At the start of the year Renaissance Capital, which has tracked this market since the 1990s, forecast 200 to 230 US listings raising $40 to $60 billion.

The dollars arrived and then some. Proceeds passed $146 billion by early September, a record. The number of companies that priced did not. Through three quarters the count is running roughly a third below last year, and last year itself was only 202.

That gap is the whole story, and it is not a forecasting failure. It is two different markets wearing one number.

SpaceX raised $75 billion in June, and $85.7 billion once the overallotment was exercised. SK hynix raised $26.5 billion in the third quarter. Between them, two listings account for most of the year.

The quarterly detail is where it becomes clear.

Second quarter: 48 listings, $104.8 billion. Renaissance counts SpaceX at its $75 billion base deal rather than the $85.7 billion it reached with the greenshoe, and on that basis roughly $30 billion remains across the other 47 companies — which Renaissance notes would still have been the strongest quarter for proceeds since 2021. Nine other deals raised a billion or more. That is a genuinely healthy quarter.

Third quarter: 30 listings, $32.8 billion. Take SK hynix out and $6.2 billion remains across 29 companies. Eighteen raised $100 million or more.

Two quarters, two different markets. And the third quarter number got worse after it ended — Renaissance revised the count down from 31 to 30 and proceeds from $34.9 billion to $32.8 billion, explicitly because of a string of postponements near the quarter's close. Deals that looked done came apart in the final two weeks.

For the clearest comparison: the third quarter of 2025 saw 64 companies price. The third quarter of 2026 saw 30.

Three companies postponed or pulled in the first quarter. Four in the second. Seven in the third.

They were not from one sector. A smart-ring maker, a nuclear components business, an insurer, a metallurgy company, a brokerage. That matters, because it rules out the easy explanation that one industry fell out of favor.

The most instructive was Oura, which postponed on September 29 having launched a week earlier. The book was reportedly four times oversubscribed. The company is profitable and grew 74% in the nine months to June. It pulled anyway.

When a deal with four times the demand it needs cannot get done, the constraint is not demand. It is price.

The mechanism is unglamorous and it is not sentiment.

The ten-year Treasury reached 5.34% on October 1, its highest since 2002, and in the three months to September it posted its largest quarterly rise since 1994. It closed Friday at 5.24%. Avery Marquez of Renaissance put the consequence plainly: the recent spike in yields and resumed rate hikes have put "some downward pressure on growth companies" — and "that's a majority of the companies in the IPO market."

The arithmetic is familiar to anyone who has built a discounted cash flow. Raise the risk-free rate and you raise the discount rate. Raise the discount rate and cash flows arriving in year seven are worth materially less today than they were in January. A company whose value sits mostly in distant growth loses more than a company earning now.

Which is why this is durable rather than a mood. It does not reverse because confidence returns. It reverses when yields fall.

Worth naming the obvious objection. A good part of this move is energy — Brent is around $106, up more than 60% on a year ago — and energy shocks can unwind quickly. A reader might reasonably conclude the discount rate comes back down on its own.

The market disagrees. It is pricing further tightening and no cut at any meeting through October 2027, which is not what gets priced when an inflation impulse is expected to fade. And no one builds a 2027 capital plan on an assumption they cannot underwrite.

Here is the part that should change how you read this market.

The conventional test for whether an IPO can price is the VIX. Below 20 and equity volatility is considered contained enough to get a deal done; well above it and the window is generally treated as shut. It is the number bankers quote and boards ask about.

The VIX closed at 14.84 on Friday — within six-tenths of a point of its low for the year, set in mid-August at 14.25. Cboe put September in a 14-to-18 range. Through the entire quarter in which seven companies postponed or withdrew, equity volatility sat comfortably in the mid-teens. In June, General Atlantic wrote that at those levels a moderate uptick was unlikely to close the equity issuance window at all.

They were not wrong about the VIX. The VIX simply stopped being the constraint.

