Summary

  • Fortinet's second-quarter revenue rose 26% to US$2.0479 billion. Product revenue increased 52% to US$773.0 million and supplied 63.2% of the total revenue increase.
  • Product's share of revenue rose by a reported 6.5 percentage points. Because product gross margin was 69.8% and service gross margin was 86.6%, the mix change pulled total gross margin down 0.5 point even though product margin improved.
  • Operating expenses increased 11%, far below revenue growth. GAAP operating margin therefore expanded from 28.1% to 33.7%; this was leverage over a larger base, not an absolute cost cut.
  • Management calls product growth a leading indicator for service growth. That claim needs attachment, renewal and deferred-revenue evidence because 90% of Q2 service revenue was already deferred at the preceding quarter-end.

Fortinet did not have a weak margin quarter. It had two different margin movements in the same quarter.

Product gross margin rose from 67.4% to 69.8%. Total gross margin fell from 80.7% to 80.2%. Both statements are true because the company sold much more of the lower-margin line.

The distinction matters beyond accounting presentation. A security appliance can produce revenue when control transfers. The related subscription and support obligations are recognised over time. A quarter led by products therefore brings forward one part of the commercial relationship while leaving another part to be proved through attachment, delivery and renewal.

Product supplied nearly two-thirds of the growth

The Form 10-Q reports total revenue of US$2.0479 billion, up US$417.9 million from a year earlier. Product revenue rose US$264.1 million to US$773.0 million. Service revenue rose US$153.8 million to US$1.2749 billion.

Product therefore supplied 63.2% of the total revenue increase. Its share of revenue moved from about 31.2% to 37.7%. Fortinet reports the shift as 6.5 percentage points from service to product.

The 52% product growth rate is not a pure volume measure. Fortinet attributes it mainly to higher hardware unit shipments, but also to higher average selling prices, demand for higher-performing models and recent pricing actions. Secure-networking hardware, term licences, AI-infrastructure deployments, upgrades, upselling and new uses all appear in the explanation.

The filing does not publish a bridge among those causes. It would be wrong to label the full increase “AI revenue”, “price” or “units”. Product revenue also includes software licences, so “hardware revenue rose 52%” would overstate what the reported segment line proves.

The narrow conclusion is stronger. Customers bought enough additional product, at a mix of volumes and prices, to make product the main source of quarterly growth. That changed the economics of the consolidated margin before it proved anything about the later service life of the installed base.

Better product margin still lowered the blended margin

Product gross profit rose US$196.2 million to US$539.2 million. Its 69.8% gross margin was 2.4 points above the prior year. Fortinet credits recent pricing actions, partly offset by a greater hardware mix and higher memory-chip costs.

Service gross profit rose US$131.9 million to US$1.1040 billion. Its gross margin slipped only 0.1 point to 86.6%, with data-centre and cloud-service expansion adding cost while revenue outpaced labour expense.

The service line still earned nearly 17 points more gross margin than product. Moving more revenue into product therefore lowered the consolidated rate even as both gross-profit pools grew. Total gross profit increased US$328.1 million, or about 25%, while the total gross margin declined 0.5 point.

This is a mix effect, not evidence that the new product dollar became unprofitable. Indeed, product margin improved. Nor is it harmless. If product remains a larger share without creating later service revenue, the blended margin can stay below a service-heavier structure.

The future cost surface also differs by line. Fortinet says full-year product margin may face higher hardware-component costs. It expects a slight service-margin decline as it expands data centres, colocation and cloud hosting for Unified SASE and SecOps.

Those investments can be productive. More cloud capacity may support more recurring service revenue, just as more products can enlarge the installed base. The return must appear after the cost, through recognised revenue and gross profit rather than through product labels alone.

Operating leverage absorbed the mix drag

The gross-margin decline did not reach the operating-margin line. Operating expenses increased US$96.8 million, or 11%, to US$953.9 million. Revenue grew more than twice as fast.

Research and development expense rose 7% to US$225.0 million. Sales and marketing rose 13% to US$669.1 million. General and administrative expense rose 7% to US$61.1 million. Fortinet increased spending in every major function; it did not produce the margin by cutting their absolute dollars.

