Report contents
Industry landscape
Where the company competes: two identity markets, one dominant rival
Okta sells identity software, the layer that decides whether a person, a device or an AI agent gets into an application and keeps a record of what happened afterward. That reads like one product. It is two markets with different buyers, different renewal dynamics and different rivals.
Workforce identity, still the larger piece, sells to a company's IT and security organization: single sign-on, multifactor authentication, lifecycle management, identity governance. Customer identity, largely the Auth0 business Okta acquired in 2021, sells to product and engineering teams that must register and authenticate outside users. Both are subscription software with gross margin near 78%, per the company's reported results, and both are sold mostly to enterprises in the United States, with international a minority of revenue.
The value chain is short, and Okta owns little of it. Upstream is public cloud capacity (Okta runs primarily on AWS) and engineering labor. There is no scarce input, no fabrication capacity, no licensed spectrum. Downstream sit two channels: direct sales to security and IT leaders, and a partner layer of resellers and global systems integrators who fold identity into larger transformation projects. Marginal cost is close to zero, so the economics turn on acquisition cost and retention, not on input prices.
Market structure: a duopoly on the workforce side, a crowd on the customer side
In employee access management, the arena where Okta built its franchise, two vendors account for roughly 70% of the market. CEO Todd McKinnon put Okta at about 30% and Microsoft at about 40% at the company's September 23, 2026 investor summit. Those are the company's own estimates, not an audited third-party measure, and the direction is consistent with what buyers see: Microsoft's share rides on Entra ID being bundled into M365 licenses that enterprises already own.
| Employee access management | Approximate share |
|---|---|
| Microsoft | ~40% |
| Okta | ~30% |
| Everyone else | ~30% |
The fringe is real but fragmented: IBM, One Identity, Ping Identity and its ForgeRock assets, CyberArk in adjacent privileged access, Google, and platform vendors such as ServiceNow that approach identity from the IT workflow side rather than the authentication layer.
Customer identity is a different map. Okta does not disclose share, and no third party publishes a reliable split. The field includes Microsoft Entra External ID, Ping, Amazon Cognito, Transmit Security and, stubbornly, homegrown builds. Developers there will swap a vendor for a cleaner SDK, which makes pricing power thinner than in workforce SSO.
Barriers are high at the top of workforce identity and considerably lower below it. Certifications, audit history and years of integration work make ripping out an identity provider a multi-quarter project for a large enterprise; management points to 17-year customer relationships as evidence of that stickiness. The technology market is global, but procurement is not: data residency rules, FedRAMP-level certification for US public-sector work and local buying preferences carve the addressable market into national pieces. R&D scale is global; go-to-market is local.
The boundary that matters most is not yet drawn. Okta argues agentic identity could become a market larger than today's IAM business, says Okta for AI Agents is the only generally available product in the category, and has added third-party identity provider support so Microsoft shops can adopt it without replacing Entra. Management said current pricing is still per user, with a 25% to 37.5% uplift over base platform pricing in sizable deals, and conceded that how to count and price an agent remains unsettled. Microsoft, IBM and ServiceNow are named as competitors.
So the question for the next three years: does agent identity become its own budget line where an independent vendor can hold price, or does it get absorbed into platform contracts that Okta does not control? The evidence does not yet settle that. What would: disclosed agent-related revenue, and whether the per-user uplift survives the shift to per-agent pricing.
Competitive advantage
Does the company have a moat? Moderate: sticky software, thin returns
Okta has a narrow moat. It is the second of two real competitors in employee identity software, and its product sits deep enough in customers' login flows that replacing it takes quarters. The harder test, sustained profit above the cost of capital, fails on Okta's own reported results. GAAP operating losses totaled roughly $2.0 billion across the five fiscal years from FY2022 through FY2026, and operating income turned positive only in FY2026, at $0.15 billion. The entry barriers are real. So far the returns they shelter have accrued mainly to Microsoft.
Market position: a durable second place with about a third of the market
Okta's core market is workforce identity: single sign-on, multifactor authentication, and the lifecycle and governance tooling around them. At the company's September 23, 2026 investor summit, CEO Todd McKinnon put Okta at about 30% of that market and Microsoft at about 40%, with the remaining third split among IBM, One Identity, Ping, CyberArk in adjacent privileged access, and a long tail. Those are company estimates, not an audited third-party measure. They match the structure buyers describe, in which Microsoft's Entra ID arrives inside Microsoft 365 licenses enterprises already pay for.