This is worth sitting with, because it explains why 2026 does not feel like 2022. That was a volatility shutdown — the VIX in the thirties, nothing could price, and everyone understood why. This is a discount rate shutdown happening in calm equity markets, which is a far more confusing thing to live through. It is how a company can be four times oversubscribed and still walk away. The demand is real. The price that demand will pay, at a risk-free rate above 5.2%, is not the price the seller had in mind.

If you want a gauge that matches the actual constraint, it is not the VIX. It is Treasury volatility.

The Renaissance IPO Index returned 15.9% year to date, ahead of the S&P 500's 12.7%. Read quickly, that says the companies which listed this year are trading well.

The quarter says otherwise. The index fell 11.5% in the third quarter while the S&P rose 2.3%, and the class that priced in those three months returned −0.1% from offer — a 6% first-day gain handed straight back in the aftermarket. Barely half closed the quarter above their issue price. Thirty-nine percent broke on the first day.

Set that against the start of the year and the direction is unmistakable. Average return from offer ran 27.9% in the first quarter, 10.8% in the second, and nothing in the third. The year-to-date number is carried almost entirely by deals that priced before June.

One split inside that data matters more than the average. Private-equity-backed IPOs returned −9% from offer in the quarter, and −15% once SK hynix is stripped out. Venture-backed IPOs returned +6%. Size offered no protection either: the ten largest deals of the quarter averaged −6%, with only four of ten finishing above issue.

So the pricing standoff is not seller stubbornness. Buyers have a quarter of evidence that the recent cohort did not reward them, and that sponsor-backed deals rewarded them least. That is the real gate on the queue — not volatility, not sentiment, but whether the last class traded well enough for the next one to price.

It also reframes the question for anyone planning a raise. You are not asking whether investors want new companies. You are asking how the five that listed before you have traded, because that is what your bookrunner will be working against.

And it is not uniform. Healthcare averaged +16.6% from offer across nine deals, almost all of them biotech. Industrials averaged −32.5% across four deals. Technology was flat. Consumer discretionary and energy were worse still, at −23.3% and −21.3%, though on two deals and one respectively. The window is not open or shut; it is open by sector, and the ordering is not the one most plans assume.

Databricks delayed its listing to 2027, with its chief executive telling Bloomberg that 2026 was a terrible year to go public. It instead raised $5 billion privately at a $190 billion valuation, disclosed in August, against a prior mark of $134 billion — a 42% increase.

Read that carefully. A company chose a private round at a higher valuation over a public listing.

That is the competitive dynamic underneath everything else. Private capital is currently paying more than public markets will, for the same assets. The IPO window has not simply closed — it has lost a bidding contest.

It does not work for everyone. Waiting is only free if someone will fund the wait at a price you like — Kraken was repriced from $20 billion to around $13.3 billion along the way.

Enterprise buyers won’t put your product near their customer data without a SOC 2 report. Twenty-one state privacy laws now hold them accountable for the vendors they share data with, so their diligence lands on you.

Sprinto gets you audit ready in 14 days, across three working sessions. AI agents connect your stack, collect the evidence auditors ask for, and close the gaps as they appear. You approve, they execute.

Sprinto also answers the security questionnaires your buyers send, using the same evidence base, so security reviews stop holding up deals.

No compliance hire. No consultant. Your auditor signs off.

Learn more and get $1,000 off.

PitchBook's third-quarter analysis counts 110 companies that looked ready to list in 2023 and are still on the sidelines. Three years. They describe it as venture's liquidity paradox: exit value running above trend while distributions back to funds sit at 7.9% of net asset value against a long-run average of 14.5%.

What is on file or openly expected for the next twelve to eighteen months sorts into five groups.

