Takeaways from the Jefferies monthly software valuation update
Rob Bartlett of Jefferies joined the panel I moderated last month on whether companies should buy, sell, or wait. This week I read his team鈥檚 monthly software valuation update, with data through September 30. I read a lot of bank decks. This one I read twice.
Here is what is new in it, and what the room added that the charts cannot.
Eighty percent of software is slow now
More than 80 percent of public software companies are expected to grow less than 20 percent next year. Five years ago, nearly half the industry grew faster than that. Today, of the 123 companies Jefferies tracks, exactly eight are projected to grow above 30 percent, and they are chips, clouds, and data platforms rather than the subscription names of the last cycle.
The growth premium collapsed, and a cliff appeared
In November 2021, the market valued a point of revenue growth at nearly five times a point of cash flow margin. Today the ratio is 1.7. That 64 percent collapse is the quiet fact underneath every repriced term sheet of the past four years.
Growth still wins. Companies growing 20 to 30 percent trade at 11.5 times forward revenue and returned 41 percent over the past year, best in the report, while companies growing under 10 percent trade at 3.5 times and lost a quarter of their value. What the averages hide, and what our panel put its finger on, is that the line between those outcomes is a cliff, not a slope. Cross 20 percent growth and the multiple roughly doubles. Nothing else a management team can do moves the number like getting over that line.
About that SaaSpocalypse
Which brings me to the word of the year. The SaaSpocalypse is the theory that AI agents will eat application software whole, and early this year it erased something like $300 billion of application software market value. The fear is priced. Application software lost 32 percent over the past year and trades at 3.8 times forward revenue, about half its pre-pandemic multiple, while security software returned 64 percent and trades at a record 10.4 times. The market pays half price for anything it thinks a model might replicate, and a record premium for whatever defends the result.
Priced is not the same as right, and Rob made the sharpest version of the counterargument on our panel. The AI dollars are not coming out of software budgets. They are competing with services and headcount, the categories where enterprises have always spent several times what they spend on software. That squeezes sales cycles, because AI is consuming the discretionary budget. It also points at a bigger prize, software plus AI absorbing work that was always billed as human labor. A category taking share from services spending is not dying.
The report backs him with two facts the obituaries skip. The share of software companies earning 20 to 30 percent free cash flow margins has doubled since 2021, while the share burning money shrank from nearly half the industry to about one in ten. It is a strange apocalypse in which the victims keep minting cash. And the Rule of 40 now explains valuations nearly twice as well as it did in late 2021. Panic does not run regressions. This market is reading income statements, and to a strategic buyer, an obituary written at 3.8 times revenue reads like a price list.
Slow software is cheap, fast software is expensive, and the market can finally tell the difference. That is not an apocalypse. That is a market.
What the room added
The charts stop where diligence begins, and the panel鈥檚 advice was mostly about what the charts cannot see.
Product is the new first question. A growth investor on our panel called product the first, second, and third diligence consideration, because whatever a buyer acquires today will look different by the end of a five-year hold. The classic SaaS metrics still matter, but they no longer carry the underwriting on their own. Win rates, roadmap, engineering bench, and whether AI strengthens the product or commoditizes it now decide where a company lands within its bucket.
Sellers should start with the buyer鈥檚 story, not their own. The advice from the strategic side of the panel was to understand what the buyer underwrites, what gap you fill, and where you sit on their roadmap, because strategic value comes from fit and synergies rather than sector multiples. The highest-value M&A narrative is not why we are great. It is why we are worth more to you than to anyone else.
And when bid and ask will not meet, the bridge gets built in the documents. Rollover equity, secondary at a discount, earnout kickers, each one a drafting exercise, and each one able to leave management misaligned at exit if overdone. One panelist put the whole negotiation in a single line. You tell me the price, I will tell you the terms. The lawyer in me adds only this. Realistic economics up front beat clever structure, because the preference stack decides who absorbs the difference either way.
The bottom line
Keep one question from all of it, the one the panel kept circling. Does your product become more valuable or less valuable as AI models get dramatically better? Answer that honestly, and everything else in this report is arithmetic you can plan around.
My thanks to Rob Bartlett and the Jefferies technology banking team. The figures above come from their Monthly Software Market Valuation and Performance Update, October 2026.
