OpenAI's enterprise business is on a $5B+ ARR run-rate in mid-2026 — roughly 25% of total company revenue — with 3M+ paid business seats and 600K+ customer accounts.
That's the short answer. The longer answer is more interesting — enterprise is now OpenAI's fastest-growing line by absolute dollars, the Fortune 500 footprint is 92%, and the average contract size is going up, not down. Below is the breakdown of the seats, contracts, and pricing that get the company to $5B and what the path to $10B looks like.
OpenAI Enterprise Revenue in 2026: The Hard Numbers
OpenAI's enterprise revenue is approximately $5B in annualized recurring revenue as of mid-2026, representing about 25% of the company's total $20B+ ARR. The business has 3M+ paid ChatGPT Enterprise and Business seats across 600K+ customer accounts, including 92% of the Fortune 500 with at least one paid deployment. Enterprise ARR roughly doubled from $2.4B at end of 2025 — a faster absolute growth rate than consumer ChatGPT subscriptions.
| Metric | End of 2024 | End of 2025 | Mid-2026 |
|---|---|---|---|
| Enterprise ARR (estimate) | ~$600M | ~$2.4B | $5B+ |
| Paid business seats | ~600K | ~2M | 3M+ |
| Business customer accounts | ~150K | ~400K | 600K+ |
| Fortune 500 paid penetration | ~75% | ~88% | 92% |
| ChatGPT Enterprise list price / seat / mo | $60 | $60 | $60 |
| Average realized price / seat / mo | $55 | $50 | $45-50 |
| Enterprise as % of total OpenAI revenue | ~15% | ~20% | ~25% |
| Net new enterprise seats per quarter | ~250K | ~400K | ~500K+ |
Estimates compiled from The Information, Reuters, Bloomberg reporting, and OpenAI public statements. ARR figures are annualized run-rates, not booked revenue. See the AI Spending dashboard for additional context on enterprise AI revenue across the frontier labs.
How OpenAI Enterprise Revenue Breaks Down by Product
OpenAI's enterprise revenue isn't one product — it's four. ChatGPT Enterprise is the headline SKU, but the API platform, ChatGPT Business, and ChatGPT Edu each contribute material ARR. Here's the rough split as of mid-2026:
ChatGPT Enterprise
~$2.5-3B ARR
$60/seat/mo list, $40-50 realized
Large-org SKU with SSO, admin console, no training on data, GPT-5 + o-series access
ChatGPT Business (Team)
~$800M-1B ARR
$25-30/seat/mo
Smaller-team SKU, 2-user minimum, fastest-growing seat count by percentage
ChatGPT Edu
~$200-400M ARR
Bespoke contracts
Universities and school districts — Cal State system at 460K users is the flagship
Enterprise API consumption
~$1.2-1.5B ARR
Per-token pricing
Built-on-API deployments at large enterprises (separate from consumer/developer API)
ChatGPT Gov
<$100M ARR
Bespoke contracts
FedRAMP-aligned deployment, US federal $1/agency loss-leader deal in market
Note: Enterprise API consumption is hard to cleanly separate from broader developer API revenue. The ~$1.2-1.5B figure represents large-org accounts with dedicated AEs and enterprise contracting terms, not the full $3-4B API line.
The Contracts That Built OpenAI Enterprise Revenue
Most of OpenAI's enterprise ARR comes from breadth — 600K accounts paying somewhere between $50/month and $50M/year — not from a small set of mega-deals. But the publicly disclosed marquee contracts matter because they set the reference price and reference architecture for the next wave of buyers. Here are the ones that actually moved the market:
US Federal Government (GSA)
$1 per agency for 1 year
August 2025 — Strategic loss leader to lock in federal AI standardization ahead of competitors. Stripe-style land-and-expand play, not a revenue contributor.
Cal State University System
460,000 students + 63,000 faculty/staff
Largest ChatGPT Edu deployment; established the higher-education reference architecture. Contract value not disclosed but estimated $20-40M ACV.
T-Mobile (IntentCX)
$100M+ multi-year per Bloomberg
Custom AI platform built on OpenAI models — first major telco platform deal. Now used as the template for enterprise platform engagements.
PwC
100,000 ChatGPT Enterprise seats
Largest single ChatGPT Enterprise deployment at announcement; rolled to all US and UK staff. Estimated $50-70M ARR at realized pricing.
Moderna
Company-wide ChatGPT Enterprise + custom GPTs
Early enterprise lighthouse account — over 750 custom GPTs deployed internally within first six months.
Klarna
Customer service automation via API
Famously claimed AI replaced workload of 700 customer service agents. Highest-profile case study OpenAI uses in enterprise pitches.
Morgan Stanley
16,000 financial advisor seats
GPT-4-powered knowledge tool for FAs — earliest large financial services deployment, now expanded across the firm.
Coca-Cola (Bain partnership)
Enterprise-wide deployment
First major CPG enterprise customer; deployment includes creative, customer service, and supply chain agents.
OpenAI Enterprise Pricing in 2026: List Price vs Realized Price
The OpenAI enterprise rate card is simple. The realized pricing is not. ChatGPT Enterprise is $60 per seat per month on the public rate card, but every large customer is negotiating down. The realized average has slipped from approximately $55 per seat per month in early 2024 to roughly $45-50 in mid-2026 as volume tiers and multi-year commits have become standard.
