ServiceNow
Subscription SaaS (multi-year contracts), historically per-seat but shifting toward usage/consumption pricing for AI (Now Assist); ~97% subscription revenue, high net-retention land-and-expand.
The thesis on this name
State of AI Compute
ServiceNow is the leading enterprise workflow-automation SaaS platform (the 'Now Platform') spanning IT, employee, customer and now agentic-AI workflows, sold to large enterprises and governments.
State of Enterprise AI SaaS
The workflow system-of-record with the strongest hybrid-monetization path: Now Assist tracking to ~$1.5B ACV (raised from $1B), $1M+ Now Assist customers +130% YoY, subscription rev +22% (ServiceNow Q1'26, Apr 2026). Unlike pure seat vendors, NOW is layering consumption/agentic SKUs ON TOP of a sticky platform, so AI is additive rather than cannibalizing. 5-for-1 split (Dec 2025) broadened the holder base; forward P/E ~23-25x is reasonable for ~20% durable growth + expanding margins. Armis/Veza M&A widens the TAM into security.
State of Enterprise AI SaaS
Market made a category error de-rating NOW 36% YTD on seat-compression fear while subscription revenue still compounds 22% and Now Assist >$1M ACV deals grew 130%.
State of Enterprise AI SaaS
Best seat+consumption hybrid; ~20% durable growth at ~24x forward. GARP ballast.
Earnings, margins, COGS & capex
Best-in-class large-cap SaaS: ~$13.3B FY25 revenue still growing 22% at ~$15B scale, 81%+ subscription gross margin, 32% non-GAAP operating margin, ~35% FCF margin, cRPO +22.5%. Q1 FY26 beat and raised across every metric. Yet the stock is down ~52% over 52 weeks on AI-agent seat-displacement fear, not on fundamentals.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~20¢ is cost of goods and ~51¢ operating expense, leaving ~30¢ of operating profit.
Revenue trend
Margins
stable/slightly eroding on AI compute + Armis integration
FY26 guide 31.5%, ~50bps down YoY on Armis dilution
flat, high quality
COGS structure
Primarily cloud hosting (multi-cloud incl. Azure), data-center/compute for AI inference, and customer support; AI (Now Assist) inference cost is the new gross-margin watch item.
Capex
Asset-light; capex a low single-digit % of revenue (own + leased data centers, plus public-cloud). No fab/heavy-infra intensity.
Latest earnings
Beat. Subscription revenues exceeded the high end of guidance by 150bps and non-GAAP operating margin of 29.5% ran ~3 points above guide; cRPO cc growth of 21.5% was more than 2 points above guidance; non-GAAP diluted EPS $0.90 vs a $0.86 Zacks consensus (+4.7%) on revenue of $3.99B vs $3.93B. Two quality caveats sit on top: the release attributes part of the topline beat to on-premise mix from strong US Federal demand that 'accelerated some on-premise subscription revenues from Q3 2026 into Q2 2026' ($149M of Q2 subscription revenue was recognized upfront on self-hosted delivery, vs $109M PY), and GAAP diluted EPS fell to $0.29 from $0.37. Price action: the stock closed -6.5% at $95.46 on print day (Jul 22, released after the close), fell a further 3.7% to $91.94 on Jul 23, then rallied to $111.23 by Jul 31 - +21% off the Jul 23 close.
Raised. FY26 subscription revenues $15,760-15,780M, +22.5% (+21% cc); non-GAAP subscription gross margin 81%; non-GAAP operating margin 31.5%; free cash flow margin 35%; weighted-average diluted shares ~1.04B. Q3 FY26: subscription revenues $3,975-3,980M, +20.5% (+20% cc); cRPO growth 19.5% (+20% cc) against an estimated ~$35M year-over-year FX headwind; non-GAAP operating margin 31%; diluted shares ~1.05B. Long-term targets reiterated from the May 4 Financial Analyst Day: $30B+ subscription revenues, 30% of ACV from AI, Rule of 60+ by 2030, and SBC below 10% of revenue by 2029; the company describes 2026 as operating to the Rule of 56.