No annual share series exists for this submarket, from Okta or from an independent body, so the five-year stability test cannot be run cleanly. What a stable position should produce is visible: gross margin of 77.4% in FY2026 and 79.6% in the quarter ended July 31, 2026, on FY2026 revenue of $2.92 billion. A vendor losing ground to a bundled rival usually shows it in price first.
Market position: Fairly Strong. Okta ranks second with roughly 30% of a two-vendor market, and the fringe is split among subscale vendors. The gap in the evidence is proof that the 30% has held; one management estimate does not establish that.
| Fiscal year (ends January 31) | FY2022 | FY2023 | FY2024 | FY2025 | FY2026 |
|---|---|---|---|---|---|
| Revenue (USD bn) | 1.30 | 1.86 | 2.26 | 2.61 | 2.92 |
| Gross margin | 69.5% | 70.6% | 74.3% | 76.3% | 77.4% |
| Operating income (USD bn) | -0.77 | -0.81 | -0.52 | -0.06 | 0.15 |
| Net margin | -65.2% | -43.9% | -15.7% | 1.1% | 8.1% |
| Free cash flow (USD bn) | 0.09 | 0.06 | 0.49 | 0.73 | 0.86 |
Profitability: under the cost of capital even as losses narrow
Take the balance sheet at January 31, 2026: $7.00 billion of equity, $0.35 billion of debt and $0.86 billion of cash, or about $6.5 billion of invested capital. After-tax operating income was roughly $0.13 billion, putting ROIC near 2%. Using EBIT of $0.26 billion less $0.02 billion of tax gets to $0.24 billion, just under 4%. Both figures sit below the 15%-25% range that would mark a clear advantage, and below even the 6%-8% that signals none.
Goodwill of $5.49 billion is 84% of that invested capital, almost all of it from the 2021 Auth0 purchase. Strip the goodwill out and returns look enormous, but that flatters the business: the goodwill is what Okta paid for customer identity, and the test is whether the purchase earned its price. Clearing an 8% cost of capital on $6.5 billion requires about $0.52 billion of after-tax operating income. Okta's cash tax rate is near zero because of accumulated losses, so that implies roughly $0.55 billion to $0.70 billion of pre-tax operating income, a GAAP operating margin of 19% to 24% on FY2026 revenue. The company reported 5% for the year and 13.6% in the July 2026 quarter. The direction is right; the level is not there yet.
Free cash flow complicates the picture. FY2026 free cash flow was $0.86 billion, 29% of revenue, a wide spread over $0.15 billion of GAAP operating income. Stock-based compensation added back in the cash flow statement and cash collected ahead of revenue recognition on multi-year contracts explain most of the difference, and Okta reports both. For a company paying a meaningful share of employee costs in equity, free cash flow overstates economic profit.
Profitability: Weak. Revenue compounded at roughly 22% a year from FY2022 to FY2026 and gross margin rose 8 percentage points, yet returns on total invested capital have not cleared the cost of capital in any year of the period.
Overall rating: Moderate
Overall rating: Moderate. Okta passes the concentration test and fails the profitability test, and the split is informative. A market where two vendors hold 70% and the follower keeps 77%-plus gross margin has genuine entry barriers. A follower that took nine years as a public company to post its first positive operating income, and that earns about 4% on capital including acquisition goodwill, has not yet converted those barriers into protected economics. Microsoft's bundling is the most plausible explanation, though the evidence here establishes the outcome rather than the cause.
Two things would move the rating. The first is a GAAP operating margin sustained in the high teens for several quarters without revenue growth falling to single digits; on today's capital base that would push ROIC toward double digits. The second is a disclosed retention or share figure showing the roughly 30% has held since before the Auth0 deal. A goodwill impairment would lift the ROIC ratio without improving the business, and should not be read as progress.
Where the advantage comes from and how long it can last: stickiness in the login, capped by the bundle
Two sources carry what protection Okta has: customer captivity in workforce identity, and scale in engineering and go-to-market that smaller rivals cannot match. Neither works against Microsoft. That asymmetry, more than anything in Okta's execution, sets the duration.
Customer captivity: Fairly Strong in workforce identity, Moderate in customer identity
The mechanism is concrete. Okta sits in the request path: every application a company runs hands the login to Okta, verifies the user, applies policy and logs the result. Removing it means re-pointing each application at a new provider, re-testing every integration, and re-documenting access controls for auditors who signed off on the old design. That work lands on internal IT, which has no incentive to volunteer for it.
