Goldman Sachs Exchanges

2026-08-03 · Hosted by Allison Nathan · Goldman Sachs

Executive Summary

The 2026 IPO market has reopened after a multi-year drought, with issuance proceeds hitting a record this year. Allison Nathan spoke with Jay Ritter (University of Florida) and Owen Lamont (Acadian Asset Management) about whether this constitutes a genuine "IPO wave" that historically precedes market downturns. Both agree it isn't a wave yet: the number of operating-company IPOs remains modest by historical standards, even though total dollars raised is record-setting.

Key Stories & Changes

1. Record IPO Issuance, But Not Yet an "IPO Wave"

  • 2026 IPO proceeds are at a record high by dollar volume, but the number of operating-company IPOs remains "fairly modest" compared with historical waves.

  • Ritter: post-dot-com-bubble years averaged ~100 operating-company IPOs/year, versus over 300/year in the 1980s-90s.

  • Lamont's rule of thumb for a true wave: an IPO happening every business day (~5/week); current pace doesn't meet that bar.

  • Two reasons for continued modest IPO counts: (1) expansion of venture capital/private equity keeping "hundreds of unicorns" private; (2) tech industry dynamics favoring scale, pushing many VC-backed exits toward trade sales rather than IPOs.

2. Historical Precedent: IPO Waves and Market Downturns

  • Ritter: high new-issue volume predicts lower future market returns, but "it works about 52% of the time" — limited predictive value.

  • Lamont calls issuance waves one of the "four horsemen" of a market bubble; the signal "worked well in 2021," which he calls "a great time to underweight US stocks."

  • Caveat: issuance waves can precede a market top by years, not signal an imminent top — cited the 1990s IPO wave and Japan's bubble, both of which lasted years after the signal appeared.

  • First-day pops (currently ~15-20%, the historical norm) have not reached 1999/2021 magnitudes, suggesting no "speculative euphoria" yet.

3. Debt Issuance and the AI CapEx Buildout

  • Companies are issuing "tons of debt, especially AI-related debt" while still repurchasing equity, per Lamont.

  • Lamont's read: debt being used to fund buybacks signals debt is overpriced relative to equity (or equity is underpriced) — a bullish equity signal for now.

  • Risk flagged: if both debt and equity issuance surge simultaneously, that would signal the whole enterprise value of companies is overpriced — "a negative sign for both credit markets and equity markets."

  • Ritter: companies "have every legitimate non-valuation reason" to issue given "huge CapEx needs" for hyperscalers and new companies alike.

4. Market Capacity to Absorb New Issuance

  • Ritter: US public companies pay out roughly $600 billion/year in cash dividends and have bought back ~$1 trillion/year in stock in recent years — $1.6 trillion of cash recycled annually, dwarfing IPO issuance.

  • Lamont notes supply/demand ultimately governs prices, but historically the market has absorbed 3-5% annual share growth (as in the 1990s) without major price impact.

  • Risk: some large tech companies that were previously big repurchasers have shifted to being net equity issuers in 2026, which could reduce the offsetting effect from buybacks/M&A.

1. Return of the IPO Market Is a Normalization, Not a Boom

Both academics frame 2026 as an "IPO drought" ending rather than a euphoric wave. The record dollar total is driven more by large individual deals than breadth of issuance, meaning the classic bubble warning signs (hundreds of IPOs, huge first-day pops) are largely absent so far.

2. AI Financing Is Shifting from Equity Discipline to Debt-Funded Buybacks

A recurring theme is that companies are tapping debt markets heavily to fund AI infrastructure while continuing to buy back equity — a pattern Lamont reads as bullish for equity valuations today, but warns it could flip negative if equity issuance accelerates alongside debt issuance.

3. Issuance Signals Are Early-Warning, Not Timing Tools

Both guests stress that even if a true IPO wave emerges, it would mark "the beginning of the bubble and not necessarily the end" — investors who short markets on an issuance signal risk exiting years too early, as seen in Japan and the 1990s US tech wave. ---

Sentiment Analysis

Overall Market Sentiment: Cautiously Reassured

Both guests conclude the current IPO resurgence is not yet a red flag for equities, though they flag specific indicators (issuance breadth, first-day pops, debt-vs-equity issuance mix) that would change their view.

Risk Factors Highlighted

Simultaneous debt and equity issuance wave: Lamont warns that if companies begin issuing heavily in both debt and equity at once, it would signal the whole enterprise (not just equity) is overpriced.

Lockup expirations: As IPO lockups expire, insiders and early investors gaining the ability to sell could pressure prices if not offset by buybacks/M&A.

Shift of big tech from net buyers to net equity issuers: Some historically large repurchasers have become net equity issuers in 2026, reducing a key offsetting force against new supply.

Limited predictive power of issuance signals: Relying on IPO volume as a market-timing tool is unreliable (52% accuracy per Ritter), risking false signals in either direction.

AI winner uncertainty: Ritter cautions there's "no guarantee" today's IPO-stage AI companies will be the sector's eventual winners, citing how 1999 IPO cohorts largely missed the internet's biggest gains (won later by Netflix, Google).

Small/unprofitable IPO underperformance: Companies below the ~$100 million revenue threshold have historically underperformed the market after listing, a risk concentrated in speculative, high-growth issuers.

Historical precedent of buyback-plus-issuance waves preceding disappointment: Ritter cites 1990s tech and 1800s railroad buildouts as precedents where heavy issuance plus buybacks preceded weak equity-holder returns.

This episode was covered in today's [The Market Signal — 2026-08-03](https://marketsignal.beehiiv.com/p/the-market-signal-2026-08-03), a cross-source synthesis of multiple podcast reports.

Keep Reading