- ServiceNow AI ACV
- Crossed $1 billion in Q2 FY26; management says it is on track to beat the $1.5B end-2026 target
- cRPO
- $13.20B, +21% YoY (+21.5% cc) - decelerating from 22.5% cc in Q1 FY26; Q3 guided to 19.5% (+20% cc)
- RPO
- $29.0B, +21% YoY (+22% cc)
- Non-GAAP diluted EPS
- $0.90 (GAAP $0.29, down from $0.37 PY)
- Customers with >$5M ACV
- 658, +~23% YoY
- Deals >$1M net-new ACV
- 123 in Q2, up nearly 40% YoY
- Renewal rate
- 98% for both Q2 and 1H FY26 - unchanged from the prior-year periods (10-Q)
- Non-seat-based share of net-new business
- ~50% (management, Q2 call)
- Net debt
- ~$0.8B at Jun 30 2026, versus a net-cash position previously (computed from the Q2 balance sheet)
Growth drivers
- Now Assist (GenAI) — >$1M ACV customers +130% YoY, deals with 3+ Now Assist products +~70% YoY
- Shift to non-seat/consumption pricing — ~50% of net-new business now non-seat-based (tokens/assets)
- Platform expansion beyond ITSM into CRM, HR, security (Armis), and agentic AI orchestration (AI Control Tower)
- New-logo ACV +50% YoY; multi-product land-and-expand with high net retention
- Federal/government and regulated-industry demand
Reported financials — SEC EDGAR
Audited GAAP figures pulled from SEC filings · latest filing 2026-01-29. The audited primary-source spine — not financial advice.
Revenue — annual (GAAP)
Margins & balance sheet — FY’25
Bull & bear
The thesis got its proof point: ServiceNow AI crossed $1B in ACV and is tracking ahead of the $1.5B end-2026 target, which is precisely the evidence that agentic AI shows up as consumption revenue rather than seat destruction. On top of that, the company beat the high end of guidance on every topline and profitability metric again, raised the full-year subscription guide, and now has a $1B+ security franchise from Armis and Veza sitting under the AI Control Tower. Durability is unchanged - 98% renewal, 658 customers over $5M ACV, $29.0B of RPO.
- ServiceNow AI crossed $1 billion in ACV in Q2 FY26 and management says the company is 'on track to beat our target of 1.5 billion ACV by the end of 2026' - the hardest evidence yet that AI is additive rather than cannibalistic (fact, release + call).
- Beat the high end of guidance across every topline and profitability metric: subscription revenues $3,877M ran 150bps above the guide, cRPO cc growth of 21.5% ran more than 2 points above, and non-GAAP operating margin of 29.5% ran ~3 points above; FY26 subscription guide raised to $15,760-15,780M (fact).
- Growth accelerated at scale: subscription +24.5% (+23% cc) and total revenues +24% on a ~$16B run-rate, with 123 deals over $1M in net-new ACV (+~40% YoY) and 658 customers over $5M ACV (+~23%) (fact).
- Security is now a real second engine: Armis (closed Apr 20 2026, ~$7.6B preliminary consideration) plus Veza (Mar 2 2026, ~$1.2B) give end-to-end cyber-asset and identity coverage under the AI Control Tower, which management describes as a $1B+ and fastest-growing product line (fact on the deals; the segment characterisation is management's).
- Retention and visibility are unchanged: renewal rate 98% in both Q2 and 1H FY26, matching the prior-year periods, with $29.0B of RPO (+21%) and $13.20B of cRPO (fact, 10-Q and release).
- Valuation still discounts the growth: 24.55x forward P/E at the $111.23 Jul 31 2026 close against a $140.25 mean target across 49 firms (fact, market data).
The forward-looking metric is decelerating while the trailing one accelerates on acquisitions. cRPO growth slipped to 21.5% cc from 22.5% in Q1 and Q3 is guided to 19.5% reported - and that is with Armis and Veza in the numbers, whose contribution the company declines to quantify. Part of the Q2 topline beat was federal on-premise revenue pulled forward out of Q3. Underneath, the economics are eroding: subscription gross margin is down 250bps year over year with the full-year guide trimmed, GAAP operating margin fell to 4%, operating cash flow declined, and the net-cash cushion is gone after Armis.
