The Issuance Pipeline Is Running Hot — and August Brings the First Real Stress Test

The underwriting window is wide open, and the numbers coming out of Wall Street’s second-quarter earnings season tell the story plainly. Goldman Sachs, JPMorgan Chase, Citigroup, and Bank of America all reported blockbuster Q2 underwriting revenue, with Goldman’s David Solomon noting that the bank’s backlog had climbed to its highest level in five years and its second-highest on record. JPMorgan CFO Jeremy Barnum described the environment as self-reinforcing: “The generally robust environment is itself begetting more activity.”^[1]

The drivers are no secret. Alphabet, Nvidia, SpaceX, and a wave of AI-adjacent companies are tapping public markets at a pace not seen since the late 1990s. UBS estimates IPO issuance could reach $200–350 billion this year, with secondary offerings potentially exceeding $400 billion — both record highs in absolute dollar terms.^[2]

That sounds alarming on its face. The question is whether the market can digest it.

The buyback offset: why gross issuance overstates the risk

The headline number is gross. What matters for share prices is net equity supply — issuance minus buybacks. Here the picture is more reassuring than the deal-flow headlines suggest.

UBS calculates that US share buybacks totaled $1.2 trillion over the trailing 12 months and expects that pace to hold through year-end. Goldman Sachs is more bullish, projecting $1.3 trillion in corporate repurchases for 2026 against $1.1 trillion in total equity issuance.^[3] If those estimates hold, the corporate sector will retire slightly more stock than it creates — net supply stays negative, and the dilution math flips in investors’ favor.^[2]

There is also a normalizing argument. Measured against total US equity market capitalization of roughly $72 trillion, projected 2026 issuance is only slightly above the long-term average as a percentage of the Russell 3000 free float.^[2] In other words, the market has grown into the deal flow. The dollar records are real, but the relative burden is closer to ordinary.

Csquare: the AI-infrastructure deal that priced below expectations

Not every deal is clearing at the terms issuers hoped for. Csquare, a Brookfield-backed North American data-center operator, priced its IPO on July 16 at $21 per share — 50 million shares for $1.05 billion in gross proceeds, with a 30-day underwriter option for 7.5 million additional shares.^[4] The deal was one of the largest US IPOs of the year by deal size, but reports noted it fell short of its share-price and valuation targets, and proceeds are earmarked primarily for debt repayment rather than growth.^[4]

That is a useful data point. When even a well-backed infrastructure name in the AI build-out theme prices below its range, it tells you the marginal buyer is not infinitely elastic. The pipeline is full, but price sensitivity is creeping in.

SpaceX’s staggered lockup: $123 billion hitting the market in tranches

The single largest overhang on the horizon is SpaceX. The company went public around June 11, 2026, at $135 per share on Nasdaq under the ticker SPCX, pushing its valuation past $2 trillion — the largest IPO in US history.^[5]

Less than 5% of SpaceX’s total shares were released into public float at launch. Rather than the standard 180-day lockup, SpaceX implemented a staggered release: tranches of roughly 7% unlock at days 70, 90, 105, 120, and 135 after the IPO.^[5]

The first major wave — unlocking 20–30% of total shares, approximately 911.5 million shares valued at roughly $123 billion at recent prices — lands in early August, roughly two trading days after the company reports Q2 earnings.^[5] A second large tranche of about 28% follows after Q3 earnings. By December 8, 2026, when the full 180-day period concludes, roughly 40% of all SpaceX shares will be freely tradable. Elon Musk’s personal holdings carry a 366-day lockup, meaning he cannot sell until approximately June 2027.^[5]

The supply-demand math here is genuinely uncomfortable. SPCX is already trading in the $124–$135 range as of mid-July, below its IPO price.^[5] The first major unlock lands simultaneously with Q2 earnings — meaning the stock will be digesting financial results and absorbing a massive supply increase at the same time. Historical analogs from the dot-com era, where lockup expirations flooded the market with insider supply after a richly priced IPO wave, are not encouraging.^[1]

SpaceX’s staggered unlock means roughly $123B in insider shares begins hitting the market in August — the first real liquidity test of the AI-era IPO cycle.

Market microstructure: the plumbing is different this time

A structural factor that did not exist in prior boom-bust cycles is worth flagging. Dark pool volume now accounts for approximately 42–45% of total US equities trading, up from roughly 14–16% in 2016 — a 240% increase in market share. The Federal Reserve’s 2025 Financial Stability Report documented this shift as one of the three most significant changes to US market architecture since 2008.^[6]

JPMorgan internal research, shared with select institutional clients, reportedly shows that 34% of daily volume in major S&P 500 constituents never appears on public order books.^[6] The execution timeline for a $50 million institutional trade has compressed from 15–45 minutes in 2016 to 2–8 minutes today, but with trade-offs in price discovery.^[6]

Roughly 42–45% of US equities volume now trades in dark pools, up from 14–16% a decade ago — reshaping how the market absorbs large supply shocks.

What this means for August’s lockup unlock is two-sided. On one hand, dark pools and algorithmic execution can absorb large block trades without immediately pressuring displayed prices — potentially smoothing the impact of insider selling. On the other hand, when price discovery is this fragmented, the visible tape may understate the actual selling pressure until it is too late for retail participants to react. The two-tier market structure that worked fine during the bull run has not been stress-tested against a concentrated, multi-hundred-billion-dollar supply event.

Putting the odds together

The base case, drawing on UBS and Goldman Sachs projections, is that net equity supply stays negative for 2026 as a whole — buybacks at $1.3T exceed issuance at $1.1T, and the market absorbs the deal flow without a meaningful valuation shock.^[2][3] Academic literature finds no consistent relationship between IPO activity levels and forward market returns; issuance is better understood as a coincident indicator of strong conditions than a leading indicator of weakness.^[2]

I would put the probability of that base case at roughly 60%. The 40% tail is the scenario where the SpaceX unlock, combined with a softening Csquare-type deal environment and continued secondary offerings from AI capital-intensive companies, creates a localized supply shock that drags the most AI-exposed names and, by extension, market sentiment. The historical analog — the 1999–2000 lockup-expiration wave — argues that timing matters more than aggregate size. A concentrated unlock hitting a stock that is already below its IPO price, into a fragmented liquidity environment, is a meaningfully different event than a smooth pipeline of well-priced deals clearing at premium valuations.

The next four to six weeks are the inflection point. If SPCX holds above its $135 IPO price through the first tranche and Q2 earnings, the base case strengthens considerably. If it does not, the signaling effect on the broader AI-issuance pipeline — and on the banks whose backlogs are at five-year highs — will be immediate.

What to watch next

• SpaceX Q2 earnings and first lockup tranche (early August): The timing overlap of results and the 20–30% share unlock is the single most important near-term data point for the issuance narrative.
• Csquare (CSQR) post-IPO trading: Whether the stock stabilizes or drifts below its $21 offer price will signal how much price sensitivity exists at the margin for AI-infrastructure deals.
• Secondary offering pace: REGENXBIO announced a $100M follow-on on July 16.^[7] The cadence of secondary filings over the next several weeks — especially from AI capital-intensive names — will indicate whether the pipeline is accelerating or plateauing.
• Buyback execution: Goldman’s $1.3T buyback estimate assumes corporates follow through. Q2 buyback announcements from mega-cap names will either confirm or erode the net-supply-negative thesis.
• Dark pool volume share: If the 42–45% figure continues climbing, price-discovery fragmentation deepens, raising the stakes for any concentrated supply event.