to be taken seriously as a signal rather than dismissed as routine strategist hedging.

The valuation concern she raises is structural. Price-to-earnings multiples on the S&P 500, by most conventional measures, remain elevated relative to long-run historical averages, particularly when adjusted for the interest rate environment. The Dow Jones, as Bloomberg's Closing Bell coverage noted, finished the most recent session lower even as the weekly trend remained positive — a pattern consistent with a market absorbing conflicting signals rather than one in orderly advance.

"Stock valuations should worry investors" — Abby Joseph Cohen, as reported by Bloomberg Markets, June 26, 2026. Cohen's observation arrives in a market environment where cross-border capital flows, rather than domestic earnings growth, are doing much of the valuation work.

South Korea's Leveraged ETF Warning: A Peripheral Signal Worth Watching

Peripheral markets often provide the earliest evidence of speculative excess before it manifests in headline indices. Bloomberg's coverage of South Korean regulators lamenting the growth of leveraged ETF activity is precisely this kind of signal. South Korean retail investors, operating in a market where leveraged and inverse ETFs have proliferated rapidly, are exhibiting the kind of amplified directional positioning that tends to appear in the later stages of speculative cycles rather than the early ones.

The Telecom Parallel: Capital Without a Destination

The most instructive historical parallel for the present configuration is not the 2008 housing crisis or the 2020-2022 crypto cycle, but the Telecom Bubble and Fibre Optic Mania (1998-2002). That episode is frequently misremembered as simply a technology story, but its deeper structure was a capital allocation story driven by monetary conditions. The late 1990s saw a prolonged period in which cheap debt, equity market euphoria, and a broadly accommodative global monetary stance directed vast sums into fibre optic infrastructure. When the cycle ended, the great majority of the laid fibre sat dark and unused, and the wreckage included WorldCom's massive accounting fraud — one of the largest in corporate history.

The parallel to today is not that AI infrastructure spending will produce identical waste, though the comparison between the dot-com bubble and AI infrastructure investment in 2024 is worth examining carefully. The parallel is structural: when central bank divergence channels global capital into a single dominant market, and when that market's valuations have already been bid above long-run sustainable levels, the capital does not simply sit idle. It searches for yield and growth, and it finds them in the narratives that are most compelling at the time. In 1999, that narrative was bandwidth. Today, it is artificial intelligence and large-cap U.S. technology. The narrative may prove more durable than fibre optic mania — or it may not. What history suggests is that the scale of eventual correction correlates closely with the degree of valuation excess and the volume of cross-border capital that needs to reverse.

For broader equity valuation context across global markets, the current environment warrants the kind of scepticism that Cohen is articulating: not panic, but a clear-eyed recognition that price levels reflect policy conditions as much as they reflect earnings fundamentals.

Why This Configuration Is Historically Unstable

Central bank divergence resolves in one of two ways. Either the lagging central bank catches up — tightening to match the leader — or the leading central bank pivots, cutting to meet deteriorating conditions. In either scenario, the capital flows that have been sustaining elevated valuations in the dominant market face reversal pressure. The timing of that reversal is, by definition, unknowable in advance. What is knowable is that the longer the divergence persists and the more capital it attracts into a single market, the more disorderly the eventual realignment tends to be.

The telecom mania of the late 1990s unravelled not because anyone suddenly decided fibre optic cables were worthless, but because the monetary conditions that had made the investment appear rational shifted. Rates rose, equity financing dried up, and the capital that had flowed in reversed. The assets — the cables, the switches, the spectrum licences — remained. Their valuations simply had to reconcile with a world in which cheap capital was no longer available to sustain them. The lesson for 2026 is not that U.S. equities are fraudulent or that AI spending is waste. It is that valuation levels which depend on persistent policy divergence for their justification are inherently fragile — and that fragility tends to become visible only after the conditions that concealed it have already changed.