Glean
B2B SaaS with a hybrid pricing model: fixed per-active-user platform fee plus consumption charges for LLM/agent usage; land-and-expand across departments (company reports 85%+ of customers deployed across 5+ departments).
Chronological priced-round trail: ~7x step-up in valuation over ~3 years (approx $1B in 2022 to $7.2B in Jun 2025), tracking ARR from sub-$50M to ~$100M+ at the Series F. The 2022 Series C mark and lead are lower-confidence than the 2024-2025 rounds, which are well-documented in primary press releases.
Earnings, margins, COGS & capex
Glean is one of the fastest-scaling enterprise-AI companies of this cycle: ~$100M ARR (FY ended Jan 2025) to $208M (Dec 2025) to ~$300M (May 2026), a ~3x increase in 15 months. Growth is expansion-led (seats + agent consumption within large accounts) with Fortune 500 logo count nearly doubling YoY. Company does not disclose profitability, gross margin, or net retention; per TechCrunch the '$300M ARR' includes consumption revenue that is not strictly recurring, so the headline slightly flatters underlying subscription durability.
Revenue trend
Margins
structurally pressured vs classic SaaS by pass-through LLM inference cost; Glean argues its context graph cuts tokens/cost per query
presumed negative - still in hypergrowth land-grab, ~1,648 employees
implied strong; expansion across 5+ departments and agent adoption cited by company as core economics driver, but not a disclosed figure
COGS structure
Primary COGS is LLM inference/model consumption (Glean is model-agnostic, routing to Anthropic/OpenAI/Google and open models) plus cloud hosting and indexing compute for the per-customer knowledge graph, and customer-success/onboarding for large-enterprise deployments. Consumption pricing is partly designed to pass model cost through to customers.
Capex
Minimal physical capex; 'investment' is R&D headcount (~1,648 staff as of May 2026) and compute. No data-center build-out disclosed (runs on public cloud).
Latest earnings
N/A (private)
No formal guidance; management narrative targets continued hypergrowth and positions Glean as the enterprise AI budget consolidator.
- ARR
- ~$300M (May 2026)
- YoY growth
- ~89% (May 2025 to May 2026)
- Employees
- ~1,648 (May 2026)
- Departmental breadth
- company-reported 85%+ of customers in 5+ departments
- Last valuation
- $7.2B (Series F, Jun 2025)
Growth drivers
- Land-and-expand — seat growth then department expansion (company reports 85%+ of customers across 5+ departments)
- Shift from Glean Assistant (search/Q&A) to Glean Agents (no-code agentic automation) as a consumption-revenue upsell layer
- Fortune 500 penetration nearly doubling YoY; large-enterprise average contract values
- Vendor-neutral positioning wins multi-ecosystem enterprises (Microsoft + Google + Slack + Atlassian + Salesforce coexisting)
- AI-budget-consolidation pitch — replace many point AI tools with one platform as enterprises rationalize AI spend
Bull & bear
Glean is the emerging neutral system-of-context for the enterprise - a real data/permissions moat, best-in-class growth, and a consumption-led agent upsell that turns AI-budget consolidation into a durable expansion flywheel.
- Permissions-aware cross-app knowledge graph is hard to rebuild and gets stickier as more apps/agents plug in - classic data-network + switching-cost moat
- ~3x ARR in 15 months to ~$300M with Fortune 500 logos nearly doubling shows enterprise product-market fit, not a pilot fad
- Vendor-neutrality is a structural wedge: the median large enterprise runs Microsoft AND Google AND Slack AND Atlassian AND Salesforce, and only Glean spans the union
- Glean Agents shifts the model from per-seat SaaS to consumption, expanding the wallet as agentic workloads scale
- AI-budget-cutting is a tailwind: consolidating many AI point tools onto one platform is Glean's explicit 2026 selling point
- Top-decile cap table and $150M fresh Series F capital give it runway to out-invest most independent rivals
Glean is a well-run horizontal app squeezed between hyperscalers who bundle 'good-enough' search/agents at near-zero incremental cost and the very model vendors it depends on - at a ~24x ARR valuation that leaves no room for growth or margin disappointment.
