A-Portfolio

The company in one breath

Okta, ticker OKTA on the Nasdaq, sells the digital bouncer that decides which employees, customers, software, and AI agents can enter a company's systems and what each one can touch. Companies pay subscription fees because one stolen identity can turn a quiet Tuesday into a very expensive hostage situation.

Why now

Okta just reported the kind of quarter that forces a second look. Revenue grew 11%, near-term contracted business grew 14%, and free cash flow jumped 40%. The business did not suddenly become a rocket ship. It became cleaner, more profitable, and harder to dismiss.

The market noticed before breakfast. The stock traded at $158.00 at 7:10 a.m. Eastern on August 27, up 17.5% from Wednesday's $134.42 regular-session close. That means buyers added roughly one-sixth to the value overnight. Good for the company, lousy for anyone who wanted a sleepy entry.

The deeper change sits in the product. Every AI agent needs an identity, permission limits, and an audit trail. Okta already guards human access, so machine identity is a logical extra room in the same house. I like that more than a random AI sticker slapped on an old product. I do not know how much agent security can add to revenue yet, and management did not give us that number.

The numbers, translated

• Revenue reached $805 million, up 11%. Revenue is the money customers paid Okta before costs. For a mature subscription software company, double-digit growth clears the basic bar, while 15% or better would show a real second growth act.
• Subscription revenue hit $793 million, up 12%, and supplied 99% of sales. That means nearly every dollar comes from recurring contracts instead of one-off projects. Good looks like more than 90% recurring, so this is the sturdy part of the story.
• Current remaining performance obligations reached $2.585 billion, up 14%. This is signed subscription business Okta expects to turn into revenue within the next year. Good looks like this backlog growing faster than current revenue, and 14% versus 11% does exactly that.
• Total remaining performance obligations were $4.858 billion, up 17%. This is the full pile of signed business not yet counted as revenue. Good looks like the pile outgrowing reported sales without management buying growth through ugly discounts, and this quarter passes the first half of that test.
• Dollar-based net retention was 107%. Take the same customers from a year ago, then include upgrades, cutbacks, and cancellations. They now spend $107 for every prior $100. Good for this kind of software would be 110% or better, so Okta is expanding inside accounts, but not fast enough to brag.
• Customers spending more than $100,000 a year rose to 5,255, up 6.3%. These are the big accounts that make sales efforts worthwhile. Good would be double-digit growth, so the enterprise base is still moving, just not sprinting.
• GAAP operating income hit $107 million, or 13% of revenue, up from $41 million and 6% a year ago. GAAP means the standard accounting scorecard, with stock compensation included. Okta kept 13 cents from each sales dollar after operating costs, and good starts around 10 cents with room to improve, so this is real progress.
• Free cash flow reached $227 million, or 28% of revenue, up from $162 million and 22%. Free cash flow is cash left after running the business and buying the equipment and software it needs. Good for an 11% grower is at least a 20% cash margin, so 28% is excellent.
• Cash and short-term investments totaled $2.299 billion, with no convertible note debt left. That means the balance sheet can fund product work, acquisitions, or repurchases without asking lenders for permission. Good is more cash than debt, and Okta clears that bar cleanly.
• Stock compensation was $114 million, equal to 14% of revenue. This is pay handed out in shares, which costs owners even when cash never leaves the bank. Good would be below 10% of revenue and falling, so this bill still annoys me.
• At $158.00, Okta carries a market value near $27.6 billion and an enterprise value near $25.3 billion. Enterprise value is what the operating business costs after subtracting net cash. That equals about 27.5 times management's midpoint for this year's free cash flow and 8.6 times expected sales. Good for an 11% grower would be closer to the low 20s on cash flow or 6 to 7 times sales, so the premarket pop ate the bargain.

What I like

First, the forward numbers beat the backward numbers. Current contracted backlog grew 14% while revenue grew 11%. That means signed work points to modest acceleration. Good turnarounds show up in the order book before they show up in reported sales, and that is happening here.

Second, profit quality improved. The standard accounting operating margin reached 13%, and the free-cash-flow margin reached 28%. Those numbers mean Okta can fund itself and still invest. Good software companies eventually stop asking investors to grade on vibes. Okta finally brought cash to the exam.

Third, management used $366 million to repurchase about 4.57 million shares during the six months ended July 31, at an average near $80.10 per share. A repurchase reduces the ownership pie, so each remaining share can claim a larger slice. Good repurchases happen below fair value, and buying near half today's premarket quote looks disciplined.

What worries me

The growth rate still starts with a one. 11% revenue growth can support a fine business. It does not deserve a heroic price by itself. Good would be a move toward 15%, led by new identity products rather than a short billing trick.

The 107% net-retention rate says existing customers expanded spending by only 7%. That is positive, but software winners usually get more lift from the customers already inside the building. Good would be 110% or higher, with both Workforce Identity and Customer Identity contributing.

