Outside AI, the last round is not a floor

For the founder of a technology company outside AI, this year's record venture totals describe a market that has mostly gone elsewhere. The PitchBook-NVCA Venture Monitor, published October 7, put US venture deal value through the third quarter at $515.8 billion, already the highest total in its dataset, and found that AI accounted for 82.7% of it.^[1] Rounds of $100 million or more were about 5% of third-quarter deals but 75% of the quarter's deal value.^[1] Our conclusion for other companies is to build a raise, a share sale or an acquisition on today's evidence, and to treat the last priced round as history rather than a floor.

Two markets inside one total

The concentration eased during the year. AI's share of quarterly deal value fell from 88.4% in the first quarter to 65.9% in the third as the largest rounds of the first half tapered off, while an estimated 5,012 deals closed in the third quarter, the second most of any quarter in the dataset.^[1] Capital is available, mostly in smaller amounts. Cooley's review of second-quarter financings found 83.6% up rounds, 4.3% flat rounds and 12.1% down rounds, with 95.8% of deals carrying a 1x liquidation preference.^[2] Pay-to-play provisions rose to 8.4% of deals from 7% in the first quarter.^[2] Those figures describe financings that closed. Companies that chose not to raise on the terms offered do not appear in them.

Listings have not reached most companies

Renaissance Capital counted 30 US IPOs in the third quarter, raising $32.8 billion, or $6.2 billion without its largest offering.^[3] Venture-backed companies accounted for nine of them and $2.4 billion; biotech made up eight of the nine, and in Renaissance's words, traditional tech unicorns were sidelined once again.^[3] The Venture Monitor estimates the backlog of companies that need to, or should have, gone public at more than two full years of IPOs, and says most companies without a path to a marquee IPO would realistically need to accept exit prices well below peak-era expectations.^[1] That is PitchBook's reading of the broad market, not a valuation of any company.

Secondary prices show where marks stand

Secondary trading gives the most current read. Forge reported on September 22 that its median trade in August cleared at a 3% discount to the last primary round.^[4] Among companies that last raised in 2021 or the first quarter of 2022, the 112 that have traded on its marketplace showed a median Forge Price 61% below that round.^[4] Company-run tender offers look calmer. Carta administered 71 in the first half, with about $3 billion of volume, and the median discount for tenders held at least a year after a primary round has been 0% in each of the last five half-year periods.^[5] The measures differ in who sets the price and which companies take part. A tender that a company runs for its own holders is not an open-market price.

Plan the transaction around the evidence

For a raise, build the valuation case on current revenue quality and recent comparable transactions, and model the terms as carefully as the price: liquidation preferences, pay-to-play provisions and redemption rights change what each class of holder receives. For a sale, expect buyers to look at secondary marks and recent acquisition prices, and check early whether existing preferences leave a sale below the last round workable for every class of holder. For employees, a modest company-run tender can relieve pressure without forcing a full repricing.

None of this assumes that the concentration in AI persists or reverses. It assumes that the price a founder can defend is the one current investors and buyers recognize.