- Microsoft (M365 Copilot) and Google (Gemini/Agentspace) can bundle enterprise search into suites 450M+ seats already pay for, structurally undercutting a standalone platform
- Undisclosed gross margin and the LLM-inference pass-through mean unit economics may be materially below classic SaaS; the moat may not convert to profit
- '$300M ARR' is partly consumption run-rate, so revenue is less recurring/predictable than the label implies and can wobble with usage
- Existential supplier risk: Anthropic/OpenAI/Google supply the models AND increasingly compete for the enterprise connector/agent layer
- ~24x ARR private mark discounts years of flawless execution; enterprise AI-budget tightening or a growth stumble invites multiple compression / down-round at IPO
- Horizontal 'AI for all your work' is a crowded thesis (Copilot, ChatGPT Enterprise, Agentspace, Moveworks/ServiceNow, Guru, Dust) - differentiation must be continuously re-earned
What it is worth
Last-priced private round + revenue-multiple cross-check (no public market price).
Hyperscaler bundling compresses growth and pricing, consumption revenue proves volatile, and thin/undisclosed margins disappoint; the ~24x mark compresses to a mid-single-to-low-teens ARR multiple, implying a flat-to-down round or a discounted IPO (materially below $7.2B).
Growth decelerates but stays strong (40-60%); market rewards durable enterprise AI leadership. Valuation holds around the ~$7.2B Series F mark, re-rating gradually as margins are disclosed and NRR is proven.
If Glean sustains 60%+ growth toward ~$600M-$1B ARR with proven gross margins and NRR while establishing itself as the neutral enterprise context layer, an IPO at a premium multiple could support a valuation well above the $7.2B mark.
Series F (Jun 2025) set a $7.2B post-money on a $150M raise led by Wellington Management, up from a $4.6B Series E (Sep 2024) and a $2.2B mark in early 2024. Against ~$300M ARR (May 2026) that is ~24x current ARR; at the June 2025 round the company had only just surpassed ~$100M ARR, so the mark was underwritten at a far richer forward multiple. Rich but roughly in line with how top-decile enterprise-AI names were priced in 2026, justified only by ~89% YoY growth and category-leadership positioning. No public float; any exit is IPO or secondary.
SWOT
Strengths
- Permissions-aware enterprise knowledge graph across 100+ connectors - a genuine data-integration moat that is slow to replicate
- Vendor-neutral horizontal platform: indexes all apps rather than favoring one suite, unlike Microsoft/Google
- Elite founder/team pedigree (Arvind Jain, ex-Google distinguished engineer in Search/Maps/YouTube and Rubrik co-founder) and top-tier cap table (Sequoia, Kleiner Perkins, Lightspeed, Coatue, DST Global, ICONIQ, Wellington, Khosla, General Catalyst, Altimeter)
- Exceptional growth durability: ~3x ARR in 15 months with expanding large-enterprise logos
Weaknesses
- Undisclosed unit economics (gross margin, NRR, burn) — consumption/inference cost structure is thinner than classic SaaS
- Headline '$300M ARR' blends consumption run-rate that is not strictly recurring (per TechCrunch)
- Horizontal product must continually justify ROI vs bundled-in-the-suite alternatives that cost customers near-zero incremental
- Deep dependence on third-party frontier LLMs it does not own
Opportunities
- Agentic automation (Glean Agents) expands TAM from 'search' to 'work execution' - larger consumption wallet
- Enterprise AI-spend consolidation cycle favors a single horizontal platform over sprawl of point tools
- International + regulated-industry expansion (Series F earmarked for global agent innovation)
- Becoming the neutral 'context layer'/system-of-record that other enterprise AI tools call into
Threats
- Microsoft Copilot bundled into M365 (450M+ commercial seats) and Google Gemini/Agentspace bundled into Workspace - distribution + zero-marginal-price pressure
- OpenAI ChatGPT Enterprise and Anthropic moving up-stack into enterprise connectors/agents
- Frontier-model vendors (its own suppliers) can build native enterprise search/graph and disintermediate
- AI-budget tightening cuts both ways - could compress deal sizes if ROI isn't crisp
- Rich ~24x ARR private valuation sets a high bar — a down-round/IPO-multiple compression risk if growth decelerates
Moats, dependencies & bottlenecks
Moats
per-user access control) Hard to replicate the breadth + permission fidelity; deepens with each connected app. Erodes if hyperscalers make native cross-suite indexing good enough.
