Pre-IPO secondary market + public comps (May 2026)

Source report: /tmp/pre-ipo-secondary-market-2026-05-12.md (316 lines).

TL;DR — $1B exit math (ARR required at each multiple)

DoorMultipleARR needed for $1B exit
IPO at incumbent multiple$167M ARR
IPO at AI premium10-15×$67-100M ARR
Strategic acquisition (Moveworks/Armis band)22-25×$40-45M ARR

AI premium forecast to amortize to par within 12-18 months → 2026-Q4 to 2027-H1 window before the premium compresses.

Public comp scoreboard — May 2025 → May 2026 (ALL FOUR DOWN SHARPLY)

CompPrice moveMultiple compression
ServiceNow (NOW)-62% from Jan 2025 peak; ~455 pre-split equiv after 5-for-1 split Dec 17, 2025)EV/Rev 13-15× → ~6× (~55% compression)
Atlassian (TEAM)-63% YoY (86.60)EV/Rev 9.7× → 2.8-3.0× (~70% compression — deepest re-rating)
Salesforce (CRM)-37% YoY (181.30)Market not paying for Agentforce story despite $800M ARR + 200 IT Service customers
Freshworks (FRSH)-45% from peak (7.50)EV/Rev 4.35×; 11% layoff + $400M buyback Q1 2026 — live “AI SaaS in transition isn’t getting paid for it” comp

ServiceNow $11.7B M&A spree decomposed (CRITICAL)

AcquisitionPriceARRMultiple
Moveworks$2.85B$100M+~22-25× ARR
Armis$7.75B$340M~22.8× ARR
Veza~$1-1.5BTBDTBD
Cuein / Logik.ai / data.world / Traceloop$0.5-1.5B (combined tuck-ins)TBDTBD

75% of dollars went to cybersecurity (Armis + Veza), NOT AI agents. ServiceNow is positioning as the AI control plane / AI governance layer that AI agent vendors integrate with, rather than as the AI agent vendor itself. See agent-tool-governance.

Aisera $1.5B “deal price” — confirmed as Series E mark, NOT deal price

  • **Pre-deal PitchBook mark: 1.55B Series E peak).
  • “$1.5B” deal-price figure circulating on LinkedIn is almost certainly the Series E mark, not the deal price (officially undisclosed).
  • Confirms wiki’s earlier Aisera entity correction (May 11 2026 errata).

Pre-IPO secondary market data

Serval

  • Only AI ITSM pure-play with an NPM SecondMarket listing (per wiki).
  • Pricing gated behind auth — no observable secondary print yet.
  • NPM listing references “47M Series A at announce; could be insider extension OR NPM data labeling inconsistency. Flagged ^[ambiguous].

Other AI ITSM

Platform-level secondary data

Index / Platform2025 performance
Forge equal-weight FPMI+85.9%
Forge AI thematic basket+165.9%
Hiive50+49.1%
Caplight 2025 volume$3.5B (+75%)
Caplight AI/ML volume+1,104% from 2023

Late-stage private AI was bull; public AI ITSM was bear — that spread IS the strategic-acquirer arbitrage.

Services-as-software multiple — no confirmed discount, but lurking risk

  • Strategic premium (22-25× ARR) exists at M&A layer.
  • Top-quartile public AI premium (10-15× NTM) exists for the best names.
  • AI-first gross margins 20-60% vs 70-90% SaaS is the lurking discount driver. Public markets will discount once Q-over-Q GM data is visible.

Notes

  • Strategic-acquirer window: 2026-Q4 to 2027-H1 (AI premium amortizing to par in 12-18 months); 1B at the 22-25× Moveworks/Armis band.
  • Atlassian’s ~70% multiple compression occurred despite Service Collection reaching $1B ARR — AI revenue alone did not hold the multiple.
  • Freshworks ran a $400M buyback + 11% layoff in Q1 2026.
  • Serval NPM listing (47M) — unreconciled; possible insider extension. Flagged ^[ambiguous].
  • Late-stage private AI valuations (Forge AI basket +165.9%) ran above public AI-ITSM comps — the strategic-acquirer-arbitrage spread.

Honest verification notes

  • Aisera $257.3M revenue is Latka-modeled, flagged ^[modeled].
  • Moveworks ARR multiple framed as 22-25× band (per SaaStr/Everest), NOT the optimistic 28× figure used earlier (see vc-sentiment-exit-comparables-2026).
  • ServiceNow May 2025 baseline derived from peak ($1199 Jan 2025) + YTD drop framing rather than a fabricated mid-month price.
  • Freshworks layoff number flagged ^[ambiguous] between Benzinga’s 500/11% and CIO’s 660/13%.