ANET · Q3 2025 Earnings

Bullish

Arista Networks

Reported November 4, 2025

30-second summary

Arista delivered $2.31B in Q3, +27.5% YoY and +4.7% QoQ, with non-GAAP gross margin of 65.2% and non-GAAP operating margin of 48.6% — both well above the Q3 guide. Management raised FY25 revenue to $8.81–8.93B (26–27% growth, from 25%), reframed AI center revenue from "ahead of $1.5B" to "at least $1.5B," and put a stake in the ground for FY26 at $10.65B (~20% growth) with AI center at $2.75B. The bullish signal is the FY26 disaggregation; the only meaningful caution is Q4 gross margin guided to 62–63%, ~300bps below Q3 actual on cloud/AI mix.

Headline numbers

EPS

Q3 FY2025

$0.75

Revenue

Q3 FY2025

$2.31B

+27.5% YoY

Gross margin

Q3 FY2025

64.6%

Operating margin

Q3 FY2025

42.4%

Key financials

Q3 FY2025

Metric	Q3 FY2025	YoY	Q2 FY2025	QoQ
Revenue	$2.31B	+27.5%	$2.21B	+4.7%
EPS	$0.75	—	$0.73	+2.7%
Gross margin	64.6%	—	65.2%	-60bps
Operating margin	42.4%	—	44.7%	-230bps

Guidance

Arista raised FY2025 revenue guidance to $8.81-8.93B (26-27% growth) and beat Q3 on revenue, gross margin, and operating margin; Q4 guidance indicates sequential revenue increase with slightly lower margins.

Guidance is issued for both next quarter and the full year. Both may appear below.

Actuals vs prior guidance

Metric	Period	Prior guide	Actual	Δ	Result
Revenue	Q3 FY2025	approximately $2.25 billion	$2.308 billion	+$58 million above guide	Beat
Non-GAAP Gross Margin	Q3 FY2025	approximately 64%	65.2%	+120 bps above guide	Beat
Non-GAAP Operating Margin	Q3 FY2025	approximately 47%	48.6%	+160 bps above guide	Beat

New guidance

Metric	Period	Guide	YoY
Revenue	Q4 FY2025	$2.3-$2.4 billion	—
Non-GAAP Gross Margin	Q4 FY2025	62-63%	—
Non-GAAP Operating Margin	Q4 FY2025	47-48%	—

Changes to prior guidance

Metric	Period	Prior guide	New guide	Δ	Result
Gross Margin	FY2025	approximately 63% to 64%	approximately 64%	+50 to +100 bps at high end vs prior midpoint	Raised
Revenue	FY2025	$8.75 billion	$8.81-$8.93 billion	+$60-180 million above prior point guidance	Raised
AI Center Revenue	FY2025	ahead of $1.5 billion	at least $1.5 billion	guidance clarified to minimum floor vs prior 'ahead of' language	Raised

Reaffirmed unchanged this quarter: Operating Margin (approximately 48%), Campus Segment Revenue ($750-$800 million)

Management tone

Q2 anchor: "Once in a lifetime opportunity" / 17%→25% raise → Q3 anchor: FY26 disaggregation and organizational restructuring around AI.

From AI-as-one-of-several-growth-drivers to AI-as-the-organizing-principle. Last quarter management broadened the AI narrative across NVIDIA, AMD, and XPUs to defuse concentration concerns. This quarter the company restructured around AI explicitly — Tyson Lamoureux now leads a dedicated cloud and AI mission reporting to Ken Duda and Hugh, with Duda promoted and recognized as leading "our top AI and cloud segment of customers." Lamoureux's own remark from the call — "I've got to say it's a lot more than I anticipated before I got here" — is the kind of hand-tipping the company rarely allows on a scripted call. The signal is that AI is no longer treated as upside optionality.

From a single 2026 target to fully disaggregated segment accountability. At Q2, management pulled the $10B revenue target forward from 2028 to 2026 but did not break it down. This quarter Arista went further: "We committed to 2.75 billion out of our new target of 10.65 billion in revenue, representing 20% revenue growth in 2026." Bumping the 2026 target from $10B to $10.65B and naming AI center ($2.75B) and campus ($1.25B) as distinct line items is a meaningful step toward holding management accountable to specific segment trajectories rather than a blended growth number.

From "ahead of $1.5B" hedge to "at least $1.5B" floor. The AI center language tightened from a vague over-shoot phrasing to a firm minimum. This is a deliberate verbal commitment — and combined with three of four tier-one customers now past the 100K GPU mark, suggests management has line-of-sight that removes the optionality previously embedded in the wording.

