The SaaSpocalypse is the name Wall Street gave to the software selloff that started in late January 2026, when investors began pricing in the risk that AI agents do the work that per-seat SaaS charges for. The SaaS Capital Index of public SaaS multiples fell from 5.6x run-rate revenue at the end of 2025 to 3.1x on June 30, 2026, its lowest month-end reading since 2011, and was back at 4.2x on September 30. For a private founder raising or selling in 2027, the headline number matters less than the spread: the market now pays several times more for fast growth and strong net revenue retention than for slow, seat-bound revenue. Growth and NRR are still the drivers of the multiple; what changed is how hard the market punishes their absence.
Data as of October 22, 2026. Every multiple below was read on the primary source on that date, and this page is refreshed each quarter when SaaS Capital, Clouded Judgement and Aventis publish new figures. This is financial-planning content, not investment advice.
If you want your own number while you read, the free SaaS valuation calculator runs the same method used in the worked example below, with every multiple sourced and dated.
Where the word "SaaSpocalypse" came from
The term is credited to Jeffrey Favuzza, who works on the equity trading desk at Jefferies, quoted by Bloomberg on February 3, 2026 in a story about traders dumping software stocks. He called it the "SaaSpocalypse". The selloff that week followed Anthropic's release of plugins for its Claude Cowork agent, including one for in-house legal work (Anthropic's own legal-industry post refers to that first legal plugin), and legal-data and publishing stocks fell hardest on the day.
The argument behind the word is the one covered in our guide to moving from per-seat to hybrid pricing: a seat is a proxy for a human doing work in the product. If an agent does that work, the customer needs fewer seats, and a vendor paid per seat earns less exactly when its product does more. Investors did not wait for that to show up in revenue. They cut the multiple first.
What happened to public SaaS multiples in 2026
The cleanest series is the SaaS Capital Index: the median market cap over annualized run-rate revenue for about 63 listed B2B SaaS companies. The month-end values below come from SaaS Capital's downloadable data file (data as of September 30, 2026):
| Month end | SaaS Capital Index (median, x run-rate revenue) |
|---|---|
| August 2021 (series peak) | 16.9x |
| December 2025 | 5.6x |
| January 2026 | 4.8x |
| February 2026 | 3.6x |
| June 2026 (2026 low) | 3.1x |
| August 2026 | 4.6x |
| September 2026 | 4.2x |
Two things stand out. The fall from December to February (5.6x to 3.6x) was 35% in two months; the 2022 rate-hike reset took the index from 13.7x to 6.5x, a 53% fall, over a full year. And the June reading of 3.1x was the lowest month-end in the series since September 2011. Other trackers agree on the shape. Aventis Advisors (August 31, 2026) says public SaaS multiples bottomed at 3.2x EV/revenue in June, which it calls the low point of the AI disruption selloff, and stood at 4.6x in August. The BVP Nasdaq Emerging Cloud Index lost 31.8% from December 31, 2025 to its April 13, 2026 low and was 26.8% above its year-end level on October 22, 2026 (our calculation from the index levels Bessemer publishes).
So the "apocalypse" part is half true. Prices fell hard and fast, then recovered most of the way. What did not recover is the old idea that all SaaS revenue is worth roughly the same multiple.
The real story: the growth premium got much wider
Jamin Ball's weekly Clouded Judgement splits public software by expected next-twelve-months (NTM) revenue growth: high growth is over 22%, mid growth 15% to 22%, low growth under 15%. Multiples are enterprise value over NTM revenue.
| Clouded Judgement post | Overall median | High growth (>22%) | Mid growth (15-22%) | Low growth (<15%) | High vs low |
|---|---|---|---|---|---|
| February 6, 2026 | 3.6x | 9.7x | 6.4x | 2.6x | 3.7 times |
| June 19, 2026 | 2.9x | 16.8x | 4.7x | 2.5x | 6.7 times |
| October 16, 2026 | 4.5x | 18.7x | 7.3x | 3.7x | 5.1 times |
Read the columns, not the overall median. Between February and October the high-growth median almost doubled while the low-growth median moved from 2.6x to 3.7x. The median company in Ball's set is expected to grow 13% over the next twelve months, with median net retention of 110% (October 16 post), which puts the typical listed SaaS company in the low-growth bucket.
Two other primary sources show the same split:
- ICONIQ, 2026 State of Scaling (September 2026): public software companies growing 20% or more rebounded to 9.2x NTM revenue in Q2 2026 from about 7x in Q1, against 2.7x for those growing under 20%. ICONIQ also finds the Rule of X (growth weighted about 2.7 times, plus free-cash-flow margin) explains more of the variation in multiples (R-squared 37%) than revenue growth alone (30%) or the plain Rule of 40 (29%), and that FCF margin is regaining weight as a driver.
