$1.4 Trillion in Buybacks Meets a Record Issuance Wave — and Half the 2026 IPO Class Is Already Underwater

The Supply-Demand Balance Is Shifting

The headline number from Goldman Sachs is straightforward: U.S. companies will repurchase approximately $1.4 trillion of shares during 2026, enough to offset roughly $700 billion in primary equity issuance plus additional supply from expiring IPO lockups.^[1] Buybacks are running at about twice the estimated level of primary issuance — corporate America remains a net buyer of its own stock.

But beneath that net-positive framing, the composition is shifting. Total equity issuance — IPOs, follow-ons, convertibles, and SPACs — reached a record $252 billion in the second quarter, surpassing the prior quarterly high of $234 billion from Q1 2021.^[1] U.S. companies raised $105 billion through follow-on offerings alone in the year through July, the strongest pace at this stage of any calendar year since 2021.^[1] Goldman’s strategists, led by Ben Snider, characterize this as “a return to normal rather than a boom,” noting that both transaction counts and issuance relative to total equity-market capitalization remain below long-term averages.^[1]

What would have to be true for each side? For the optimists: buyback authorizations hit a record of nearly $1 trillion year-to-date, S&P 500 repurchase growth is running at roughly 11% year over year, and Goldman finds “no abnormal sign of indigestion” in offering discounts or post-pricing performance.^[1] For the skeptics: issuance is accelerating, AI-related transactions already account for about 40% of U.S. follow-on equity issuance, and consensus forecasts put hyperscaler capital expenditure at $1.1 trillion in 2027 — exceeding operating cash flow by roughly $150 billion before those companies are expected to return to free-cash-flow positivity in 2028.^[1] If capex runs above consensus, as many investors expect, the equity-financing requirement grows.

The IPO Pipeline: Defense Tech, Biotech, and Chinese Listings

The U.S. IPO calendar this week is thin but telling. Renaissance Capital lists two deals: Lyntris (LYNX), a defense-technology roll-up backed by Trive Capital, and Advasa Holdings (ADBT).^[2]

Lyntris is the larger story. The Falls Church, Virginia-based company plans to raise $492 million by offering 24 million shares at $19 to $22, targeting a market value of roughly $2.4 billion at the midpoint.^[3] But the structure warrants attention: nearly 80% of the offering — 19.1 million of the 24 million shares — comes from selling stockholders rather than the company’s treasury.^[3] That means most of the capital is an exit for existing backers, not fresh fuel for operations. The company was established in 2026 as a roll-up of defense-tech connectivity businesses providing sensor architecture, hardware, and data platforms for the U.S. Department of Defense and allied militaries.^[3]

Vogenx (VOGX), which priced last week, offers a different profile. The clinical-stage biopharmaceutical company sold 6.25 million shares at $13 — the high end of its range — raising approximately $81.3 million in gross proceeds.^[4] The Raleigh, North Carolina-based company is developing treatments for metabolic disorders including post-bariatric hypoglycemia, a condition with no FDA-approved therapies.^[4] Biotech IPOs that price at the high end signal genuine demand for clinical-stage risk, but the proof comes in the aftermarket.

Meanwhile, Chinese listings continue to populate the global pipeline. On August 11, three Chinese issuers debuted with striking first-day gains: Chengdu Ultra Pure Applied Materials closed 662% above its offer price on the Shenzhen ChiNext board, CIQTEK rose 419% on the Shanghai STAR Market, and JAKA Biotech advanced 139% on the Beijing Stock Exchange.^[5] Londian Wason New Energy Tech (FOIL) priced an upsized NYSE offering of roughly 4.3 million ADSs at $22, raising approximately $94.3 million.^[5] These first-day pops are notable, but as Public Markets Review observes, they reflect price discovery against a low offer price — not necessarily durable demand. Turnover, public float, and allocation concentration all matter more than the headline percentage.^[5]

The Aftermarket Problem

Here is where the balanced case meets an inconvenient data point: roughly half of the 203 companies that have gone public on the NYSE or Nasdaq in 2026 now trade below their offering price.^[6] Global IPO proceeds reached approximately $190.9 billion across 530 deals in the first half of 2026, already topping the full year of 2025.^[7] But volume is not the same as performance.

