Wall Street firm believes the AI stock market boom is 'nearing an end.' Here's why

Capital Economics expects the S&P 500 to rise about 8% from roughly 7,650 to 8,250 by end-2026, but then fall 21% to 6,500 by end-2027 as the AI-led equity bubble bursts. The firm cites dot-com-era extremes in earnings-growth expectations and index concentration, alongside record foreign ownership of U.S. equities and renewed net equity issuance. While leverage and volatility are not yet at late-bubble extremes, the analyst says the market peak could be months rather than years away, with technology earnings the key downside trigger.

Analysis

The actionable asymmetry is not an immediate index short: muted leverage and contained constituent dispersion imply that a melt-up can persist while late-cycle supply is absorbed. The more vulnerable exposure is cap-weighted beta, where incremental earnings disappointment from a small set of AI-linked leaders can mechanically drive broad index de-rating. A shift from buybacks to secondary issuance, IPOs, and employee-share monetization would remove an important technical bid precisely when passive ownership has concentrated marginal demand.

Over the next 1-3 months, monitor whether new listings and follow-on issuance are priced aggressively and whether they trade below issue price after lockup expiration. That sequence would matter more than headline valuation because it signals weakening risk appetite among the natural buyers of high-duration growth. The second-order risk is a feedback loop through foreign flows: a weaker dollar can initially cushion local-currency returns, but an equity drawdown combined with dollar softness would reduce the diversification case for U.S. mega-cap exposure and increase redemption pressure.

For 6-18 months, the key distinction is between an earnings reset and a multiple reset. If AI spending remains elevated but monetization lags, hyperscalers and semiconductor leaders may retain revenue growth while seeing free-cash-flow conversion and return-on-capital compress; that is sufficient to impair cap-weighted indices even without recession. The contrarian case is that the absence of extreme leverage means a correction may be shallow and rotational rather than systemic, favoring equal-weight, value, and cash-generative defensives over a wholesale equity exit.

Key Decisions for Investors

• Reduce cap-weighted U.S. equity beta versus equal-weight exposure: pair long RSP / short SPY over the next 3-6 months. This targets concentration and duration risk rather than recession risk; exit if mega-cap earnings revisions continue to broaden upward and the RSP/SPY ratio breaks materially lower.
• Add convex protection rather than an outright near-term short: buy 6-12 month SPX or SPY put spreads, financing only modestly with far-out downside puts. Structure for a 10-15% index decline, where a multiple reset becomes meaningful; avoid oversized premium spend while realized volatility remains subdued.
• Watch new-issue performance as a risk-on/risk-off trigger: if high-profile AI, software, or semiconductor IPOs and follow-ons begin trading below issue price within 20 trading days, increase the SPY hedge and consider short exposure through QQQ versus XLP. The thesis is falsified if issuance is absorbed cleanly and aftermarket performance remains consistently positive.
• Favor FCF-resilient defensives and value cash flows—XLP, XLV, BRK.B—against long-duration growth proxies such as QQQ on a 6-18 month horizon. The relative trade loses if AI monetization produces broad, sustained upward revisions to free cash flow rather than merely revenue and capex growth.