Summary

  • JFrog's cloud revenue reached US$87.5 million, up 53% and equal to 53% of total revenue, versus 45% a year earlier.
  • GAAP gross margin rose from 76.3% to 77.9%, even as third-party hosting costs increased US$4.2 million. That proves blended operating leverage, not a cloud-only margin.
  • The statutory cost line combines SaaS with self-managed subscriptions, so investors cannot assign the US$36.2 million cost balance to either deployment model.

A majority business without a separate cost ledger

JFrog reported US$163.8 million of Q2 revenue. Its US$87.5 million cloud operating metric was 53% of that total, crossing the halfway line after representing 45% a year earlier. The remaining US$76.3 million is an arithmetic residual, not a disclosed self-managed segment: the operating metric and accounting lines need not share every classification rule.

The statutory statement uses different boundaries. It combines self-managed and SaaS subscription revenue in one US$155.5 million line, then reports US$8.2 million of self-managed licence revenue separately. Cost of revenue combines self-managed subscriptions and SaaS in a single US$36.2 million line. There is no cloud-only cost, gross profit or capital intensity measure.

That missing split matters because delivery changes who carries infrastructure. In SaaS, JFrog hosts and manages more of the service. In self-managed deployment, the customer's environment carries more of the runtime surface while JFrog supplies software and support. The two models can have different hosting, support, security and renewal economics even when they use the same platform.

The disclosed blended result is constructive. GAAP gross profit rose to US$127.6 million and gross margin to 77.9% from 76.3%. JFrog says cost of revenue increased US$6.0 million, including US$4.2 million of additional third-party hosting costs driven mainly by SaaS growth and US$1.4 million of personnel expense. Revenue grew fast enough to produce operating leverage across the pooled ledger.

It would be wrong to reverse-engineer a cloud margin by assigning all hosting growth to cloud and all other costs to self-managed products. Hosting supports shared architecture; personnel and acquired-intangible amortisation also sit in the combined line. The filing establishes the direction of one cost driver, not a complete allocation rule.

Product breadth overlaps deployment mode

Enterprise+ represented 59% of total revenue, up from 55%. That share cannot be added to the 53% cloud share. Enterprise+ describes subscription breadth; cloud describes delivery. One customer can occupy both sets. Treating them as separate revenue pools would double count.

Contract indicators show depth without solving the margin question. Net dollar retention was 121%, based on an average of four same-customer quarterly comparisons. Customers above US$1 million ARR rose from 61 to 97. About US$30.3 million of the US$36.6 million year-on-year revenue increase came from existing customers, leaving roughly US$6.3 million attributable to new customers on the disclosed arithmetic.

RPO reached US$659.0 million: US$385.8 million billed and US$273.2 million unbilled, with 67% expected within twelve months. Yet RPO excludes usage fees above minimum SaaS commitments. A consumption-led cloud expansion can therefore appear in revenue and retention before it is fully visible in the contract balance.

Cash is the strongest cross-check. Q2 operating cash flow was US$57.1 million and free cash flow US$53.7 million. The company still recorded a US$13.2 million GAAP operating loss versus US$32.6 million of non-GAAP operating income; Q2 share-based compensation was US$39.6 million. Cloud scale is advancing, but the public record supports a blended economics claim, not a standalone cloud-profit claim.

Sources