Stock compensation still consumed 14% of revenue. Management bought back shares faster than it issued them during the first half, which helps, but I refuse to call share pay free because the accounting statement prints it in a different neighborhood. Good means this cost drops below 10% while employee retention stays healthy.

Then there is trust. Okta sells security and has suffered security incidents. A locksmith can survive a burglary, but not if he keeps leaving the window open. Good means clean controls, quick disclosure, and no repeat event that damages customer confidence.

What would make this a buy

I want one of two doors.

The price door opens around $140 or lower, my threshold as of August 27. At $140, the operating business would cost about 24 times the midpoint of management's full-year free-cash-flow guide. That is still not cheap, but it gives a decent company room to become a better one. Good entries leave space for one ordinary quarter without turning the thesis into roadkill.

The proof door opens if revenue growth reaches 12% to 15%, current contracted backlog stays above 15%, net retention returns to 110% or better, and free-cash-flow margin holds above 25%. That combination would mean customers buy more, future sales strengthen, and cash stays real. I would pay more for that proof because the business would have earned it.

What makes me walk away

If current contracted backlog falls below 10% growth, the forward engine has stalled. Good needs backlog to grow at least as fast as revenue, not trail it.

If net retention falls to 100% or less, existing customers have stopped expanding in total. Good software gets more valuable after installation, so flat spending would punch a hole in the platform story.

If management cuts full-year free-cash-flow guidance below a 25% margin, the new profit discipline has cracked. Good would be holding the current 28% to 29% guide while still funding product work.

One more serious security failure with slow disclosure ends the conversation. The exact acceptable number is zero preventable trust disasters. For an identity company, good security is not a feature. It is the floor.

Earnings On Deck

• Okta (OKTA), August 26 after the close. The print is in. I watched current contracted backlog, net retention, standard accounting profit, and cash. Good meant backlog growth above revenue growth and a cash margin above 25%, which Okta delivered. Bad would have meant another growth slowdown hidden behind adjusted profit.
• Rubrik (RBRK), August 27 after the close. Watch subscription annual recurring revenue, which is the yearly value of repeat contracts, and free cash flow. Good means recurring revenue stays fast while cash losses narrow. Bad means growth cools before the business becomes self-funding.
• SentinelOne (S), August 27 after the close. Watch annual recurring revenue and operating loss. Good means security demand stays strong while each sales dollar carries a smaller loss. Bad means the company spends harder just to keep growth from slipping.
• Affirm (AFRM), August 27 after the close. Watch merchandise volume, funding costs, and credit losses. Good means more purchases run through the network without borrowers falling behind. Bad means volume grows because Affirm accepts weaker credit and gets stuck with the bill.

The verdict

Score: 77 out of 100, watch. Okta earned points for independent identity leadership, accelerating contracted backlog, clean standard accounting profit, excellent cash generation, net cash, and disciplined repurchases. It lost points because revenue growth still sits at 11%, net retention is only 107%, stock compensation remains heavy, security history still matters, and the overnight rally removed the easy entry.

The bucket is watch. This is a credible turnaround and a possible identity layer for the AI-agent era. At $158, the business deserves respect and the stock demands patience.

The DBTech read

Current remaining performance obligations, usually called cRPO, are signed subscription dollars a company expects to recognize as revenue within one year. Think of revenue as meals already served and cRPO as paid reservations still waiting for a table. When cRPO grows faster than revenue, tomorrow may look better than today. Good cRPO growth comes from real contracts and stays ahead for several quarters. One hot quarter helps, but a streak earns conviction.

Sources and as-of date

Research completed August 27, 2026 at 7:10 a.m. Eastern.

The $158.00 premarket price came from Yahoo Finance's Nasdaq feed at 7:10:17 a.m. Eastern on August 27, 2026. The $134.42 regular-session close came from the same Yahoo Finance feed for 4:00 p.m. Eastern on August 26, 2026. Alpaca's IEX broker feed showed a last trade of $139.29 at 4:01:04 p.m. Eastern on August 26, but its free feed did not provide the live premarket quote, so I used Yahoo for the current price.

Financial statements, customer metrics, share counts, debt, stock compensation, and repurchases came from Okta's Form 10-Q filed with the U.S. Securities and Exchange Commission on August 27, 2026: https://www.sec.gov/Archives/edgar/data/1660134/000166013426000069/okta-20260731.htm

Quarter results and management's fiscal 2027 outlook came from Okta's earnings exhibit filed on August 26, 2026: https://www.sec.gov/Archives/edgar/data/1660134/000166013426000068/okta-7312026_ex991.htm

The earnings calendar came from Nasdaq's calendar for August 27, 2026. The market reaction cross-check came from CNBC's August 26 earnings coverage. The customer-use context came from Capterra's 2026 Okta review page, which listed 944 verified reviews when checked. Community research covered Reddit, Hacker News, and YouTube from July 28 through August 27, 2026. The social sample was thin, so I used it for color, not proof.