AI and its infrastructure. Anthropic filed confidentially on June 1 and has not yet filed publicly; bankers have discussed a valuation anywhere from $1.8 to $3 trillion against $965 billion in May, and marketing is reported as possible from the week of November 9, which implies a public filing around October 23. DayOne Data Centers filed on October 5 for an estimated $3 billion. OpenAI has moved to 2027. Databricks is 2027. Cohere, SambaNova and Perplexity are later still.

Aerospace and defense. Anduril at around $61 billion with no filing. Shield AI and Saronic in the same position.

Energy and deep tech. Quantinuum. Standard Nuclear filed in July. Holtec withdrew.

Private-equity-backed assets, the largest group and the least discussed. Dealogic counts 23 North American sponsor-backed IPOs worth $21.5 billion through September 29, against 14 worth $8.2 billion a year earlier — volume more than doubled. Queued behind that: Wella, owned by KKR, and Aggreko, owned by TDR Capital and I Squared, have both filed publicly. Copeland is a possible year-end candidate. Mavis Tire moved to 2027 after acquiring Pep Boys, and Cumberland Farms may follow. CVC's Bamboo Insurance postponed its September listing on valuation pressure. On the software side, Thoma Bravo's Proofpoint, Genesys under Permira and Hellman & Friedman, and UKG are profitable and well run but grow more slowly than the AI names, and need an answer on what AI does to their own models before they can price.

Everything else, which is the problem. Stripe at $159 billion after February's tender offer, and in no hurry. Canva at $42 billion, moved to 2027. Kraken, repriced. Discord, filed confidentially in January. Strava, filed confidentially, targeting spring. Plaid, exploring. Revolut.

Look at what is missing. Renaissance's own description of the second quarter named the sectors that got away — aerospace and defense, AI infrastructure, energy, biotech — and then noted that venture-backed tech largely remained absent. Enterprise software, which has been the backbone of the IPO pipeline for two decades, has barely filed at all.

For sponsor-owned companies there is a gate that has nothing to do with sentiment. Nick Williams, head of US equity capital markets at Deutsche Bank, told Mergermarket that a viable candidate has to have deleveraged far enough that the listing does not require an excessively large primary issuance — and that a significant number of assets stay out of reach until late 2027 or beyond for that reason alone. With the ten-year around 5.25%, leverage and maturity walls are live topics again, and what passed for an optimal capital structure in 2021 does not now.

That is a balance-sheet test rather than a market test, which makes it the one a CFO can run tonight. If the primary proceeds needed to bring leverage to a level public investors will accept exceed what the float can reasonably bear, the company is not waiting for a window. It is waiting for its debt to come down.

There is a second restraint that is easy to miss, because it is not about the deal in front of them. A sponsor holding several assets that will eventually need public markets is not pricing one listing — it is protecting the reception of the next four. "You don't want one underperforming transaction to undermine your credibility when bringing future assets to market," Williams said. Cleary Gottlieb's Lesley Janzen makes the same point from the other side: right-sizing valuation expectations is what produces the post-listing appreciation that matters for later sell-downs.

A corporate issuer carries no such constraint, and that asymmetry is worth understanding. Sponsor-backed deals ahead of you in the queue may hold back for reasons that have nothing to do with your own prospects — and their restraint is part of why the calendar looks thinner than the pipeline is.

It is open, and it has been open all year, for a specific kind of company.

Large. Comprehensible. Earning now, or in a sector where investors have already decided they want exposure regardless. Able to price where the buyer wants rather than where the last private round implies. And ideally not actually in need of the money.

For a venture-backed software company at a 2021 valuation with four years of accumulated preference stacked above the common, it has not been open at all. Not in the first quarter, not in the second when everything else was working, and not now.

Our take: the useful question for the next eighteen months is not whether the window opens. It is whether your company resembles the ones that got out this year or the ones that did not. That test is uncomfortable but it is answerable today: can you price where buyers are rather than where your last round was, do you earn money now, and could you walk away if the answer came back wrong. Three yeses and the window is open to you. One no and you are in a queue of 110 that has not moved in three years.