Those expenses occupied less of a much larger revenue base. GAAP operating income rose US$231.3 million to US$689.3 million, and operating margin expanded 5.6 points to 33.7%.

Fortinet provides the bridge. Sales and marketing supplied 3.6 percentage points of operating leverage, R&D supplied 1.9 points and administration supplied 0.6 point. The 0.5-point gross-margin decline offset part of those gains, leaving the reported 5.6-point improvement.

The arithmetic also reconciles in dollars. The US$328.1 million increase in gross profit, less US$96.8 million of additional operating expense, equals the US$231.3 million increase in operating income.

This is stronger evidence than a margin created by a shrinking expense base. Yet it remains sensitive to the revenue cycle. If product growth slows before services accelerate, operating expenses may no longer fall as quickly as a share of revenue.

The service ledger runs on an older clock

Fortinet's recognised service revenue does not mainly arise from products shipped in the same quarter. The filing says 90% of service revenue in both compared second quarters was already included in deferred revenue at the preceding 31 March.

That figure is not a renewal rate or an attachment rate. It identifies the timing origin of recognised revenue. Most of the quarter's service line came from obligations sold and invoiced earlier, then recognised as FortiGuard subscriptions, other security services and FortiCare support were delivered.

This delay is central to management's outlook. In the prepared remarks, Fortinet says first-quarter 2026 was the trough in service-revenue growth and calls faster product growth a key leading indicator for the second half.

The mechanism is plausible. A new or upgraded appliance can carry security subscriptions, technical support and later expansion. A higher-performance device used around an AI data centre can enlarge the amount of protected traffic and the operational value of attached services.

But a leading indicator is not a conversion receipt. The checked materials do not disclose product-to-service attach rates, renewal rates, the average delay from shipment to service recognition, or expansion by product cohort.

Security-subscription revenue grew 13% and technical support and other services grew 14%. Those are positive results, but they are well below the 52% product rate and mostly reflect an earlier contract book. The current product cohort still has to enter that book.

Billings improve visibility without closing the test

Billings rose 33% to US$2.3721 billion. The reconciliation consists of US$2.0479 billion of GAAP revenue plus a US$324.2 million increase in deferred revenue. A year earlier, deferred revenue added US$149.2 million to the calculation.

Total deferred revenue reached US$7.6757 billion, about 17% above the prior year and 8% above December. The majority represents unrecognised subscription and support revenue. That balance provides a substantial delivery schedule and helps explain the cash result.

Yet Fortinet itself names the limitations. Billings change with contract term and do not determine when service revenue will be recognised. Another company may also calculate a similar measure differently. A longer agreement can lift billings without proving stronger consumption or a better renewal.

Operating cash flow reached US$1.0436 billion and company-defined free cash flow reached US$965.6 million. Management attributes the improvement to higher billings, better quarter linearity and working-capital discipline.

Cash is important counterevidence to a fragile-growth thesis. It is not the same receipt as service conversion. Collections and deferred revenue can produce cash before all obligations are delivered, while gross margin records the cost of that delivery later.

The installed base is now the acceptance surface

Fortinet describes demand across Secure Networking, Unified SASE and AI-driven Security Operations. Its investor presentation assigns 66%, 24% and 10% of Q2 billings to those pillars and reports growth of 34%, 35% and 25% respectively.

Those are useful portfolio signals, not GAAP segments. FortiSASE billings grew more than 100%, according to management, but no absolute amount is provided. An eight-figure AI-data-centre win shows commercial relevance without identifying recognised revenue, margin or repeatability.

The market question is therefore not whether buyers wanted products in Q2. They did. It is whether that demand enlarges a durable service relationship rather than ending at the shipment.

A constructive path would join four movements. Service revenue growth would accelerate after its stated trough. Deferred revenue would grow without relying only on longer terms. Service gross-profit dollars would outpace the added cloud-delivery cost. Product cohorts would renew and expand into SASE, support and security operations.

The adverse path is subtler than a sales collapse. Product revenue could remain respectable while component costs rise, service attachment disappoints and the blended margin stays product-heavy. Billings could still look strong for a time if contract duration lengthens.

Fortinet has earned the right to make the conversion claim. Product margin improved, operating leverage was broad and cash generation was strong. It has not yet published the cohort evidence that would make conversion observable from outside.

Sources