JFrog released its second-quarter 2026 report on Thursday, 6 August 2026, and we are writing about it today.
Verification: the report was filed with the SEC on Form 8-K on 6 August 2026, carrying Item 2.02 (Results of Operations and Financial Condition) and Item 9.01, accession number 0001193125-26-337998. The earnings release was furnished as Exhibit 99.1 (frog-ex99_1.htm) and signed by CFO Eduard Grabscheid. The quarter ended 30 June 2026, and the earnings call was held the same day at 2:00pm Pacific.
Every number in this article comes straight from that filing, and only from it. Even so, errors, inaccuracies or omissions are possible, and the figures may change after publication. Spotted something that looks wrong? Write to me and I will correct it.
One note for anyone going to look the filing up: JFrog is incorporated in Israel, but it reports to the SEC as a US domestic filer, meaning Forms 8-K and 10-Q, rather than the Form 6-K used by most Israeli companies listed on Wall Street. Its EDGAR identifier is CIK 0001800667, which I confirmed from two independent sources rather than assumed.
What the company does
Modern software is built out of thousands of pieces. Some are written in house, but the overwhelming majority are pulled in from outside: open-source libraries, runtimes, models, vendor components. Once code is compiled it becomes what the industry calls an artifact, a packaged, ready-to-run piece of software. All of those pieces need somewhere to live: to be stored, versioned, scanned and shipped out to servers around the world.
That is exactly what JFrog sells. The company provides a platform that acts as the organization's central warehouse for software components, together with tools that scan them for security holes and problem packages and then distribute them into production. The field is called DevOps, which in plain language means everything that connects the developer typing a line of code to the moment that line runs at a customer. The other recurring term is the software supply chain, the path every external component takes into your product, which happens to be exactly where attackers like to walk in.
There are two commercial models. Some customers run the software on their own servers, which is called self-managed, and some consume it as a service from JFrog's cloud, which is SaaS. Both are sold on annual subscriptions. The top tier is called Enterprise+, the full end-to-end platform, and the company likes to point at it because it marks customers who have committed deeply.
Two more terms that come up below. ARR is annual recurring revenue, meaning what a customer pays on an annualized basis. Net dollar retention answers one simple question: take the group of customers you had a year ago, add no new logos at all, and ask what they pay today versus back then. Above 100 percent means existing customers are expanding.
The quarter
| Metric | Q2 2026 | Q2 2025 | Change |
|---|---|---|---|
| Revenue | 163.8 million dollars | 127.2 million dollars | Up 29 percent |
| Cloud revenue | 87.5 million dollars | Not disclosed in dollars | Up 53 percent |
| Subscription revenue (self-managed and SaaS) | 155.5 million dollars | 121.1 million dollars | Up 28.5 percent |
| License revenue (self-managed) | 8.2 million dollars | 6.1 million dollars | Up 33.8 percent |
| GAAP gross profit | 127.6 million dollars, 77.9 percent | 97.0 million dollars, 76.3 percent | 160 basis points of expansion |
| Non-GAAP gross margin | 83.2 percent | 83.1 percent | Just 10 basis points |
| GAAP operating loss | 13.2 million dollars, 8.1 percent of revenue | 26.0 million dollars, 20.4 percent of revenue | Loss narrowed by 12.7 million dollars |
| Non-GAAP operating income | 32.6 million dollars, 19.9 percent | 19.4 million dollars, 15.2 percent | Up 68.4 percent |
| GAAP net loss | 4.2 million dollars | 21.7 million dollars | Loss narrowed by 17.5 million dollars |
| GAAP loss per share, basic and diluted | 0.03 dollars | 0.19 dollars | Still a loss |
| Non-GAAP net income | 34.3 million dollars | 22.0 million dollars | Up 55.7 percent |
| Non-GAAP diluted EPS | 0.27 dollars | 0.18 dollars | Up 50 percent |
| Cash flow from operations | 57.1 million dollars | 36.1 million dollars | Up 58.1 percent |
| Free cash flow | 53.7 million dollars | 35.5 million dollars | Up 51.6 percent |
| Net dollar retention | 121 percent | 118 percent | Up 3 points |
| Customers with ARR above 1 million dollars | 97 | 61 | Up 59 percent |
| Customers with ARR above 100 thousand dollars | 1,291 | 1,076 | Up 20.0 percent |
| Enterprise+ share of revenue | 59 percent | 55 percent | Up 4 points |
| Share-based compensation | 39.6 million dollars, 24.2 percent of revenue | 38.0 million dollars, 29.9 percent of revenue | Intensity down sharply |
| RPO, remaining performance obligations | 659.0 million dollars | Not disclosed in dollars | Up 38 percent |
Two notes on the table. First, on the loss lines I have not computed a percentage change, because both periods are losses and a percentage off a negative base is a meaningless number. The honest comparison is how many dollars the loss narrowed by. Second, on cloud and RPO the company gave only the growth rate, not the year-ago dollar amount. Working backwards from the percentages gives roughly 57 million dollars for cloud and roughly 478 million for RPO, but those are my calculations and not filed figures, which is why they are not in the table.
