Markets · Opinion
Stop discounting your boring software; the card data says buyers kept paying
Stripe's index shows non-AI B2B revenue up 30%+ through the sell-off. Sellers who cut prices to match the stock chart are giving away money, and buyers waving that chart have less leverage than they think.
By early February 2026 nearly $1 trillion of enterprise software value had gone, and perhaps $2 trillion by the time the dust settled. Any procurement lead holding that chart walked into spring renewals expecting vendors to be frightened. Stripe's new weekly index says the vendors had no reason to be. Across an average of about 72,000 non-AI software businesses a week, revenue is growing above 30% year over year, at or above the highs of summer 2023, and the index sits about 4% above its pre-sell-off trend.
I expected to dismiss this. Jason Lemkin's write-up lists five reasons to, and most of them are fair. The sample is mostly self-serve, card-on-file companies, and the large public vendors that lost the value barely appear in it. A company that winds down to zero drops out of a same-store index rather than counting as a 100% decline. The index is nominal, so price rises count as growth. The median company only got back to trend, UK growth has slowed since summer 2025, and the rest of the world followed the same pattern. I accept all of it. None of it rescues the buyer's position, and some of it makes the seller's position stronger.
Take the nominal point first. Lemkin reads it as a warning: revenue can rise 20% while seats fall, if price per account rises faster. Look at what has to be true for that to show up in card payments. Customers had to pay the higher price. Nobody forced them to, and these are monthly card subscriptions that can be cancelled in a click. If price is doing much of the work, buyers accepted the increases during the very months the stock market was telling them their vendors were doomed. Anyone planning to negotiate on the strength of the sell-off should look at that first.
Now the best objection, which Lemkin himself makes: the companies that sold off sell to enterprises on annual contracts, paid by wire, and Stripe barely sees them. True. The sell-off priced cash flows five to ten years out, and Tedeschi says strong revenue does not prove the market was wrong. I make no claim that enterprise vendors are fine; the buyer simply has no evidence yet that they are weak. Leverage at a renewal comes from a credible alternative, and the alternative in this story, agents doing the work that seats used to do, shows up in a negotiation only when a buyer can say "we cut seats last quarter and nothing broke." Few buyers can say that today. Public net revenue retention will tell us who is right, quarter by quarter, and until it falls buyers have no number to put on the table at renewal.
The data also shows where the budget went. Companies less than a year old are growing fastest, and by Stripe's proxy, which uses signals like MCP requests and agentic CLI usage, they are now 4 points more likely than mature firms to be adopting agent infrastructure. Mature firms led by 5 points at the start of 2026, and that gap flipped in April. Software spend is moving to different vendors, and I read that as the real threat to incumbents. A buyer who tells a mature vendor that AI makes its product worthless is overstating it, because the money is still going to software, just newer software.
For sellers, the lesson is narrower than "everything is fine." Mature businesses stopped slowing, but only back to their mid-2025 pace, which is where the slowdown started. A stabilised business that cuts list price in response to a stock chart is paying for a problem its customers have not shown up with. The sensible move is to split 2026 growth into seat expansion and price per account, which Lemkin also recommends, and to decide which of the two you are actually selling. If your seats are flat and price is carrying you, the risk is real, and the fix to look at is a different pricing unit, with the discount left out.
Buyers should do the same sum from the other side. Pull your 2026 renewal quotes, compare the uplift with your seat count, and ask whether the vendor's leverage rose or fell. If you paid more for the same seats, the sell-off did nothing for you, and I would not expect a better deal in January 2027 than the one you took this year.
Prompted by Stripe’s New Index Shows Non-AI B2B Revenue Up 30%+ Year Over Year Through the SaaSpocalypse, SaaStrAI.