From supply chain as headwind to supply chain as the binding constraint. Last quarter inventory and deferred revenue volatility were framed as ongoing nuisances. This quarter, in response to direct questions from BofA and JPM on the sequential growth deceleration (Q2 +10% → Q3 +5% → Q4 implied +0–4%), management was explicit: "We will continue to have some variability in future quarters as a reflection of the combination of demand for our new products and the lead times from our key suppliers." Lead times of 38–52 weeks on memory and merchant silicon were cited. The reframe matters because it positions decelerating sequential growth as a shippable-vs-demanded problem, not a customer-demand problem.

From TAM-expansion-narrative to $100B+ TAM. Q2's TAM framing was generic. This quarter Tom Sanjay Kapadia explicitly anchored the company's strategic addressable market at "over 100 billion in forthcoming years" — context for why aggressive 2026 segment targets are defensible and why organizational investment is accelerating.

Recurring themes management leaned on this quarter:

AI network architecture dominance and system-level platform consolidationOrganizational restructuring with dedicated cloud/AI leadership hierarchyEthernet speed scaling (800G to 1.6T) as enabler of multi-petabit AI networksDeferred revenue growth ($4.7B, +$625M Q-o-Q) signaling strong forward booking momentumEcosystem standardization (ESUN, UEC) positioning Arista as open yet proprietary leaderTAM expansion narrative anchored to terawatt-scale AI infrastructure buildouts

Risks management surfaced:

Tariff and trade restrictions (mentioned in forward-looking statement boilerplate and Q4 guidance at 62-63% gross margin 'inclusive of possible known tariff scenarios')Supply chain lead time variability from key suppliers affecting inventory managementGross margin pressure from potential tariff scenarios in FY26 (62-64% range vs. FY25 ~64%)Deferred revenue volatility independent of underlying business drivers due to customer-specific acceptance clausesCompetition from white box commoditization and customer preference for lower-cost alternatives in 'good enough' use cases

Q&A highlights

Tal Liani · Bank of America

Asked about sequential growth deceleration (10% Q2 to 1.6% Q4) and underlying drivers, seeking clarification on demand outlook for next year.

Management emphasized no demand concerns; attributed deceleration to supply constraints rather than demand weakness. Reiterated 20% growth commitment for next year despite revenue base increase from $8.75B to $8.87B. Stated shipments follow supply availability, not demand.

Sequential growth deceleration: 10% (Q2) → 5% (Q3) → 1.6% (Q4)2026 growth commitment: 20% (unchanged)Revenue base increase: $8.75B to $8.87BSupply-driven constraints identified as primary cause of variability

Aaron Rakers · Wells Fargo

Questioned gross margin guidance of 62-63% for current quarter, implying product margins below 60%, and asked about tariff and mix impacts with forward guidance implications.

Management explained product margin compression is driven by cloud and AI titan mix shift, which historically occurs when AI/cloud volume increases. Stated margins are managed well and within historical ranges. No structural change to business model. Manufacturing team discipline managing mix effectively.

Gross margin guidance: 62-63%Product margin: significantly below 60% (due to cloud/AI mix)Services/software margin: stated to be lower than analyst assumptionsHistorical precedent: margin pressure expected with heavy AI/cloud mix

Atif Malik · Citi

Asked about decision-making ownership on networking hardware for OpenAI and Anthropic partnerships, and whether Arista's market share within cloud titan customers remains stable.

Management reported market share is stable with no significant changes up or down. Described decision-making as joint responsibility: LLM providers (OpenAI, Anthropic) guide design specs, while cloud titans handle procurement of data centers/power/cooling. Market characterized as rising tide benefiting all participants.

Arista share within cloud titans: stable, no significant changesCompetitive environment: coexist with NVIDIA bundling and white box alternativesMarket dynamic: massive and growing, 'rising tide rises all boats'Decision-making structure: joint responsibility between LLM providers and cloud titans

Samik Chatterjee · JP Morgan

Probed variability in shipments and whether tier-one customer 100K+ cluster progression plans have changed, driving Q4 guidance reduction. Asked about supply-driven constraints.

Management confirmed variability is largely supply-driven, not demand-driven. Confirmed three of four tier-one customers have crossed 100K GPU mark; fourth approaching. Long lead times on components (38-52 weeks) cited as constraint. Purchasing commitments increasing, which management said reflects underlying demand.

