Funding & M&A Digest — September 11, 2026: Fresha’s $1B Non-AI Unicorn, a16z Doubles Down on Procurement, and the NRR Chasm Widens

Two things happened today that should be read together, not separately. OpenAI shipped GPT-6 Astra and floated AGI-onset language. And KKR wrote an $80M check into Fresha — a booking marketplace for salons and spas — at a $1 billion valuation. No frontier model. No agentic workflow. Just scheduling, payments, and a category lock in a fragmented vertical.

That juxtaposition is the whole market in one frame. Capital is bifurcating: one pool chasing AI-native categories before the model layer eats them, another pool retreating to businesses whose moat is workflow gravity and payment rails — things a foundation model cannot casually replicate.

What Printed Today

Fresha — $80M from KKR NGT Growth, $1B valuation

The headline isn’t the amount. It’s the absence of AI in the thesis. Fresha operates a vertical booking marketplace for beauty and wellness — a category with millions of micro-merchants, terrible incumbent software, and embedded payment monetization. KKR’s Next Generation Technology growth arm is not buying multiple expansion on an AI narrative; it’s buying take-rate compounding on transaction volume.

For founders, the read-through is unglamorous and useful: vertical SaaS with payments attached still clears $1B in 2026. When the median private B2B SaaS company is growing 26% (down from 30% in 2022), a business with merchant-level switching costs and non-discretionary usage is a defensible asset — not a consolation prize.

Lio — $30M Series A, led by a16z

AI enterprise procurement. And critically, this is the second procurement-automation round this quarter. When a16z leads a Series A in a category that already saw a funded competitor within 90 days, that’s not discovery — that’s a land grab. Procurement is being priced as a winner-take-most category, and the consolidation window is roughly 18 months before one of these companies becomes an acquisition target for Coupa, SAP Ariba, or a private-equity roll-up.

Watch the burn discipline here. Procurement software sells to CFO organizations, which means long cycles, security reviews, and CAC that compounds badly if you miss. With CAC up 14% year-over-year across the sector and CAC payback at $1–10M ARR sitting at 15 months, a $30M Series A into a contested category is a bet that speed beats efficiency. Sometimes it does. Often it doesn’t.

Catena Labs — $30M Series A · GEEIQ — $6.8M

Catena Labs continues the AI-native financial infrastructure thesis. GEEIQ — brand analytics for gaming and virtual worlds — is the kind of $6.8M round that tells you the seed-to-Series-A gap has widened: it’s a bridge to proof, not a scale round.

The unicorn firehose is still running

Per the Crunchbase Unicorn Board, 29 companies joined in August, adding roughly $63 billion in board value. That averages to ~$2.2B per addition. Paper value, yes — but it signals that late-stage private markets have not repriced the way public comparables have. That gap is where 2027’s down rounds live.

The Metric That Explains All of It: NRR at 101%

The 2026 Benchmarks Report puts annual net revenue retention for private B2B SaaS at 101%. Read that number slowly. The median company now expands almost exactly as much as it contracts. Expansion revenue — the mechanic that powered 2019–2021 valuations — is no longer a growth engine for the median. It’s a leak-plugging exercise.

Now the contrast. Snowflake printed 126% NRR on $6B in revenue with 37% growth that is accelerating (FQ2’27). A company 100x the size of the median is growing faster and retaining better. That is not a performance spread. That’s a chasm.

The practical implication for anyone raising: investors are no longer underwriting “we’ll fix retention with a customer success hire.” NRR is now read as a product-market-fit signal, not an ops signal. If you’re under 105%, the diligence conversation is about whether your category has structural expansion headroom at all.

The benchmark stack you should actually be measured against

• LTV:CAC floor has moved from 3:1 to 4:1. Three-to-one now reads as “you are technically not losing money.” Four-to-one is the new healthy.
• Monthly churn averages 3.5% in B2B SaaS; top performers hold under 2%. Recurly’s cross-sector data puts total SaaS churn at 3.22% — split 2.16% voluntary, 1.06% involuntary. That involuntary third is failed payments and expired cards, and it’s the cheapest churn you will ever recover.
• CAC payback: 15 months at $1–10M ARR. Top performers land 5–7 months.

The India Angle: Capital Is Getting More Specific

Indian SaaS funding this cycle is notably less generalist. Graph AI — an AI-native patient safety platform — raised $13.3M led by Insight Partners, a cross-border check into a narrow clinical-workflow category. That’s the pattern: US growth funds are underwriting Indian companies that own a regulated, high-consequence workflow, not horizontal productivity tools.

The structural advantage remains intact. Indian SaaS companies build to global ARR with domestic engineering cost bases, which means CAC payback math that a Bay Area competitor structurally cannot match. When sector CAC is up 14% and payback sits at 15 months, an India-headquartered company running 8–10 month payback isn’t just efficient — it’s holding an arbitrage that widens as US go-to-market costs inflate. Expect more Insight, Accel, and Peak XV checks into Delhi, Bengaluru, and Chennai companies with US logos and Indian cost structures.

The Risk Nobody Priced This Morning

Two governance shocks worth logging. First, Anthropic was designated a Pentagon supply-chain risk and a $200M contract collapsed — triggered by control terms over autonomous weapons and domestic surveillance, not by model performance. Federal ARR carries governance clau