Procurement reinforces the technical lock. Large buyers select identity vendors on evidence of past operation: FedRAMP authorization for US public-sector work, SOC 2 reports, a reference list of comparable customers. Those credentials take years to assemble, which raises the cost of being new more than it raises the cost of switching. Identity is also a small line in an IT budget next to the cost of an outage or a failed audit, which damps price sensitivity at renewal.
The visible evidence fits. Gross margin held between 74% and 79% over five years while revenue compounded at about 22%, and management points to customer relationships running 17 years. The counter-evidence is growth: revenue rose 11% to $1.58 billion in the first half of FY2027 from $1.42 billion a year earlier. Slower growth shows the renewal base alone no longer carries the number. It does not by itself show share loss, and Okta does not break out retention in its financial statements, so the size of any churn is not publicly measurable.
The two halves of the business differ. In workforce identity the buyer is a security organization that prizes continuity; captivity is Fairly Strong. In customer identity, the Auth0 side, the buyer is a product team that will migrate for a cleaner SDK and better pricing, and Amazon Cognito and Microsoft Entra External ID anchor price low. There captivity is Moderate at best.
Economies of scale: Moderate, and useful mainly against the smaller rivals
Engineering is the fixed cost in identity. Okta has spent a decade building coverage across single sign-on, multifactor authentication, lifecycle management and governance, and spreading that development budget over $2.92 billion of revenue gives a per-dollar cost that Ping, One Identity or IBM cannot match. That gap shows up as breadth: a mid-market buyer can source one platform instead of stitching three products together.
Relative share decides whether scale protects anything, and Okta's share is about three-quarters of Microsoft's. Bundling lets Microsoft spread Entra's development cost across an installed base of Microsoft 365 seats no identity vendor can replicate, and lets it treat the identity component as free at the margin. Against the leader, Okta's scale is a disadvantage; against the fringe, it is a real barrier, because matching the platform would require years of spending with no visible return. The partner channel through resellers and systems integrators is a smaller version of the same story: costly to build, replicable in time, and no defense against a bundle.
Economies of scale: Moderate. Scale keeps Okta ahead of everyone except the one competitor that matters.
Durability: the bundle sets the ceiling and agents could move the floor
Microsoft's bundling is the binding constraint. For a Microsoft-centric enterprise the marginal price of a competent identity provider is already zero, and Okta's answer is depth and neutrality: stronger governance, multi-cloud coverage, and an access layer not owned by the productivity vendor. Independence wins deals among buyers who dislike that concentration, and there are enough of them to support a 30% share. Independence does not stop Microsoft from pricing low, and it does not create switching costs of its own. It is positioning, not a moat.
Cheap developer identity caps the Auth0 half. Amazon Cognito and Entra External ID cover basic registration and authentication at low cost, and small engineering teams still build their own. That is a good-enough substitute for a meaningful slice of the customer identity market and a direct constraint on price in the business carrying $5.49 billion of goodwill.
Agent identity could reset the unit of account. Management argues the category could grow larger than today's access management, says Okta for AI Agents is the only generally available product in it, and has added support for third-party identity providers so Microsoft shops can adopt it without replacing Entra. Current pricing stays per user, with a 25% to 37.5% uplift over base platform pricing in sizable deals; management concedes how to count and price an agent is unsettled. If seats stop being the unit, Okta's installed base of named users matters less and the competitive contest restarts on a field where Microsoft, IBM and ServiceNow are all present.
The workforce moat should hold through the next five years in large enterprises, because the switching cost is embedded in audit trails and application integrations that no buyer wants to rebuild. The customer identity half is weaker and faces a free substitute. Note where the protection actually operates: it shows up at renewal, when a customer weighs the cost of leaving, and much less at the point of sale, where Microsoft's bundle sets the reference price. Whether agent identity becomes a separate budget line where an independent vendor can hold price is the open question, and the evidence so far only shows that Okta shipped a product and that its own pricing model is unresolved.
Generated by DeepVal AI. Data and models may contain errors or be outdated. For research only; not personalized investment advice.
Continue reading the full report
- Comparison and summaryHow Okta Inc compares with its closest rivals, and our moat rating
Free to sign up · 20 research points · unlocking this report uses 10
20 points cover 2 full reports. No card needed.
Research reflects its generation date. Preview the first two modules.
More on Okta Inc
Okta Inc
Business, moat and key questions
Open company page