Large GPU cluster status: 3 of 4 customers have crossed 100K markComponent lead times: 38-52 weeks (memory, chips, merchant silicon)Demand assessment: greater than ability to shipPurchase commitments: increasing (reflects underlying demand)

David Vogt · UBS

Questioned early 2026 guidance raise and implied margin for core business outside AI, asking whether core enterprise/campus business will remain flat or decline given AI growth assumptions.

Management stated 2026 outlook includes optionality and doesn't assume 100% achievement of AI/campus goals. Core business growth status uncertain (flat, single-digit, or higher) - cannot yet predict mix. Emphasis that AI and campus will grow substantially in double digits, with existing large base showing lesser but unspecified growth.

2026 guidance philosophy: includes optionality, not 100% assumption of AI/campus targetsCore business growth outlook: unknown (flat, single-digit, or above single-digit)AI and campus: double-digit growth expectedMix prediction for 2026: cannot yet determine

Answers to last quarter's watch list

Q3 revenue beat vs. $2.25B guide wider than Q2 beat its guide? — Q3 came in at $2.308B, a $58M / +2.6% beat. Q2 beat its guide by a similar order of magnitude. This was a beat-and-raise, not a blowout, and FY25 was lifted by 100–200bps of growth to 26–27%.

Resolved positively

Backend AI revenue tracking the $750M sub-target inside the $1.5B AI center FY25 number? — Management did not provide a specific backend AI revenue dollar figure this quarter, but firmed the AI center language from "ahead of $1.5B" to "at least $1.5B" and disclosed three of four tier-one customers have crossed 100K GPUs. The underlying trajectory is clearly intact; the specific $750M backend disclosure framework wasn't repeated.

Continue monitoring

Enterprise/neo-cloud customer count moving above 30 by year-end? — Management did not provide an updated cohort customer count on this print. The diversification narrative was instead carried by the FY26 disaggregation and the organizational restructuring.

Continue monitoring

Gross margin path to the 63–64% FY guide; size of the H2 give-back from Q2's 65.6%? — Q3 printed 65.2% (only a modest give-back), but Q4 is guided to 62–63%, which is the real step-down. FY blended is now targeted at ~64% (up from prior 63–64% range). The give-back is back-end-loaded into Q4 on cloud/AI mix plus tariffs. Status: Resolved negatively for Q4 specifically, neutral for the full year.

Analyst Day disclosure on $10B 2026 target composition? — Partially answered on this print rather than at Analyst Day: 2026 revenue target raised to $10.65B, AI center $2.75B, campus $1.25B explicitly disclosed. Core business not quantified (per UBS Q&A).

Resolved positively

Deferred revenue trajectory and any duration disclosure? — Deferred revenue grew to $4.7B, +$625M QoQ, with management explicitly attributing the increase to higher evaluation inventory reflecting new product uptake. Purchase commitments + inventory totaled $7.0B vs $5.7B at end of Q2. No duration buckets disclosed.

Continue monitoring

What to watch into next quarter

Whether Q4 non-GAAP gross margin lands at the high end of the 62–63% guide or below it — a print at or below 62% would suggest the mix headwind is structural into FY26, not a one-quarter cloud/AI titan concentration.

Whether FY26 revenue guidance of ~$10.65B holds or is raised at Analyst Day / Q4 print — Arista has a recent pattern of raising mid-year by hundreds of bps; an early refresh of the $10.65B number would signal the AI center $2.75B target is itself conservative.

Concrete dollar disclosure on FY25 AI center revenue achievement vs. the "at least $1.5B" floor — needed to validate whether the firmed language reflects modest upside or a meaningful overshoot.

Purchase commitments + inventory direction from the $7.0B Q3 print — further growth implies supply is still the binding constraint; flattening would suggest demand catching up to capacity, with implications for Q1/Q2 FY26 shipments.

Whether the fourth tier-one cloud customer crosses 100K GPUs and any disclosure on a fifth named AI customer — would directly speak to the durability of the FY26 AI center target.

Core (non-AI, non-campus) business trajectory disclosure — currently a $6.65B black box inside the $10.65B FY26 target; UBS's question wasn't answered quantitatively and remains the largest modeling gap.

Sources

Arista Networks Q3 2025 Earnings Release, filed with SEC: https://www.sec.gov/Archives/edgar/data/1596532/000159653225000284/ex991q325-earningsrelease.htm
Arista Networks Q3 2025 earnings call (management prepared remarks and Q&A).
Arista Networks Q2 2025 brief (Tapebrief), for cross-quarter comparison.

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.