The cash position. At 30 June 2026 the company held 824.5 million dollars in cash, cash equivalents and short-term investments (96.7 million in cash and 727.8 million in short-term investments), against 704.4 million dollars at 31 December 2025. The June 2025 comparative simply is not in this filing, so there is no valid year-over-year comparison to make.
Growth accelerated, and the guidance already asks it to slow
This is the real story of the report, and it is not in the headline.
First, the acceleration is genuine. In the first quarter of 2026 revenue grew 25.79 percent (154.0 million dollars against 122.4 million), and in the second quarter it grew 28.73 percent. A software company at this scale that speeds its growth rate up, instead of gently decelerating the way almost all of them do, has done something right. I derived the first-quarter figure myself from the six-month column in the release, so the source is clear.
Then the guidance arrives, and it asks for the opposite. Third-quarter revenue guidance is 164 to 166 million dollars. The midpoint, 165 million, is only about 0.75 percent above the 163.8 million just reported. In other words the acceleration is not guided to continue. The quarterly step is close to flat.
The sharper part is profitability. Third-quarter non-GAAP operating income is guided at 27 to 29 million dollars. The second quarter came in at 32.6 million. Management is guiding to a lower operating profit than the one it just reported, at the peak of the growth acceleration.
The full-year arithmetic makes it plainer. Full-year non-GAAP operating income guidance is 116 to 120 million dollars. The company already booked 65.5 million in the first half. Subtract, and the second half is only 50.5 to 54.5 million dollars, materially below the first half. Same on EPS: full-year non-GAAP diluted EPS guidance of 0.96 to 1.00 dollars against 0.54 already delivered in the first half implies 0.42 to 0.46 in the second half. That arithmetic is mine, but every input to it is filed in the release.
In short, this report is nicely wrapped at the front and carries margin pressure in the back. The release does not explain what creates that pressure, and it is the first thing I would be looking for on the call.
What is driving the growth: cloud, security and large customers
Cloud is the engine. Cloud revenue reached 87.5 million dollars, up 53 percent, and now stands at 53 percent of total revenue against 45 percent a year ago. That is a real mix shift: the company is moving from selling software the customer runs to selling a service it runs itself, with revenue accruing on consumption.
The large-customer numbers are the most impressive part of the report. Customers paying more than one million dollars a year went from 61 to 97, up 59 percent. That is far faster than revenue growth itself, and it says the company is going deeper inside big organizations rather than merely adding new logos. The 100-thousand-dollar tier grew from 1,076 to 1,291, up 20 percent.
Net dollar retention improved to 121 percent from 118 percent a year ago. The company attributes it to demand for its security solutions and to cloud consumption. Enterprise+ also rose to 59 percent of revenue from 55 percent.
The forward-looking number is RPO. Remaining performance obligations, meaning money customers have already contractually committed but which has not yet been recognized as revenue, stood at 659 million dollars and grew 38 percent. That outpaces the 29 percent revenue growth, and it is the strongest positive datapoint in the release: the backlog is filling faster than it is being drawn down.
