Goldman expects 100 IPOs totalling $160 billion will come to market in 2026

Goldman Sachs now expects 100 IPOs totaling $160 billion in 2026, versus a prior 120-IPO forecast, as volatility and geopolitical uncertainty weigh on issuance. The average 2026 IPO is up 19% on its first trading day, and 53 companies have filed to go public year to date, more than double last year’s comparable period. The article is broadly constructive on the IPO pipeline, though it flags software-heavy backlog risk and emphasizes mega-IPO uncertainty, including SpaceX’s potential $75 billion debut.

Analysis

The first-order read is that the market is rewarding scarcity and complexity: when IPO windows open, capital crowds into the names with the cleanest growth narratives and the shortest path to profitability. That setup favors the “picks-and-shovels” beneficiaries more reliably than the largest prospective listings themselves, because the real monetization often comes from the ecosystem trade around underwriting, execution, hedging, and index inclusion rather than from the IPO event alone. Goldman’s data also implies a supply/demand imbalance in issuance: if the pipeline stays heavy but software-heavy backlog keeps slipping, the market may continue to bid up a narrower set of high-quality tech/AI exposure while punishing lower-quality issuers on debut.

The bigger second-order effect is that a successful mega-IPO calendar can become a liquidity drain on adjacent high-beta tech. New issuance tends to absorb marginal risk capital for several weeks, which can pressure secondary multiples even when the headline market is bullish. That makes the current enthusiasm for AI-linked memory and compute names more interesting: these are the likely operating leverage beneficiaries if the market keeps paying up for AI infrastructure, but they are also the most vulnerable if sentiment rotates from “own the enablers” to “fund the new issue.”

The contrarian miss is that strong first-day pops do not necessarily validate long-only ownership of the IPO class; they often signal that private-market pricing is still too conservative and that public-market demand is chasing a near-term squeeze. In that environment, the highest expected value trade may be to own the infrastructure and execution beneficiaries while fading late-cycle enthusiasm in the most crowded momentum names. For the names in scope, SMCI has the clearest second-order upside from AI capex intensity, APP is the cleaner expression of monetized ad-tech AI momentum, and GS is a more defensive way to own the issuance and deal-flow uptick without taking single-name IPO risk.

Key Decisions for Investors

• Long SMCI on dips over the next 1-3 weeks; use it as a levered proxy for continued AI compute capex and meme-like momentum in high-beta tech. Risk/reward improves if the market keeps rewarding infrastructure over application-layer names; cut if IPO demand starts pulling liquidity out of crowded AI semis/hardware.
• Pair trade: long GS / short a basket of lower-quality recent IPOs or unprofitable software issuers for a 1-2 month horizon. The thesis is that underwriting and advisory volumes should benefit from an active issuance window while weaker new listings remain vulnerable to post-debut mean reversion.
• Long APP into any broad tech weakness over the next 2-6 weeks. APP is a cleaner expression of AI-driven monetization than pure story stocks, and it should outperform if capital rotates toward names with visible cash generation and operating leverage.
• Use call spreads on SMCI rather than outright shares for the next earnings/AI-capex catalyst cycle. The stock can overshoot on momentum, but defined risk is preferable because new-issue supply or a rotation out of crowded AI trades can compress multiples quickly.
• If IPO enthusiasm accelerates further, reduce exposure to the most crowded high-beta tech winners and rotate toward GS as a volatility buffer. The risk is not a market crash but capital dilution across new issuance, which can cap upside in crowded longs for several weeks.