Once Glean indexes all systems and 5+ departments rely on it, rip-and-replace is painful; agents built on Glean deepen lock-in.
Structural wedge vs suite vendors, but neutrality is a positioning advantage, not a technical barrier.
Elite team and funding enable out-investment, but not a durable defensive moat on their own.
Query and agent-usage data could compound relevance; not yet demonstrated as decisive vs frontier-model quality.
Dependencies
Supplier + potential competitor Model quality, price, and availability drive both product quality and COGS; these same vendors are moving into enterprise search/agents.
Infrastructure supplier Hosting and inference compute; Azure/GCP are also owned by direct competitors Microsoft/Google.
Microsoft 365, Google Workspace, Atlassian, Salesforce, ServiceNow, Notion, etc.) Integration surface Product value depends on connector access; a suite vendor could restrict/deprecate APIs or bundle rival search to disadvantage Glean.
Concentrated in big-ticket enterprise deals; sensitive to AI-spend rationalization (cuts both ways).
Well-funded post-Series F, but hypergrowth spend implies ongoing capital reliance until profitability.
Advantages
- Broadest permissions-aware index across heterogeneous enterprise app stacks
- Vendor-neutral single pane over the union of a company's tools
- Category-leading growth and reference logos (Databricks, Reddit, Pinterest, Samsung, Duolingo, Instacart)
- Platform breadth: search + assistant + no-code agents, expanding from knowledge access to work execution
- Deep-pocketed, blue-chip investor base and founder credibility
Weaknesses
- Opaque unit economics (margin, NRR, burn undisclosed)
- Consumption-blended revenue less predictable than pure subscription
- Sandwiched between bundling hyperscalers and its own model suppliers
- Premium valuation leaves little margin for error
- Horizontal thesis faces persistent, well-capitalized competition
Bottlenecks
- Inference cost per query/agent run — margin scales only if the context graph keeps token usage down as agent workloads grow
- Enterprise sales + deployment cycle and security review for indexing sensitive corpora limits velocity
- Reliance on external model roadmaps for capability gains it cannot fully control
- Proving hard ROI vs 'free/bundled' suite copilots inside cost-conscious buying committees
Top signals & trends
Top signals
Demonstrated enterprise PMF and expansion, not a pilot bubble.
Up-market penetration and larger ACVs.
Expands wallet (bull) but makes revenue less recurring and margin more inference-dependent (bear).
Bundled-suite copilot conversion remains a small share of the installed base, leaving room for a best-of-breed neutral platform.
Supplier-turned-competitor disintermediation risk.
Rich multiple; IPO or next round must be supported by sustained growth + margin proof.
Trends
Directly Glean's 2026 selling point as AI budgets tighten.
Grows TAM and consumption revenue if Glean owns the context layer agents call into.
Near-zero incremental price pressures standalone platforms.
Improves gross margin over time but also lowers the barrier for competitors and for build-your-own.
Favors Glean's permissions-aware architecture over bolt-on solutions.
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.
Frontier LLM provider (also a competitor).
Frontier LLM provider (also a competitor).
Models + cloud infra (also a competitor).
Cloud/model infra option (also the top competitor).
Cloud hosting/inference infrastructure.
Underlying GPU compute powering inference (indirect).
Named enterprise customer (TechCrunch).
Named customer (TechCrunch).
Named customer (TechCrunch).
Named large-enterprise customer (TechCrunch).
Cited Glean customer.
Cited Glean customer.
Early customer.
Bundled into 450M+ M365 seats; primary distribution threat, though paid Copilot conversion is still ~20M seats (~4-5% of the base as of Apr 2026).
Agentspace targets multi-agent enterprise workflows; bundled into Workspace.
Moving up-stack into enterprise connectors/knowledge; also a key model supplier to Glean.
Enterprise Claude with connector ecosystem; supplier-turned-rival dynamic.
Acquired Moveworks (2025) to fold agentic enterprise assistant into its platform.
Agent platform anchored in CRM data; overlaps on agentic work automation.
Rovo enterprise search/agents across Atlassian + connected apps.
Enterprise AI search / knowledge management competitor.
Agent/assistant platform positioned as a Glean alternative.