- Bessemer's cloud index constituents (65 companies, October 22, 2026, our calculation from the per-company data Bessemer publishes): median EV/annualized revenue 5.5x for the whole index, 4.0x for the 23 companies growing under 15%, 5.8x for the 24 growing 15% to 25%, and 13.0x for the 18 growing over 25%. The index's simple average is 8.6x, which a few very high multiples pull well above the median, so do not quote it as "the" SaaS multiple. Growth, free cash flow margin and Rule of 40 for each of these companies, computed from their own SEC filings, are in our public SaaS companies table.
That is the part of the SaaSpocalypse that will still be true in 2027. The market did not decide software is worthless. It decided that slow-growing software with exposure to seat compression deserves a multiple close to a mature industrial business, and that fast growers, many of them AI products, deserve a large premium.
What it did to private SaaS multiples
Private companies are not marked to market every day, so the effect arrives through deals and term sheets with a lag. Aventis Advisors tracks private SaaS M&A with a disclosed revenue multiple (543 deals from 2015 to 2026, Mergermarket data):
| Year | Median private SaaS EV/revenue (Aventis) |
|---|---|
| 2021 | 6.3x |
| 2022 | 5.4x |
| 2023 | 3.7x |
| 2024 | 2.9x |
| 2025 | 3.8x |
| Q1 2026 (small sample) | 3.1x |
| 2015-2026, all deals | 4.5x median (25th percentile 2.4x, 75th percentile 8.1x) |
The private median for Q1 2026 fell back to 3.1x, but it is based on a small number of closed deals and Aventis labels it as such. Deals that closed in Q1 were mostly priced in late 2025, so the full effect of the February selloff will show in the 2026 full-year figure. Aventis also notes that companies guide to 11-12% growth this year with consensus below 10% for 2027, and that premiums go to larger ARR, strong NRR and efficient unit economics.
Private multiples follow public ones through two channels. Acquirers that are themselves listed pay with stock whose multiple just fell. And VCs price rounds against an exit they expect to sell into the public market. SaaS Capital's private valuation framework makes the link explicit: its formula starts from the public index and adds growth and NRR terms, which is the method our calculator uses for private companies.
AI-native vs legacy SaaS: what actually changed
The narrative says AI-native companies are taking the value that seat-based SaaS loses. The data supports a narrower version of that.
- Retention is where AI-forward companies separate. In ICONIQ's 2026 State of Scaling, "Pacesetters" (top-quartile growth, AI-native or AI-driven companies; about 60% are AI-native) under $100M ARR have median net dollar retention of 120% in 2026, against 102% for the other companies in ICONIQ's data at the same scale. The Pacesetter group at that size is small (five companies), so read it as a direction, not a benchmark.
- AI products still run lower gross margins. ICONIQ's State of AI 2026: The Builder's Economy (July 2026, survey of about 300 executives, average N=287) puts gross margin on AI products at 45% in 2025, projected 53% in 2026 and 59% in 2027. That is far below the classic 75-80% software margin, which is why buyers now ask about inference cost per customer. Our guide to AI COGS and gross margin shows how to model it.
- Seats did not disappear. In the same ICONIQ survey, AI pricing models (select all that apply) moved between Q4 2025 and Q2 2026 as follows: subscription or platform 58% to 57%, consumption 35% to 42%, seat-based 23% to 28%, outcome-based 18% to 23%. Companies used 1.7 models on average, up from 1.5. The move is toward hybrid pricing, a base fee plus a meter, not toward abandoning seats. Outcome-based pricing is growing from a low base.
For a private founder, "AI-native" is not a label that earns a multiple by itself. What earns it is the metrics AI-native leaders happen to post: fast growth, high NRR and a pricing model whose revenue rises with usage.
Worked example: two $3M ARR companies, one market
Take two private SaaS companies, each at $3M ARR with a -5% EBITDA margin, valued on the private-company method in our SaaS valuation calculator. That method is SaaS Capital's published formula (from its 2022 framework): multiple = -3.2 + 0.32 x SaaS Capital Index + 8.26 x ARR growth + 2.62 x NRR, with growth and NRR as decimals, run with the index at 4.2x (September 30, 2026). The low and high ends scale the middle multiple by Aventis's private-deal spread (2.4x and 8.1x around a 4.5x median).
| Company A | Company B | |
|---|---|---|
| ARR | $3,000,000 | $3,000,000 |
| ARR growth | 30% | 15% |
| Net revenue retention | 105% | 95% |
| Multiple (middle) | 3.37x | 1.87x |
| Valuation, middle | $10.1M | $5.6M |
| Valuation, low to high | $5.4M to $18.2M | $3.0M to $10.1M |
The arithmetic: A = -3.2 + 0.32 x 4.2 + 8.26 x 0.30 + 2.62 x 1.05 = -3.2 + 1.344 + 2.478 + 2.751 = 3.373x, so $3M x 3.373 = $10,119,000. B = -3.2 + 1.344 + 8.26 x 0.15 + 2.62 x 0.95 = -3.2 + 1.344 + 1.239 + 2.489 = 1.872x, so $5,616,000. Same revenue, a $4.5M gap in value.