Shlomi Ben Haim, CEO and Co-founder of JFrog, said in the release: "Q2 reflected the strength of our strategy and the disciplined execution of the JFrog team". On the industry backdrop he added: "AI is fundamentally reshaping how software is created, accelerating the volume of software artifacts flowing into production". That is an interesting claim from a company that charges precisely for managing that volume, but the release itself contains no figure quantifying what AI contributed to revenue.
A profit that leans on what was excluded from it
This is where we need to be precise. The 32.6 million dollars of non-GAAP operating income is what you get after adding back roughly 45.8 million dollars of expenses: 39.6 million of share-based compensation, 4.7 million of intangible amortization and 1.6 million of acquisition costs. On a full GAAP basis, the one that forces the company to include everything, the quarter ended with an operating loss of 13.2 million dollars and a net loss of 4.2 million.
Note the relative size: share-based compensation alone, at 39.6 million dollars, is larger than the entire 32.6 million dollars of non-GAAP operating income. The adjusted profit is smaller than the expense excluded to create it.
And in fairness, there is a real improvement here. Share-based compensation intensity fell from 29.9 percent of revenue to 24.2 percent. That is a meaningful move in the right direction. But it does not make the company GAAP profitable, and that crossover has not happened yet.
One more thing worth knowing about the bottom line. The net loss narrowed by 17.5 million dollars, but not all of that came from the business. Operations contributed 12.7 million, interest and other income 1.4 million, and tax 3.4 million: from a tax expense of 2.0 million a year ago to a tax benefit of 1.4 million this quarter. Roughly a fifth of the improvement in the bottom line came from tax, not from operations. A tax benefit in a loss-making quarter is not a repeatable source of earnings.
The dilution continues, and the guidance assumes more of it. GAAP weighted-average shares rose from 115,250 thousand to 121,816 thousand, about 5.7 percent in a year. On a non-GAAP diluted basis they rose from 120,578 thousand to 127,677 thousand, about 5.89 percent. This is why non-GAAP EPS grew 50 percent while non-GAAP net income grew 55.7 percent: the pie grew faster than the slice. Third-quarter guidance assumes roughly 130 million diluted shares, so management expects the dilution to keep going.
Cash flow quality: how much of this is timing
The cash flow looks excellent, and it deserves a look under the hood. Operating cash flow was 57.1 million dollars against 36.1 million a year ago, and free cash flow 53.7 million against 35.5 million. Free cash flow, on the company's own definition, is operating cash flow less capital expenditure, which was 3.3 million dollars this quarter against 0.6 million a year ago.
But most of the improvement is timing, not profit. Deferred revenue, meaning cash collected from customers that has not yet been recognized as revenue, built by 44.3 million dollars in the quarter against only 6.3 million in the year-ago quarter. That gap alone, about 38.0 million dollars, is larger than the entire 21.0 million dollar year-over-year increase in operating cash flow. Essentially all of the first half's 43.8 million dollar deferred revenue build landed in the second quarter.
And in the other direction: accounts receivable was a 22.3 million dollar drag on cash this quarter, against a 1.2 million dollar source last year, while accrued expenses contributed 18.0 million against 10.7 million. All of this is billings and collection timing layered on top of profit, and it is hard to expect a repeat at this magnitude.
The deferred revenue balance stood at 385.8 million dollars at 30 June 2026 (351.5 million current and 34.2 million noncurrent), against 342.0 million at the end of 2025.
Where the spending grew
This is where the operating leverage lives, and where one line stands out. Total GAAP operating expenses rose to 140.9 million dollars from 123.0 million, up just 14.5 percent against 29 percent revenue growth. That is exactly the leverage investors look for in a software company: revenue running faster than costs.
| Line item | Q2 2026 | Q2 2025 | Change |
|---|---|---|---|
| Research and development | 54.0 million dollars | 47.4 million dollars | Up 13.9 percent |
| Sales and marketing | 61.1 million dollars | 55.4 million dollars | Up 10.3 percent |
| General and administrative | 25.7 million dollars | 20.1 million dollars | Up 27.6 percent |
| Total operating expenses | 140.9 million dollars | 123.0 million dollars | Up 14.5 percent |
The outlier: general and administrative. It is the fastest-growing expense line at 27.6 percent, close to the revenue growth rate itself, while R&D grew 13.9 percent and sales and marketing 10.3 percent. Inside that line, share-based compensation jumped 43.9 percent, from 6.3 million dollars to 9.0 million, while share-based compensation in every other line item actually fell year over year. The release offers no explanation.
