ANET · Q2 2025 Earnings
BullishArista Networks
Reported August 5, 2025
30-second summary
Arista delivered $2.21B in Q2 revenue, +30.4% YoY and +10% QoQ, with non-GAAP operating margin of 48.8% and non-GAAP operating income crossing $1B for the first time. Management raised FY25 revenue guidance to $8.75B (25% growth, up from the 17% guide that opened the year) and pulled forward the $10B revenue target to 2026 from 2028. The print recasts Arista from a hyperscaler-concentrated AI beneficiary into a multi-segment platform story with enterprise campus, neo-cloud, and multi-accelerator AI all contributing.
Headline numbers
EPS
Q2 FY2025
$0.73
Revenue
Q2 FY2025
$2.21B
+30.4% YoY
Gross margin
Q2 FY2025
65.2%
Operating margin
Q2 FY2025
44.7%
Key financials
Q2 FY2025| Metric | Q2 FY2025 | YoY |
|---|---|---|
| Revenue | $2.21B | +30.4% |
| EPS | $0.73 | — |
| Gross margin | 65.2% | — |
| Operating margin | 44.7% | — |
Guidance
Prior quarter data unavailable — comparison not possible.
Product revenue
Q2 FY2025| Segment | Q2 FY2025 | YoY |
|---|---|---|
| Product | $1.877B | +31.9% |
| Service | $0.328B | +22.7% |
Management tone
Management's posture this quarter is significantly more expansive than Arista's typical conservative cadence. The language shifts from incremental progress to platform redefinition, and the willingness to raise FY25 mid-year by eight percentage points of growth — and to pull the $10B target forward by two years — is itself the signal.
From pragmatic 17% start-of-year guide to 25% "transformational" framing. McAdam's framing on the call: "We began the year with a pragmatic guide of 17%... as the year has progressed, we recognize the potential to build a truly transformational networking company... This feels to us like a unique once in a lifetime opportunity." Arista raising mid-year on customer momentum (rather than holding conservative buffer) is a meaningful break from the company's historical pattern of beating-and-raising-modestly.
From single-accelerator AI bet to multi-accelerator standard. The AI commentary explicitly broadened beyond NVIDIA: "Arista continues to be the premier and preferred AI networking platform of choice for all flavors of AI accelerators. While majority today is NVIDIA GPUs, we are entering early pilots connecting with alternate AI accelerators, including startup XPUs, the AMD MI series." Combined with the UltraEthernet Consortium spec release in June, this reframes Arista's AI exposure from "leveraged to NVIDIA's ramp" to "leveraged to Ethernet displacing InfiniBand across the accelerator ecosystem."
From hyperscaler-concentrated AI to enterprise/neo-cloud diversification. "AI is also spreading its wings into the enterprise and Neo Cloud sectors, and we are winning approximately 25 to 30 customers to date." Customer count in this cohort grew from ~15 to 25–30. This directly addresses the concentration risk created by the loss of the fifth major AI customer earlier this year — management is signaling the long tail now offsets the lost top-five slot.
From product-line view to integrated "Arista 2.0" platform. "As we enter the next phase of Arista 2.0, growing from $5.8 billion in 2023 to a forecasted $10 billion revenue in 2026." The narrative now bundles data center, AI, campus, enterprise, and WAN (including VeloCloud) into a single platform thesis rather than separate segment stories.
Recurring themes management leaned on this quarter:
Risks management surfaced:
Q&A highlights
Amit Daryanani · Evercore
What specific factors enabled Arista to raise FY guidance from 17% to 25% growth? What markets or vectors are driving this? Is the higher growth durable given the once-in-a-lifetime AI opportunity?
Management attributed the raise to activity translating into confirmations across all three sectors: enterprise campus (record demand, largest TAM), AI (despite losing fifth customer, targeting $750M backend revenue and exceeding $1.5B for year with strong activity in four major customers plus new enterprise/neo-cloud deployments), and cloud (front-end upgrades driven by back-end AI pressure requiring 100G to 400G to 800G migrations). Long-term growth expected in mid-teens to low double-digits CAGR.
