Key takeaway
Split Cisco's quarter into two lines. AI infrastructure revenue went from about $1 billion in FY2025 to about $4 billion in FY2026 to a $9 billion FY2027 target, while product gross margin fell from 67.2% to 63.7%. The growth and the dilution are two sides of one coin; expense leverage only bought the operating margin back temporarily.
Why this needs to be split apart
Cisco's first-quarter fiscal 2027 report looks self-contradictory at the headline level: revenue up 24%, the fastest in more than fifteen years; a record 37.8% non-GAAP operating margin; and a non-GAAP total gross margin that fell from 68.1% to 65.6%, with product gross margin dropping from 67.2% to 63.7%.
In one quarter, a growth record, a margin record and gross margin compression all happened at once. Read any single number alone and you reach the opposite conclusion from reading another.
This comparison puts AI infrastructure-driven revenue next to AI hardware-driven margin dilution to test whether they are the same phenomenon seen from two sides.
The table: growth side versus dilution side
| Dimension | Q1 FY2026 | Q1 FY2027 | Change | Basis |
|---|---|---|---|---|
| Revenue | $14.883 billion | $18.517 billion | +24% | GAAP |
| Product revenue | $11.077 billion | $14.667 billion | +32% | GAAP |
| Services revenue | $3.806 billion | $3.850 billion | +1% | GAAP |
| Networking revenue | About $7.78 billion | $11.132 billion | +43% | GAAP |
| Non-GAAP total gross margin | 68.1% | 65.6% | -2.5pt | Non-GAAP |
| Non-GAAP product gross margin | 67.2% | 63.7% | -3.5pt | Non-GAAP |
| Non-GAAP services gross margin | 70.7% | 72.9% | +2.2pt | Non-GAAP |
| Non-GAAP operating expenses | $5.017 billion | $5.140 billion | +2% | Non-GAAP |
| Non-GAAP operating expense ratio | 33.7% | 27.8% | -5.9pt | Non-GAAP |
| Non-GAAP operating margin | 34.4% | 37.8% | +3.4pt | Non-GAAP |
| AI infrastructure revenue, hyperscalers | — | FY2027 target $9 billion | Prior target $7.5 billion | Company guidance |
| AI infrastructure revenue base | About $1 billion in FY2025 | About $4 billion in FY2026 | — | Company disclosure |
Difference one: growth came almost entirely from products, and specifically from networking
Of the 24% revenue growth, products grew 32% and services 1%. Within products, networking rose 43% to $11.132 billion and was the only line growing quickly; security rose 7%, collaboration 8% and observability 2%, all below the company average.
The growth was therefore not a broad recovery but a single product line pulling the whole company. That line is where AI infrastructure orders land.
Difference two: margin compression was concentrated in products while services improved
Product gross margin fell 3.5 points on a non-GAAP basis; services gross margin rose 2.2 points. On the same statement, the two lines moved in opposite directions.
The explanation is straightforward: AI infrastructure involves hardware-heavy system-level delivery with substantial silicon photonics, switching and memory content, and unit margins below traditional enterprise networking gear. Services rely on renewals from an installed base, where scale is flat but mix improves, so margin rises.
Total gross margin got dragged down by products. This matches the explanation the company itself offered: rising component costs, particularly memory, and a higher hardware share of revenue.
Difference three: expense leverage covered the margin loss and then some
This is the most easily missed layer. Non-GAAP operating expenses grew just 2%, and their share of revenue fell from 33.7% to 27.8%, a 5.9-point drop. Gross margin fell 2.5 points, the expense ratio fell 5.9 points, and the net effect was a 3.4-point gain in operating margin to 37.8%.
So "gross margin compression" and "record operating margin" are not contradictory: the first is a product-mix problem, the second is a scale-against-fixed-cost problem. The problem is that the expense ratio has a floor. Once it stops falling, gross margin dilution flows straight through to operating margin.
Difference four: the three steps of the AI target carry different margin costs
The disclosed AI infrastructure revenue path has three steps: about $1 billion in FY2025, about $4 billion in FY2026, and a $9 billion target for FY2027, up from $7.5 billion.
The steeper the curve, the higher the hardware share of total revenue and the greater the pressure on gross margin. Raising the FY2027 target from $7.5 billion to $9 billion — a 20% increase — also means more gross margin dilution than originally planned.
Management's Q2 guidance implies non-GAAP gross margin of 65%-66%, below this quarter's 65.6%, meaning the company itself expects a bit more compression.
Difference five: orders and backlog are the validation point
Total product orders grew 21% year over year, with double-digit growth in every region. RPO was $46.169 billion, up 8%, with product RPO up 7% and services RPO up 9%; deferred revenue was $30.058 billion, up 7%.
These three figures point to the same thing: orders are growing, but backlog growth of 8% trails revenue growth of 24%. That gap means part of this quarter's revenue came from delivering existing contracts rather than entirely from new orders. If revenue growth stays above 20% for several quarters while RPO growth stays in the single digits, the backlog gets consumed and growth needs new orders to replace it.
Conclusion in three sentences
First, $9 billion is not the problem; the gross margin attached to $9 billion is. Raising the revenue target is good news on the demand side, but the margin dilution per $10 billion of AI infrastructure revenue will determine the quality of this growth cycle.
Second, expense leverage is a buffer, not a cure. A 37.8% operating margin is impressive, but it rests on a 5.9-point decline in the expense ratio. That ratio cannot fall indefinitely, so dilution eventually reaches the operating line.
Third, the validation point is the Q2 gross margin guide. If the 65%-66% range lands at or below the low end, dilution is accelerating; if it returns above 66%, the product mix is improving. That number says more about the economics of this AI buildout than revenue does.
What cannot be directly compared
- GAAP and non-GAAP figures cannot be mixed: this quarter's GAAP total gross margin was 63.7% against 65.6% non-GAAP, in the same three months;
- The "about $7.78 billion" networking figure for Q1 FY2026 is back-calculated from the 43% growth rate and was not directly disclosed;
- The $9 billion AI infrastructure figure is full-year guidance, not recognized revenue; the roughly $1 billion and roughly $4 billion figures are prior company disclosures on an approximate basis;
- RPO spans multiple years and cannot be divided by a single quarter's revenue;
- The company did not split how much of the margin decline came from memory costs versus hardware mix;
- This article makes no valuation judgment and does not forecast the share price.
Sources
- Cisco earnings release (November 12, 2026): https://newsroom.cisco.com/c/r/newsroom/en/us/a/y2026/m11/cisco-reports-first-quarter-earnings.html
- Cisco Q4 FY2026 earnings presentation (AI revenue base and the prior $7.5 billion FY2027 target): https://s21.q4cdn.com/812015656/files/doc_earnings/2026/q4/presentation/Q4-26-Earnings-Slides.pdf
- Related on this site: Cisco Q1 FY2027 financial data, Cisco Q1 FY2027 earnings FAQ
