Median horizontal B2B software now trades at 2.7x trailing-twelve-month revenue — less than a third of what infrastructure commands. That figure, drawn from JPMAM data in a16z Growth's September State of Markets deck, frames the sharpest valuation split in public software: cloud, data, and AI infrastructure sits at 9.1x, security and identity at 6.8x, vertical software at 4.6x, consumer and transactional platforms at 4.0x, and horizontal apps at the bottom of the stack. Every category compressed. Horizontal apps compressed the most.

The market is pricing defensibility, and the deck treats this explicitly. If you sell a horizontal app an AI-native competitor could plausibly rebuild, the market assumes someone will. If you own the data layer, the security layer, or a vertical workflow with real switching costs, you get paid more than 3x as much per dollar of revenue. For founders, the implication is direct: a horizontal company's next fundraise or M&A conversation starts from 2.7x, so either build a credible infrastructure or vertical story or plan around that multiple.

Growth Determines Multiple

The same slide splits public software by growth rate on TEV/forward revenue. Companies growing 20–40% trade in roughly the 9–13x range. Companies growing 10–20% slid to 4–5x and stayed there. a16z's own summary: high-growth multiples have normalized; medium-growth multiples haven't recovered. The gap between 25% growth and 15% growth is worth 2x to 3x on valuation.

About 75% of public software is now profitable, and only about 30% of it grows 20% or more. a16z describes the shift as software having "traded growth for profitability," and notes the market then priced those companies exactly like profitable companies growing in the low teens. Profitability earns no premium when three-quarters of the comp set already has it.

Growth-by-cohort data shows the decline has stopped. The 90th percentile sits around 29–30%, the 75th percentile at 20–22%, the median at 12–13%, and the 25th percentile at high single digits. Operating leverage metrics — OpEx/revenue, margin change, revenue per employee — are stable or improving. a16z calls the fundamentals "remarkably stable" through the SaaSpocalypse selloff: stock prices moved more than the businesses did.

The Stripe Numbers Need Careful Reading

The deck's most interesting slide is also its most misleadingly titled. a16z claims B2B accelerated into the SaaSpocalypse using Stripe payment data, but the chart splits companies by age. B2B companies under one year old ran year-over-year growth in the low hundreds of percent through 2025, then went nearly vertical to about 500–600% by early 2026, cooling slightly but staying above 500%. Mature B2B companies peaked around 32% in early 2025, decelerated to about 19% by January 2026, and have recovered to only about 24%.

The acceleration is almost entirely brand-new companies. Established B2B — the companies most founders run — went through a real slowdown, not just a sentiment shift. The competitive threat for an incumbent is a company that didn't exist 12 months ago and is growing 5x. The deck's recommendation: assign someone to track new entrants in your category by name every quarter.

Unicorns Split Into Two Populations

a16z titles its unicorn slide "Mostly Profitable, Mostly Growing ≤ 20%." The SVB data it cites supports the second half but not the first. For U.S. VC-backed tech unicorns in 2026: 42% grow 0–20% and another 15% are shrinking, so 57% sit at 20% or below. Only about 25% are actually profitable — 4% with margins above 25%, 21% at 0–25% — not "mostly." And 55% have under two years of runway: 26% at 0–1 years, 29% at 1–2 years.

The herd is aging. Median unicorn age is up 7% to roughly 15 years, while new unicorns are 37% younger than a few years ago. Two populations have formed: a large, aging cohort of low-growth, mostly unprofitable companies burning toward a deadline, and a small cohort of young AI companies reaching $1B+ faster than ever.

The Raise Bar Hasn't Moved

Startups that raised recently grow roughly 60–70% with deeply negative margins — in line with peak ZIRP. Median scaled startups across fintech, enterprise, and consumer internet grow 15–30% with modestly negative margins. As of Q2 2026, a company growing 25% at breakeven has a good business that probably can't raise a priced equity round at a good valuation.

Carta data covering 2,773 funds and about $119B in committed capital as of Q1 2026 shows the 2024 vintage at 40.5% net IRR for the 90th percentile and -3.3% at the median — a 44-point spread. Median DPI is 0.00x for every vintage from 2021 on. LPs have received essentially no cash back, and fund position on that curve affects founders: a fund with 0.00x DPI will be slower on follow-ons.

On AI adoption, only 20% of organizations cite cost as a constraint, per McKinsey data, and AI spend remains below cloud's 4% of IT budgets in year two. Adoption is wide but shallow — a small sliver of power users consumes orders of magnitude more AI than everyone else. And Revelio Labs/Ramp data contradicts the entry-level elimination story: 24 months after adoption, high-intensity AI adopters increased entry-level hiring share by 1.15 points while low adopters cut it by 0.52 points. Jason Lemkin's summary on X was blunt: "All That Matters is Growth. Do gross margins matter? No. Does it matter if revenue recurs? No."

Whether the 500%+ growth of AI-native entrants pulls mature B2B growth back above 24% — or compresses horizontal multiples further — is the question the next State of Markets installment will have to answer.