Four things every CFO needs to know this week.

· Earnings season opens Tuesday with the big banks, and expectations are extraordinary. FactSet put the S&P 500's estimated third-quarter growth rate at 29.6% on Friday, and on the five-year average rate of beats expects the actual figure to land above 35%. Hold that against the analysis above: public investors can buy thirty percent earnings growth in companies they already own, which is a large part of why a new issue has to price where the buyer wants.

· M&A had its first sub-trillion quarter since mid-2025. Global deal value fell 41% from the second quarter to $993 billion, with ten deals above $10 billion — the fewest megadeals since late 2024. Year to date still stands at $3.9 trillion, up 28% and the strongest since 2001, but deal count fell 8% to a level last seen in 2020. Clifford Chance's Sarah Jones called it "a normalization rather than an end of a cycle." Fewer deals, larger, financed at a long end that just posted its biggest quarterly rise since 1994.

· The SEC moved to widen who can invest in private markets. Release 2026-96 proposes changes to performance-fee eligibility, performance-compensation disclosure, interval funds and multi-class closed-end funds — and separately asks for comment on new pathways to accredited investor status, including a FINRA exam and credentials such as the CPA, CFA and CFP. The pathways are a request for comment rather than a proposed rule, which is the distinction worth holding onto. The second-order effect is the one that matters here: a deeper private bid is exactly what has been outbidding the public market all year.

· THE RATE PATH — the hold everyone expects is a pause, not a peak. Markets put the odds of no change on October 27–28 at about 82%, and there is no Summary of Economic Projections at this meeting, so the next dot plot is December 8–9. December is the one to model. Futures price roughly a 79% chance the policy rate is higher by then, and about 81 basis points — three more hikes — across the next nine meetings. The probability of a cut at any meeting between now and October 2027 is currently zero. If your 2027 plan has rates coming down, it is not modeling the market's view. It is modeling a reversal the market has stopped pricing at all.

1. What is your last round worth today at current yields? Rebuild it at a discount rate reflecting a ten-year around 5.25% rather than the one in the original model. For a company whose value sits in later-year cash flows the difference is not marginal. That number, not the last round, is your realistic starting point.

2. Can you price where buyers are? The deals getting done are priced where the market wants them. If your board or your preference stack makes that impossible, you do not have a timing problem — you have a structural one, and waiting does not solve it.

3. Is private capital cheaper than public right now? For some companies this year it has been, visibly. Price both. If a private round clears at a better valuation with less disclosure and no lockup, that is not a fallback — it is the better option, and it should be assessed rather than assumed.

4. Could you walk away? Companies that got out this year could mostly afford not to. If you cannot fund another eighteen months, your window is defined by your runway rather than by the market, and that is worth knowing before you start rather than during.

SPONSOR THIS NEWSLETTER

Your name, in front of the CFOs and senior finance leaders who read The CFO Desk every Tuesday. The primary slot is available for Q4. Write to [email protected].

A record year in which the record belongs to two companies is not really a record year. It is a very good year for a small number of businesses and an ordinary one for everybody else, reported as a single number because that is how the number is reported.

The queue behind it is not clearing. It is being repriced while it waits, and in at least one case being paid more to stay private.

Next week we turn to the other exit. M&A has just had its first sub-trillion quarter in over a year while the year to date runs at its strongest since 2001 — fewer deals, much larger, and financed at a long end nobody modeled. One banker's summary of the sponsor problem is worth sitting with until then: you cannot sell everything to a strategic, you cannot put everything into a continuation vehicle, and you cannot sell a minority stake in everything. Some of it has to find a buyer.

One question before you go: if a listing is in your plan, what is it waiting for?

One click. And if you have a view on why, reply — we read everything.

The Editors hold no position in any company mentioned.

— The CFO Desk
[email protected] · thecfodesk.com
CAPITAL · FINANCE · LEADERSHIP