Analysis
The US market has completed 238 IPOs so far in 2026 against 248 at the same point in 2025, a decline of about 4%, per StockAnalysis's running tally. That number sits awkwardly next to the year's headline: SpaceX listed in June at a $1.77 trillion valuation, the largest offering in history, and Anthropic is expected to follow in October. Both things are true, and the gap between them is the actual story of this IPO year.
The explanation is mix, not volume. Deal count is dominated by micro-caps, SPACs and foreign private issuers doing $20 million to $75 million offerings, and that cohort has thinned. The dollar total is dominated by a handful of enormous deals. A year can therefore set a record for capital raised while listing fewer companies, and 2026 is doing exactly that. It also means the count is a poor proxy for whether the window is open for a venture-backed software company, because almost none of the 238 are venture-backed software companies.
September illustrates the concentration. Renaissance Capital's pricing table shows a single completed pricing so far this month -- TurboGen on September 3, which closed its first day down roughly 60%. The calendar then jumps straight to four deals in the week of September 15 totaling about $1.8 billion: Holtec Nuclear, Bamboo Insurance, Electra Therapeutics and Orion180. Long quiet stretches punctuated by clustered supply is what an issuance market looks like when underwriters are waiting for specific windows rather than running a continuous book.
“Deal count is dominated by micro-caps, SPACs and foreign private issuers doing $20 million to $75 million offerings, and that cohort has thinned.”
The rate environment explains the waiting. With the 10-year Treasury at 5% and an 88% market-implied probability of a Fed hike this week, the discount applied to a pre-profit issuer's terminal value is the harshest it has been since 2023. Issuers with near-term cash flows -- insurers, energy infrastructure, specialty finance -- can price into that. Issuers asking buyers to underwrite 2030 are choosing to stay private, which is why the four deals this week include two insurance companies and zero software companies.
TurboGen's 60% first-day decline is the cautionary data point the pipeline is reading. A broken deal in a thin month does disproportionate damage to the next issuer's book-building, because the sell-side has to explain it in every call. That single print is likely worth more delay in the pipeline than any macro variable this month.
For venture funds, the practical consequence is duration. If the public market is absorbing asset-heavy issuers and deferring growth ones, the exit path for a 2021-vintage software portfolio runs through M&A and secondaries rather than listings, and both clear at discounts to the marks on the books. The DPI conversation with LPs does not improve because SpaceX and Anthropic printed large numbers; those are two companies, and they are not in most portfolios.
