NextFin News - The smart-ring maker Oura had investors lining up: its order book was roughly four times oversubscribed, it was profitable, and revenue was forecast to jump 90% this year. Yet on September 29 it pulled its planned $2.2 billion Nasdaq listing anyway, joining nuclear-services firm Holtec Nuclear and insurer Bamboo Insurance in a late-September exodus that has turned Wall Street's eagerly awaited fall IPO window into a waiting room. The question is no longer whether companies can find buyers. It is whether they can get the price they came for.
The Window Is Open, but the Door Has Narrowed
The sequence of withdrawals reads like a roll call of the market's new selectivity. Oura, the Finnish-founded wearable-health company, had marketed 50 million shares in a $40-to-$44 range that would have raised up to $2.2 billion at a fully diluted valuation of $15.62 billion. Holtec Nuclear suspended its roughly $900 million offering on September 16 amid heightened scrutiny of the capital pouring into AI-adjacent sectors, then withdrew its registration statement on September 25. Bamboo Insurance Services, backed by CVC Capital Partners and White Mountains, postponed a roughly $700 million IPO on September 22, a day before it was set to price at a $3.24 billion valuation on $18-to-$20 shares.
The timing is the point. September is traditionally one of the strongest months for new listings, yet only three companies completed US IPO debuts after the Labor Day holiday on September 7. Through mid-September, 109 offerings had raised $146.5 billion this year, but that headline is distorted by concentration: a single listing, SpaceX, accounted for roughly 70% of all 2026 IPO proceeds to date, and just over 100 companies have gone public in the US this year, well below the 15-year annual average of more than 200.
The pullbacks are not confined to the US. EG Group, the UK petrol-station and convenience-store operator, delayed a planned New York IPO until 2027, abandoning an attempt to raise around $1 billion at an approximately $9 billion valuation as takeover interest from infrastructure investor Stonepeak emerged. In London, Airtel Money priced a 529 million-pound offering at a $7 billion valuation, a deal that an IPOX Schuster analyst described as evidence that issuers must now remain realistic about valuation expectations.
The market's own barometer agrees. The Renaissance IPO ETF, which tracks newly listed companies, is down about 10% from its late-June high near $59.42 even as the major indexes sit at records. And the aftermarket has been unforgiving: five of the ten largest US listings this year trade below their IPO prices, and the year's offerings have returned a weighted average 13%, below the S&P 500's roughly 14% gain. Issuers are not fleeing an empty market. They are fleeing a market that is paying less than they hoped.
Why the Rates Shock Hit IPOs First
The trigger is a repricing of the risk-free rate that underpins every valuation model on Wall Street. On September 16, the Federal Open Market Committee voted unanimously, 12-0, to raise the federal funds target by 25 basis points to 3.75%-4.00%, the first rate increase since July 2023, and its dot plot signalled one more quarter-point hike before year-end. The bond market then did the rest: the 10-year Treasury yield touched an intraday 5.344% on September 30, its highest level since April 2002, while the 30-year bond reached levels not seen in 24 years.
That matters disproportionately for IPO candidates. A company going public is, by definition, a claim on cash flows that sit further in the future than those of an established blue chip. When the discount rate rises, the present value of those distant cash flows falls fastest. "A rising rate environment hits the value of future cash flows and you have the higher cost of debt to do things like build data centers, so it is kind of a double whammy," said Matt Kennedy, a senior strategist at Renaissance Capital.
The mechanism runs through two channels at once. First, the math: for a long-duration growth asset, a 100-basis-point rise in the discount rate can erase roughly a fifth or more of present value, because the bulk of the asset's worth sits in profits expected years away. Second, the plumbing: many of the companies dominating this year's pipeline, from AI infrastructure to data centers to nuclear services, are themselves capital-intensive, and higher borrowing costs compress the returns on the very projects that justified their valuations. The same force that shrinks the price investors will pay also raises the cost of the growth the issuer must deliver to justify it.
There is also a sequencing problem. IPOs are priced on a timetable, with books opening, demand gauged and the deal struck, while the macro backdrop can move inside that window. Oura launched its offering and, eight days later, postponed it before pricing. In that span the 10-year yield climbed more than 17 basis points in a single week, and the MOVE index of Treasury volatility jumped from 80 to 104 before settling at 96. A deal that looked financeable on Monday can look expensive by Friday.
The AI Valuation Edifice Meets a Higher Discount Rate
Beneath the rate move lies a second, more uncomfortable question: what are AI-related assets worth when money is no longer free? The 2026 IPO revival was built largely on the promise of artificial intelligence, and the concentration is stark. SpaceX's dominance of proceeds is one symptom; the pipeline is another. Anthropic is advancing what could become one of the largest offerings ever, with formal marketing potentially beginning the week of November 9 and a valuation discussed around $2 trillion. OpenAI, by contrast, has ruled out a 2026 listing, its chief executive saying important safety work remains.
The tension between those two stories defines the moment. On one side, a company investors may underwrite at a $2 trillion valuation on the strength of an AI infrastructure buildout that itself depends on cheap debt. On the other, a retreat by issuers whose valuations were calibrated in a near-zero-rate world and who now find the arithmetic does not clear. Holtec Nuclear's withdrawal, explicitly tied to scrutiny of capital deployed in AI-related sectors, is the clearest signal that the trade has moved from "any AI exposure earns a premium" to "show me the returns at these financing costs."