Where the gap comes from, in this formula:
- 15 points of growth = 1.24x on the multiple (8.26 x 0.15), about $3.7M on $3M ARR. If Company B grew 30% with its 95% NRR, it would be valued at $9.3M.
- 10 points of NRR = 0.26x directly (2.62 x 0.10), about $0.8M. If B kept 15% growth but lifted NRR to 105%, it would be valued at $6.4M. The direct NRR term looks small, but NRR is also part of growth: a company at 95% NRR has to replace 5% of its ARR with new customers before it grows at all, so in practice fixing retention is often the cheapest way to lift the growth term too. See net revenue retention for the mechanics.
- The market moved both companies by the same amount. The index term is 0.32 x index. The fall from 5.6x (December 2025) to 3.1x (June 2026) took 0.78x off every private multiple in this formula, about $2.4M at $3M ARR. For Company B that is a third of its December value; for Company A, about a fifth. Slow growers lose a larger share of their value in a selloff because the market term is a bigger part of their multiple.
On public comparables the gap is wider still. By Clouded Judgement's buckets (October 16, 2026), A falls in the over-22% bucket at 18.7x and B in the 15-22% bucket at 7.3x; by Aventis's Rule of 40 slope (+1.02x per 10 points around a 4.6x median at a score of 26), A's score of 25 gives 4.50x and B's score of 10 gives 2.97x. Public multiples are an upper bound for a $3M ARR private company, not a target, but they show the direction: a buyer in 2027 will pay most for the metrics that put a company in the top bucket.
What to fix before a raise or sale in 2027
- Get NRR above 100% and show the bridge. Investors now separate companies that grow inside their base from companies that refill a leaking bucket. Present starting ARR, expansion, contraction and churn as a monthly bridge, not just the ratio. The NRR guide covers the definition buyers expect.
- Show your seat exposure honestly. If revenue is per seat and your customers' teams are shrinking, a diligence team will find it. Model what happens if seat counts fall 20-40%, and show the pricing response (base plus usage, credits or outcomes). Our per-seat to hybrid pricing guide has a worked migration model, and our Entrata S-1 teardown shows a 2026 IPO candidate that bills per apartment unit instead of per seat.
- Know your gross margin per product line. If you sell AI features, separate inference cost from hosting and show cost per customer at the median and at heavy users. A blended margin that hides a 40% AI line will be found in diligence. See AI COGS and gross margin.
- Report growth plus profitability together. ICONIQ's finding that FCF margin is regaining weight matches what buyers ask for: a Rule of 40 or Rule of X score with the inputs shown, not just a growth rate.
- Anchor your valuation expectation to a dated number. A founder quoting 2021 multiples, or even December 2025 ones, starts the negotiation from a number the market no longer pays. Run your figures through the SaaS valuation calculator and check the date on each source.
The process side (timing, round size, SAFE versus priced round, data room) is in the SaaS fundraising guide, and the full method behind the multiple is in how to value a SaaS company.
Each lever above is a line in your operating model, so the useful question is how much each one moves next year's ARR. Adlega builds a driver-based 36-month SaaS model with NRR, growth, gross margin and Rule of 40 tracked as you change assumptions, and its AI CFO shows the formula behind each number when you ask why the forecast moved.
FAQ
What is the SaaSpocalypse?
It is the nickname for the selloff in software stocks that began in late January 2026, driven by the fear that AI agents will reduce the number of paid seats customers need. The SaaS Capital Index of public SaaS multiples fell from 5.6x at the end of 2025 to 3.1x on June 30, 2026 and recovered to 4.2x by September 30, 2026.
Who coined the term SaaSpocalypse?
It is credited to Jeffrey Favuzza of the equity trading desk at Jefferies, quoted by Bloomberg on February 3, 2026, the week Anthropic released plugins for its Claude Cowork agent and software stocks fell sharply.
Is SaaS dead?
No. Public software multiples recovered most of their 2026 fall, and high-growth software traded at 18.7x NTM revenue on October 16, 2026 (Clouded Judgement). What is under pressure is slow-growing, seat-priced software: the low-growth bucket traded at 3.7x on the same date. The data points to a repricing by growth and retention, not to the end of subscription software.
What are SaaS valuation multiples in 2026?
The SaaS Capital Index was 4.2x run-rate revenue on September 30, 2026. Clouded Judgement's overall public software median was 4.5x NTM revenue on October 16, 2026, ranging from 3.7x for companies growing under 15% to 18.7x for those growing over 22%. Aventis puts the median private SaaS deal at 3.8x revenue in 2025 and 3.1x on a small Q1 2026 sample.
Do private SaaS companies get public multiples?
Usually not. Private companies are smaller, less liquid and riskier, so they generally trade at a discount to listed peers. SaaS Capital's private-company formula starts from the public index and adjusts for growth and NRR; at today's index, a $3M ARR company growing 30% with 105% NRR comes out at about 3.4x, well under the 18.7x public high-growth median.