A first share repurchase that appears only in the statements, never in the text. Financing activities show a new line, repurchases of ordinary shares of 2,028 thousand dollars, and the balance sheet shows treasury shares of the same amount. Every prior comparative period shows zero, so this is the first time. The release text, the CEO quote and the highlights bullets make no mention of a buyback or an authorized program at all. At 2.0 million dollars against 824.5 million of cash it is immaterial, and it comes nowhere near offsetting roughly 5.7 percent of annual dilution. But readers should know the line exists, and that the company did not bother explaining it.
The outlook
Guidance was given in the release under the heading covering the third quarter and fiscal year 2026.
| Metric | Q3 2026 | Full year 2026 |
|---|---|---|
| Revenue | 164 to 166 million dollars | 648 to 652 million dollars |
| Non-GAAP operating income | 27 to 29 million dollars | 116 to 120 million dollars |
| Non-GAAP diluted EPS | 0.22 to 0.24 dollars | 0.96 to 1.00 dollars |
| Assumed diluted share count | About 130 million | About 129 million |
Two notes. First, the company gave no GAAP guidance at all, stating that a reconciliation of the non-GAAP guidance to the corresponding GAAP measures is not available on a forward-looking basis without unreasonable effort. That is standard industry language, and in practice it means there is no forecast for the accounting loss. Second, as set out above, this guidance implies a second half that is weaker than the first half on both operating income and earnings per share.
הזווית שלי
דעה אישית של אילן אברמוב - לא ייעוץ ולא המלצה
This is a good quarter, and I am not trying to take that away from it. Growth accelerating from 25.8 percent to 28.7 percent, cloud running at 53 percent, a jump from 61 to 97 customers paying over a million dollars a year, and a contracted backlog up 38 percent. These are not the numbers of a company in trouble. For me the single most important figure in the whole report is RPO, because a backlog growing faster than revenue means today's selling is tomorrow's revenue.
Even so, what caught my eye was the guidance. When a company reports 32.6 million dollars of adjusted operating income and immediately guides to 27 to 29 million for the next quarter, then sets a full-year target that requires a second half weaker than the first, that is not a technical detail. It says management knows something about the cost structure of the back half that it did not spell out in the release. I do not know whether it is mix shift toward cloud that costs more to run, hiring, AI investment or something else, and I would rather say I do not know than guess.
What bothers me most is the gap between the adjusted profit and the accounting reality. Adjusted operating income of 32.6 million dollars that rests on excluding 39.6 million of share-based compensation is not the same thing as profit. I give real credit for the drop in that intensity from 29.9 percent to 24.2 percent of revenue, which looks like a sustained improvement rather than a trick, but as long as the share count keeps rising by about 5.7 percent a year, the shareholder pays for that profit in dilution.
On the cash flow, I read it carefully. Free cash flow of 53.7 million dollars is a nice number, but when a 44.3 million dollar deferred revenue build accounts for the bulk of the improvement, against only 6.3 million from the same line a year ago, that is billing timing rather than a doubling of cash generation. The tax benefit that supplied roughly a fifth of the narrowing loss is not something that repeats either.
What I am watching next quarter. If the company meets the lower operating profit it guided or beats it, and if third-quarter growth lands above the midpoint, I will read this guidance as management conservatism. If it lands exactly in the range, I will take it that the margin pressure is real. Three numbers are on my desk: the cloud growth rate, the count of customers above a million dollars, and that general and administrative line that jumped 27.6 percent with no explanation. And I would like to know whether the tiny buyback that showed up this week is the start of something or a one-off line. Nothing here is a recommendation.