SpaceX, the year’s marquee listing, has fallen roughly 49% from its post-listing peak.^[6] The slow decline from its debut highs — crossing below its IPO price within about a month — has become a reference point for the broader newly-public universe.^[6] The pattern is consistent with a market where scarcity-driven first-day pops fail to hold once the initial float rotates into a broader holder base.

What would have to be true for the aftermarket to improve? Lower deal cadence allowing each IPO to absorb more investor attention, companies pricing with more discipline rather than maximizing the first-day pop, and a stable macro backdrop that keeps risk appetite intact. None of those are guaranteed in the second half.

The Overnight Plumbing Changes

On August 5, 2026, the SEC approved Amendment No. 27 to the Limit Up-Limit Down (LULD) Plan, establishing temporary price-band protections for overnight trading.^[8] The move comes as exchanges prepare for 23-hour, five-day-a-week trading sessions — an expansion into periods historically characterized by lower liquidity, wider spreads, and heightened price volatility.^[8]

The framework, set to take effect December 6, 2026, works as follows:^[8]

This is a measured first step. The 20% bands without automatic pauses mirror protections already used by overnight ATSs, which should be familiar to current overnight traders.^[8] The two reference prices — including the 7:45 p.m. consolidated last sale — account for post-close news and trading.^[8] The SEC’s explicit expectation is that Phase 1 data will inform Phase 2 revisions.^[8]

The design raises a question worth tracking: 20% is a wide band. For a stock that normally moves 1-2% overnight, it provides substantial room for price discovery. But for a small-float newly-public stock or a high-beta AI name — exactly the securities most likely to see overnight activity — 20% may still permit the kind of volatility that prompts calls for tighter guardrails.

The Liquidity Warning Nobody Expected

The unwind of Leopold Aschenbrenner’s Situational Awareness hedge fund is a market-structure event that bears watching, even though its direct market impact was contained. The fund, which grew to as much as $45 billion in assets, was forced to sell all of its public stock holdings after steep losses in AI infrastructure names including SK Hynix and CoreWeave triggered margin calls.^[9] Citadel, Ken Griffin’s hedge fund, reached a deal to buy the stock portfolio.^[9] The fund’s assets reportedly fell to around $10 billion.^[9]

Two things stand out. First, the unwind of a leveraged book reportedly as large as $16 billion in public equities produced “barely a ripple beyond a few days of volatility,” according to Deutsche Bank Research.^[10] That is a testament to market depth — the system absorbed a forced seller without a cascading dislocation. Second, the event exposed how concentrated and leveraged the AI trade had become. When a single fund’s deleveraging can be absorbed without contagion, it tells you the plumbing held. It also tells you the plumbing was tested.

Citadel Securities, in its August market note, described the period as one of the “most technically challenging trading environments” of the year, with positioning having since normalized.^[10] Their view is that fundamentals matter again after a reset in levered positioning.^[10] Bloomberg Law’s reporting adds nuance: while the S&P 500’s volatility curve shows traders pricing daily swings of less than 0.8% for the rest of August, options positioning reveals how quickly sentiment can switch between fear and greed.^[10]

What to Watch Next

• Lyntris (LYNX) pricing and first trade. A $492 million defense-tech deal that is 80% secondary tests whether investors are willing to buy existing holders’ exits at a $2.4 billion valuation. Watch the discount to range and first-week turnover.

• The buyback-issuance gap. Goldman’s framework has buybacks at roughly $1.4 trillion versus ~$700 billion in primary issuance.^[1] If hyperscaler capex beats consensus — and many investors believe it will — the issuance side of that equation grows. Track follow-on filing volume in September.

• Overnight price band implementation. Phase 1 takes effect December 6, 2026.^[8] The first quarter of overnight trading data will shape Phase 2 rules. Any stocks that hit the 20% bands in the early weeks will become the test cases.

• 2026 IPO aftermarket breadth. With roughly half the class underwater,^[6] the question is whether the second-half IPOs price with more discipline — or whether the pipeline thins as issuers wait for better conditions.

• Anthropic IPO watch. The Situational Awareness fund retained its private Anthropic stake through the unwind,^[9] and ION Analytics reports growing fears around the Anthropic IPO amid increasing Chinese competition.^[6] The next mega-IPO will be a real test of whether AI-linked issuance still has legs.

This article is research commentary, not investment advice. All figures are sourced from the references cited above.