- cRPO - the 12-month forward visibility metric - is decelerating: 21% reported / 21.5% cc in Q2 versus 22.5% cc in Q1, and Q3 is guided to 19.5% reported / 20% cc against an estimated ~$35M FX headwind. That is the first sub-20% reported cRPO guide (fact).
- The 24.5% headline is partly inorganic and the organic rate cannot be verified from the filings: the 10-Q states that pro-forma results for the 2026 acquisitions 'have not been presented ... because the effects of these acquisitions were not material to our consolidated results of operations,' so ServiceNow quantifies neither Armis's nor Veza's contribution (fact, 10-Q).
- Part of the beat was borrowed from Q3: the release attributes the on-premise mix outperformance to US Federal demand that 'accelerated some on-premise subscription revenues from Q3 2026 into Q2 2026,' and $149M of Q2 subscription revenue was recognized upfront on self-hosted delivery versus $109M PY (fact, release + 10-Q).
- Margin structure is degrading on both AI compute and deal accounting: non-GAAP subscription gross margin 80.5% vs 83% PY with the FY26 guide trimmed to 81%; GAAP operating margin collapsed to 4% from 11% on $655M SBC, $219M purchased-intangible amortization and $75M of business-combination costs; GAAP diluted EPS $0.29 vs $0.37 (fact).
- The balance-sheet cushion is gone: net cash flipped to ~$0.8B net debt after the Armis close, funded by a $4.0B term loan (refinanced into senior notes) plus $2.1B of commercial paper at a 3.98% weighted-average rate - and Q2 operating cash flow FELL to $587M from $716M PY (fact).
- The valuation asymmetry has substantially closed: the stock is +21% off its Jul 23 post-print close and forward P/E is now 24.55 - so the same growth debate is now being had at a higher price (fact, market data).
What it is worth
Reverse-DCF / multiple sanity check vs. consensus (web-grounded Jun 2026).
~$70-85
growth decelerates into the teens as agents compress seats, margin erodes on AI compute, multiple stays compressed/de-rates further; near the 52-week low (~$81).
~$110-130
~18-20% growth, ~35% FCF margin, multiple modestly re-rates from depressed ~14x EV/EBITDA as the AI-displacement fear proves overdone.
~$160-185 (split-adj)
growth holds 20%+, AI consumption ARR scales (toward $1.5B Now Assist), multiple re-rates to ~28-32x forward FCF — back toward prior-year levels.
At ~$100B EV with ~35% FCF margin on ~$15B revenue, FCF is ~$5.3B → ~19x EV/FCF and ~14x EV/EBITDA (historic low), forward P/E ~20.6, P/S ~7. The price implies durable ~20%+ growth fading to low-teens with stable ~35% FCF margin — i.e. the market is pricing only modest agentic-AI upside and meaningful seat-displacement risk. If NOW holds 20%+ growth and converts AI to consumption ARR, the multiple is too low; if growth breaks toward the teens on seat erosion, ~20x forward is fair-to-rich.
SWOT
Strengths
- Category-leading, sticky platform of record for enterprise IT/workflow — high switching costs, ~98%+ renewal
- Rare combo of scale + growth: $13B+ revenue still +22%, with 32% op margin and 35% FCF margin
- Deep AI distribution — Now Assist embedded across an installed base of large enterprises; NVIDIA + Microsoft partnerships
- Strong balance sheet (net cash, $5.44B operating cash flow, $5B buyback)
Weaknesses
- Per-seat licensing legacy exposes it to the agentic-AI 'fewer seats' narrative
- Multi-year growth deceleration (30%+ → low-20s) as it approaches ~$16B scale
- AI inference cost is a structural gross-margin headwind; Armis adds near-term margin dilution
- Heavy reliance on large-enterprise/government deal timing (e.g. delayed Middle East on-prem deals cost ~75bps in Q1)
Opportunities
- Reposition as the agentic-AI 'control tower' / orchestration layer — sell more agents, not fewer seats; consumption pricing can decouple revenue from headcount
- CRM land-grab against Salesforce; security expansion via Armis
- International + public-sector penetration
- AI ARR scaling toward a stated $1.5B+ Now Assist ambition
Threats
- 'SaaSpocalypse' — AI agents (e.g. Anthropic Claude Cowork, Microsoft/Salesforce Agentforce) displacing licensed seats → multiple compression across enterprise SaaS
- Microsoft bundling/co-opetition risk (partner and competitor)
- Macro/geopolitical deal-slippage (Iran/Middle East noted in Q1)
- Valuation de-rating risk if growth dips below ~20%
Moats, dependencies & bottlenecks
Moats
NOW becomes the system of action for IT/workflow across an enterprise; rip-and-replace is painful, renewals ~98%+.