The private-market overhang makes the public window harder to open. Bamboo's backers, CVC and White Mountains, had bought their stake at a $1.75 billion valuation less than a year before the postponed offering; the planned $3.24 billion pricing would have been an 85% markup in under twelve months. Blockchain.com is pitching investors on a roughly $500 million IPO at a $4 billion-to-$6 billion valuation, well below its $14 billion private-market peak in 2022. For issuers, the choice is binary: accept a down round in public, or stay private and wait. Most are choosing to wait.
"While the major indices are at or near all-time highs, the breadth of the rally and the underlying trading environment have been less consistent," said West Riggs, head of equity capital markets at Truist Financial Corp. "There's also a lot for markets to digest between geopolitics, oil and the Fed, so companies with flexibility are choosing not to force a transaction into periods of heightened volatility."
That last clause, "companies with flexibility," is the hinge of the whole episode. Oura can afford to wait because it is profitable and growing. Holtec can wait because private capital and strategic interest remain available. The issuers who cannot wait are the ones with near-term maturities, burn rates tied to expensive debt, or investors who need liquidity. The freeze, in other words, is selective, which is what makes it more damaging than a blanket shutdown. It separates the haves from the have-nots, and the have-nots are the ones most likely to become forced sellers.
Cyclical Freeze or Structural Repricing?
Is this a temporary pause or a regime change? The answer is both, and confusing the two is the most common error investors make in moments like this.
The cyclical case is strong and concrete. The window slammed shut after a discrete shock, the September 16 rate hike and the Treasury selloff that followed, layered onto event risk from geopolitics and oil. Volatility has been concentrated in bonds, not stocks: the VIX finished the week ended September 25 flat at 14.87 even as the S&P 500 rose 1.2% and the Nasdaq Composite gained 2.1%. Equity indices at or near records are not a backdrop in which capital markets close permanently. IPOX Schuster's Lukas Muehlbauer put it directly: higher rates are making investors more selective around growth valuations, but the window remains open. History supports the cyclical read. IPO markets are episodic by nature, and windows that close on rate shocks tend to reopen when yields stabilise.
But the structural case is at least as compelling, and it is the one the market is still discounting. Three pieces of evidence point to a regime shift rather than a pause. First, the 2026 recovery was an illusion of concentration: with SpaceX responsible for roughly 70% of proceeds and five of the ten largest listings already underwater, the broad market never really healed. Second, the private-market valuation edifice was constructed on near-zero rates; at a 10-year yield above 5%, a large share of that paper wealth cannot be realised anywhere near its last marked value. Third, the exit mechanism itself has atrophied. Private-equity-backed exits decelerated in the second quarter, and while most private-equity assets leave via sale rather than IPO, the public window is the price-setting margin that anchors every private negotiation.
The distinction matters because it changes who benefits. If the freeze is cyclical, it is a timing problem: issuers wait, banks build a backlog, and a wave of listings arrives once yields stabilise, a bullish signal for the banks and exchanges that sit on that backlog. If it is structural, the backlog never fully clears; valuations reset lower, and the winners are the buyers of last resort, strategic acquirers such as Stonepeak and public companies that can use stock as currency to buy private assets at a discount to their private marks.
The strongest counter-thesis is that the IPO drought is simply a calendar artefact, that companies delayed around the Fed meeting and will return in force once the dot plot is digested, and that the AI capital-spending cycle will keep demand for listings insatiable. That view has data on its side: with the S&P 500 and Nasdaq at records and AI revenue run rates still accelerating, the fundamental appetite for growth exposure has not vanished. The rebuttal is that appetite is not the constraint; price is. Demand at the old price is not demand. Oura's four-times-oversubscribed book proved that, and the company still walked away. When the best-quality issuer of the month declines to sell at any price, the market is not short of buyers. It is short of a price at which both sides agree.
What to Watch Next
Three signals will separate the cyclical read from the structural one. First, the 10-year Treasury yield: if it sustains below roughly 4.5%, the discount-rate pressure that drove the September withdrawals eases materially and the cyclical-pause thesis gains ground; a move back toward 5.3% would confirm the structural repricing. Second, Anthropic: if it prices successfully in November at or near the $2 trillion valuation under discussion, the window is demonstrably open for prime assets; a further delay or a pricing more than 20% below expectations would signal that even the strongest issuers cannot clear the new bar. Third, the backlog itself: more than two dozen companies with top-tier banks have filed publicly since July, but only nine have listed. A cluster of clean debuts in October would indicate pent-up supply finding a clearing price; a second wave of withdrawals would indicate the opposite.
The near-term outlook is for continued selectivity: strong issuers with pricing flexibility will wait for calmer windows, while weaker credits face a narrowing set of options. Over a medium horizon, the direction of long-term yields will determine whether the backlog converts into a wave or dissolves into write-downs. Structurally, the era of pricing IPOs as if the 10-year yield would remain anchored near zero is over, and with it the era in which a private-market markup could be exported to public investors without a fight.
The fall IPO window has not closed. It has simply stopped paying the prices that made it worth opening, and until yields and valuations renegotiate, the most rational thing a company can do is wait.
Explore more exclusive insights at nextfin.ai.
