The $1.2 Trillion Stock Shock Coming in 2027: Why Big Tech AI Stocks Are Poised to Underperform

JPMorgan estimates net equity issuance will swing to about $200 billion in 2026 and then surge 500% YoY to roughly $1.2 trillion in 2027, marking the largest two-year period of stock supply since the late 1990s. The article argues that buyback-driven share scarcity is reversing as AI firms and hyperscalers fund roughly $750 billion of infrastructure spending through IPOs and secondary offerings, including proposed financings tied to SpaceX, OpenAI, Anthropic, Alphabet, Meta, and Oracle. The shift is not inherently bearish for AI, but it raises the bar for revenue growth, free cash flow, and ROIC, and could pressure valuations if capital raises outpace returns.

Analysis

The market implication is less about headline dilution and more about a regime change in who sets marginal price. For years, buybacks acted like a structural bid that masked weak underlying demand; a sustained swing to net issuance means the market must now clear a much larger equity overhang just as passive flows are already concentrated in a handful of mega-cap AI names. That creates a subtle but important vulnerability: even if AI fundamentals keep improving, multiples can compress because scarcity itself is no longer doing the work.

The second-order effect is that capital intensity becomes a scoreboard, not a footnote. Companies with the lowest incremental ROIC on AI capex will be punished first because the market will increasingly price them like utility builders with execution risk, while beneficiaries shift toward picks-and-shovels suppliers that earn returns regardless of platform winner. NVDA remains the cleanest relative winner because it monetizes spending from all camps; by contrast, GOOGL, META, and ORCL face a higher hurdle to prove that every additional dollar of investment is accretive rather than simply impressive.

The timing matters: this is likely a months-to-years story, but the first tradable window comes when secondary deals and pre-IPO financing chatter start becoming a recurring headline cycle. Consensus is probably underestimating how quickly buyback support can vanish from the tape once boards prioritize AI funding and compensation dilution over repurchases. The contrarian angle is that the market may still be early in rewarding the supply-chain winners and late in paying up for the capital raisers, especially if the next earnings season shows revenue lagging capex by another quarter or two.

A key tail risk to the bearish supply thesis is that strong index and retail demand can digest issuance longer than fundamentals would suggest, creating a painful squeeze for shorts. But if returns on AI capex disappoint by mid-2026, the market could abruptly rerate the entire cohort from compounder to capital sink. In that scenario, the highest-beta issuers would likely underperform first, while equipment, power, and semiconductor suppliers remain comparatively insulated.

Key Decisions for Investors

• Reduce exposure to GOOGL/META/ORCL on strength and rotate toward NVDA or the broader AI supply chain; use a 3-6 month horizon and expect relative performance to favor the names monetizing capex rather than funding it.
• Pair trade: long NVDA / short ORCL or META into issuance headlines; target 8-12% relative outperformance if AI spend continues outrunning monetization, with the short leg acting as the capital-intensity hedge.
• Buy downside protection on GOOGL and META via 6-9 month put spreads into any post-earnings rally; the trade is attractive if the market starts discounting future dilution and buyback crowd-out before revenue catches up.
• Avoid initiating new long positions in pre-IPO AI-adjacent names until pricing clears and secondary supply is absorbed; the risk/reward skews worse in assets dependent on narrative rather than cash flow.
• Watch for a tactical long in semiconductor and power-infrastructure suppliers on any broad AI pullback, as they should retain pricing power even if platform multiples compress.