ChatGPT Plus (consumer)
$20/seat/mo
Not enterprise, but the price anchor every enterprise buyer references
ChatGPT Business (Team)
$25-30/seat/mo
2-user minimum, no training on data, mid-market sweet spot
ChatGPT Enterprise (list)
$60/seat/mo
Published rate card — what shows up in initial quotes
ChatGPT Enterprise (1,000+ seats)
~$40-50/seat/mo
Standard enterprise discount with 1-year commit
ChatGPT Enterprise (10,000+ seats)
~$30-40/seat/mo
Multi-year commit with annual escalators
ChatGPT Enterprise (Fortune 100)
Bespoke, often <$30/seat/mo
Custom contracts include API credits, support tier, dedicated capacity
The gap between $60 list and $30-40 realized at the top end is what makes the simple revenue math hard. A 100,000-seat customer at list would be $72M ARR; the same customer at the realized large-deal price is closer to $36-48M ARR. That ~40% discount is what investors and competitors should be modeling, not the published rate.
OpenAI Enterprise Revenue vs the Competition
OpenAI's enterprise business is the largest among the frontier AI labs, but Anthropic is closing the gap faster than the headline numbers suggest. Microsoft Copilot bundles into a different motion entirely — it's technically the largest AI-as-a-feature business but most of it isn't direct AI revenue. Here's the comparison as of mid-2026:
| Provider | Enterprise ARR (est.) | Paid Seats | Per-Seat Price |
|---|---|---|---|
| OpenAI (ChatGPT Enterprise + Business) | ~$5B+ | 3M+ | $40-60 |
| Microsoft Copilot for M365 | ~$4-5B (AI-attributable) | 10-12M paid | $30 add-on |
| Anthropic Claude for Enterprise | ~$2.5-3B | Estimated 600-800K | ~$60-75 |
| Google Gemini for Workspace | ~$800M-1.2B (AI-attributable) | Bundled into Workspace | $0 add (in Business+) |
| AWS Q Business + Bedrock | ~$500M-1B | Consumption-based | $20-25 + usage |
| Perplexity Enterprise Pro | ~$80-120M | ~150K | $40 |
Estimates compiled from press reports and earnings call disclosures. Microsoft and Google numbers separate the AI-attributable portion of bundled SKUs and are inherently approximate.
Why OpenAI Enterprise Revenue Growth Is Accelerating
The cleanest way to read the OpenAI enterprise number is to separate seat growth from per-seat revenue growth. Seat counts are growing roughly 3x year-over-year — from ~1M paid seats in early 2025 to 3M+ in April 2026. Realized per-seat pricing is slipping ~10-15% annually as discounts and volume tiers take hold. Net of both, ARR roughly doubles per year — and that pace is sustained by four specific dynamics:
Land-and-expand by department
Median Fortune 500 customer started with one BU pilot in 2024; the 2026 cohort is rolling org-wide. The seat-per-customer ratio has roughly 4x'd from early adopters.
Custom GPTs and Workspaces
Customers that build 50+ internal GPTs see 2-3x higher expansion in year 2 vs customers that don't. Moderna's 750+ custom GPTs is the canonical case study OpenAI shows in deal cycles.
Agent-based new SKUs
Operator (autonomous agent) and Deep Research are starting to appear as add-on line items in renewals at ~$10-20 per seat per month uplift.
Government and education wins
Cal State (460K users), the US federal $1 deal, and 20+ state and city agreements are converting students and federal employees into future enterprise users.
What the OpenAI Enterprise Revenue Number Doesn't Include
A few important things sit outside the $5B enterprise ARR figure and are worth flagging when modeling the business:
- →The $13B+ Microsoft commitment — that's a strategic investment and capacity arrangement, not enterprise SaaS revenue. It shows up in different ways in OpenAI's P&L.
- →Stargate compute commitments — the $500B SoftBank/Oracle/OpenAI infrastructure deal is a capex story, not enterprise ARR.
- →Developer API revenue from SMBs and consumer apps built on the API — that ~$3-4B sits in the API line, not enterprise.
- →ChatGPT Plus consumer subscriptions ($20/month consumer SKU) — generates ~$10-11B ARR but is consumer, not enterprise.
- →Apple Intelligence revenue share — minimal as of mid-2026 and not yet a material line.
The Path From $5B to $10B Enterprise ARR
OpenAI is publicly targeting $100B+ total revenue by 2029. Internal projections shared with investors have enterprise hitting $25-30B of that — meaning enterprise needs to roughly 5-6x from today. The path to even half that, $10B in enterprise ARR by end of 2027, requires three things to hold:
1. Seat growth holds at 2x
Need to get from 3M to ~6M paid seats by end of 2027 — slower than today's 3x but harder as the base scales
2. Per-seat ARPU doesn't collapse
Add-on agents and Operator need to offset core ChatGPT Enterprise discount pressure; current ARPU floor is ~$35-40
3. Anthropic doesn't take net new
Anthropic enterprise share doubled in 2025; if OpenAI loses 30%+ of net new Fortune 500 deals, the $10B target slips
None of those are guaranteed. The realized-pricing trajectory in particular is the one to watch — every quarter that ARPU slips faster than seats grow narrows the path. But unlike most software businesses at this scale, OpenAI has the advantage of being the default starting point for new enterprise AI buyers, which keeps the top-of-funnel intact even as competition tightens.
$5B in enterprise ARR, 92% of the Fortune 500 paying, and 3M+ seats — but the real story is the gap between $60 list and $45 realized.
Enterprise discounting is now the single biggest variable in OpenAI's 2027 revenue model.
Track AI revenue and valuation trends on the AI Valuations dashboard and the AI Spending tracker at Value Add VC. Originally published in the Trace Cohen newsletter.