Samik Chatterjee · JPMorgan
How is Arista meeting the $750M AI backend revenue target despite losing the fifth (sovereign) customer? Is this from larger cluster sizes or distributed deployments? What's the status of the previously-slow fourth customer?
Two customers have approached/will reach 100,000 GPUs with focus shifting from sheer GPU count to distributed clusters for training/inference. Third customer medium pace, fourth customer still early (target 100k GPUs for next year). Offsetting fifth customer loss through increased aggregate contribution from 15 to 25-30 enterprise and neo-cloud customers. Emphasis moving beyond GPU counting to locality, distribution, radix, multi-tenancy, and resilience factors.
Michael Ng · Goldman Sachs
What drove the deferred revenue/billings growth? Did deferred revenue contribute materially to guidance raise? What's management's confidence in the $10B revenue target for 2026?
Deferred revenue growth driven by new products, use cases (especially AI), spanning 12-18-24 month durations across product and services. Deferral volatility expected due to size of new use cases and product introductions. Management will not guide deferred revenue separately. Announced $10B revenue goal for 2026 (two years ahead of previously-guided 2028 schedule), to be discussed further at Analyst Day. Long sales cycles and high experimentation with GPUs, traffic patterns, and brownfield integration driving extended deferral periods.
George Nodder · Wolf Research
How does Arista view the competitive environment given investor concerns about Celestica strength and NVIDIA's networking business? What is Arista's differentiation strategy?
Management stated competitive landscape unchanged from historical patterns (Cisco, whitebox competitors, NVIDIA participation already acknowledged). Perceived negative chatter attributed to slower meta share growth vs. prior years, not environmental change. Differentiation highlighted at three levels: (1) platform performance and features innovation at strongest point ever, (2) customer intimacy with appreciation for support/quality/problem-solving approach. No material change in competitive position despite external speculation.
Tal Liani · Bank of America
Is Tomahawk 6 delay correlated with current growth (pull-forward buying)? At 49% operating margin, won't customers switch to cheaper whitebox alternatives? How does Arista justify premium pricing vs. whitebox economics?
On Tomahawk 6: Broadcom is good partner, not late, complex silicon technology, product in labs, new products expected next year—no specific correlation to current growth surge implied. On margins and whitebox: 49% margin reflects operational efficiency (doing more with less, limited layers/hierarchy) not just product value. Customers can trade capex for opex: cheap box requires hundreds of internal engineers to support; Arista model with congestion control, buffering, EOS reduces support burden. Whitebox coexistence expected in leaf-only scenarios; premium justified by value, support, quality, and total cost of ownership for customers appreciating Arista's approach.
What to watch into next quarter
Whether Q3 revenue beats the $2.25B guide by a wider margin than Q2 beat its guide — would signal the 25% FY raise is itself conservative and another raise is coming on the Q3 print.
Backend AI revenue progress toward the $750M FY25 target — needs to track at ~$190M/quarter run rate in H2; any softness here directly challenges the "fifth-customer loss is fully offset" narrative.
Whether enterprise/neo-cloud customer count moves above 30 by year-end — the diversification thesis depends on this cohort continuing to broaden, not just deepen.
Gross margin path to the 63–64% FY guide — Q2 printed 65.6%; the implied H2 step-down to ~62% in some quarters would mean tariff/supply tailwinds are reversing. Watch the size of the give-back.
Any Analyst Day disclosure on the composition of the $10B 2026 target — specifically the AI vs. campus vs. cloud mix, given the target was pulled forward two years.
Deferred revenue trajectory and whether management starts disclosing duration buckets — long-cycle AI deployments are inflating deferred balances; opacity here is a growing investor concern.
Sources
- Arista Networks Q2 2025 Earnings Release, filed with SEC: https://www.sec.gov/Archives/edgar/data/1596532/000159653225000214/ex991q225-earningsrelease.htm
- Arista Networks Q2 2025 earnings call commentary (management prepared remarks and Q&A).
Get the next brief, free.
We publish analyst-grade earnings briefs the same day or morning after every call — headline numbers, segment KPIs, Q&A highlights, and tone analysis. Free during beta.
This is not investment advice.