Cross-departmental workflow data and integrations deepen with each module; Now Assist trains on customer workflow context.
Moderate-Strong $13B revenue funds AI R&D and a large enterprise sales motion smaller rivals can't match; NVIDIA/Microsoft alliances extend reach.
Trusted for governance, security and gov/regulated deployments — reinforced by AI Control Tower + Armis.
Dependencies
Infrastructure / COGS AI inference scaling raises hosting cost and is the key gross-margin variable.
AI supplier / partner Apriel/Nemotron model integration and NVIDIA AI Factory reference architectures underpin its agentic-AI roadmap.
Partner + competitor Deep Agent 365 / AI Control Tower integration, but Microsoft is also a bundling competitor in enterprise AI.
Deal timing sensitive to macro and geopolitics; concentrated in big multi-year contracts.
Advantages
- Sticky platform-of-record with ~98% renewal and land-and-expand
- Rare scale-plus-growth (22% at $15B run-rate) with 35% FCF margin
- Largest installed base to distribute agentic AI into; consumption pricing already ~50% of net-new
- Net-cash balance sheet + buyback optionality into a depressed multiple
Weaknesses
- Seat-model legacy exposed to the agentic-AI displacement narrative
- Structural growth deceleration toward the low-20s
- Gross-margin erosion from AI compute + M&A dilution
- Sentiment/valuation overhang disconnected from operating results
Bottlenecks
- AI inference cost-to-serve constraining gross-margin expansion
- Sales cycle length and large-deal timing (lumpy cRPO conversion)
- Re-pricing the seat-based book toward consumption without disrupting renewals
- Armis integration execution and near-term margin dilution
Top signals & trends
Top signals
Operating performance accelerating to 22% even as the stock falls.
Evidence AI is monetizing as consumption, not cannibalizing seats.
Market pricing a secular seat-displacement threat to per-seat SaaS.
Multiple reset hard for a 22% grower; asymmetry if growth holds.
AI compute + Armis dilute the margin trajectory.
Trends
The central swing factor — existential threat if it shrinks seats, large tailwind if NOW becomes the agent control/orchestration layer.
EV/Sales fell ~5.6x→~4.2x sector-wide; NOW de-rated to historic-low multiples.
AI Control Tower + Armis position NOW for the agent-oversight market across Microsoft/NVIDIA ecosystems.
Favors scaled incumbents like NOW if it converts the AI narrative.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Compute + reasoning models (Apriel/Nemotron, AI Factory reference architectures) powering NOW's agentic AI.
Cloud hosting + Agent 365 integration; key infra and AI-ecosystem supplier.
Cloud infrastructure provider for portions of the Now Platform / partner ecosystem.
~85%+ of the Fortune 500 / Global 2000 across banking, telecom, public sector; e.g. financial institutions, healthcare, federal agencies.
Both a large customer and a major implementation partner deploying Now Platform at scale.
Consultancies use and resell ServiceNow; representative of the SI-driven enterprise demand base.
Closest large-cap peer; Agentforce ($800M ARR +169%) is the application-layer agentic play. Slower growth (~10-13%) but more direct seat-displacement exposure.
Partner AND competitor — Copilot/Agent 365 bundling into enterprise; Azure is also NOW's hosting partner.
HR/finance SaaS overlap with NOW's employee-workflow expansion; same seat-model AI overhang.
ITSM/dev-workflow competition (Jira Service Management) in the mid-market.
Not a direct SaaS competitor but the catalyst — Claude Cowork (Feb 2026) crystallized the agent-displaces-seats fear; an AI-model/agent-platform substitute risk.