Arista's Q3 2025 Earnings Call: Contradictions in Supply Chain, Growth, and Margins Emerge

Generated by AI AgentEarnings DecryptReviewed byAInvest News Editorial Team
Wednesday, Nov 5, 2025 6:48 am ET4min read
Aime RobotAime Summary

-

reported $2.3B Q3 revenue (27.5% YOY), driven by AI/cloud demand with 18.7% AI-related revenue.

- Gross margin dipped to 65.2% (vs 65.6% prior quarter) due to product mix shifts and tariff impacts, though above guidance.

- Management attributed shipment delays to supply chain constraints (38-52 week lead times), not weakening demand, reaffirming 20%+ growth targets.

- FY2026 guidance shows $10.65B revenue target (20% growth) with AI center investment rising to $2.75B, while campus expansion and blue-box adoption remain key focus areas.

Date of Call: November 4, 2025

Financials Results

  • Revenue: $2.3B, up 27.5% YOY, (transcript notes 'above our guidance of $2.5B')
  • EPS: $0.75 diluted EPS, up 25% YOY (diluted shares 1.277B)
  • Gross Margin: 65.2%, above guidance of 64%, down from 65.6% sequentially, up from 64.6% YOY
  • Operating Margin: 48.6% of revenue (operating income $1.12B)

Guidance:

  • Q4 FY2025: Revenue $2.3B–$2.4B; gross margin 62%–63% (inclusive of known tariffs); operating margin ~47%–48%; effective tax rate ~21.5%; ~1.281B diluted shares.
  • FY2025: Full-year revenue growth ~26%–27% (~$8.87B midpoint); campus ~$750M–$800M; AI center at least $1.5B; gross margin ~64%; operating margin ~48%.
  • FY2026: Target revenue growth ~20% to $10.65B; campus target $1.25B; AI center target $2.75B; gross margin ~62%–64%; operating margin ~43%–45%.

Business Commentary:

  • Revenue Growth and AI Integration:
  • Arista Networks reported revenue of $2.3 billion for Q3 2025, up 27.5% year-over-year.
  • Growth was driven by strong demand in AI and cloud technologies, especially networking for AI, with AI-related revenues accounting for approximately 18.7% of total revenue.

  • Gross Margin and Product Mix:

  • The non-GAAP gross margin for Q3 was 65.2%, slightly below the previous quarter's 65.6%.
  • The decrease was partially attributed to the mix of products, with cloud and AI titans driving volume and higher enterprise customer contributions.

  • Enterprise and Campus Expansion:

  • The enterprise segment, notably driven by the VeloCloud acquisition, contributed significantly to growth.
  • The company is focusing on expanding its presence in the campus segment, leading to increased investment in this area.

  • Leadership and Organizational Changes:

  • Arista announced the promotion of Ken Duda to President and CTO, reflecting his critical role in AI and cloud segments.
  • The appointment of Todd Nightingale as President and Chief Operating Officer was aimed at driving growth in the campus and enterprise markets.

  • Product Deferred Revenue and AI Demand:

  • Product deferred revenue increased by approximately $625 million compared to the previous quarter.
  • This increase is influenced by new customer-specific acceptance clauses and AI-related projects, indicating a ramping up of new products and use cases.

    Sentiment Analysis:

    Overall Tone: Positive

    • Management framed the quarter as the "19th consecutive record quarter," highlighted "momentum continues," and stated confidence "for the remainder of FY'25 and through FY'26." They reiterated targets (e.g., AI center $1.5B for 2025, $2.75B for 2026) and emphasized strong demand despite supply-driven shipment variability.

Q&A:

  • Question from Tal Liani (BofA Securities): There is quarter-to-quarter deceleration versus prior year; what are the underlying drivers and should we be concerned about growth going forward?
    Response: Demand remains strong; observed deceleration reflects shipment/supply constraints and quarter-to-quarter variability, not weakened demand—company remains committed to its 20% growth target.

  • Question from Aaron Rakers (Wells Fargo Securities): Can you unpack the Q4 gross margin guidance drivers—tariffs versus other dynamics and implication for forward expectations?
    Response: Q4 guidance (62%–63%) is driven by mix shifting toward cloud/AI (lower product margins) and known tariff scenarios; this is a mix/timing dynamic, not a structural margin change, and is within historical guidance ranges.

  • Question from Michael Ng (Goldman Sachs): How is Arista positioned for full-rack solutions and addressing compute-network convergence?
    Response: Arista is engaged in multiple early full-rack designs via partnerships (JDM/blue-box and reference designs), handling cabling/co‑packaging and expects revenue to come from both network sales and integrated solutions as UEC/ESUN standards mature (2026–2027).

  • Question from Atif Malik (Citigroup): Who drives networking-hardware decisions among LLM providers, cloud titans and does Arista's share remain stable?
    Response: Share is stable; specifications and design are a joint process—LLM/software players guide specifications while cloud titans manage procurement and deployment decisions.

  • Question from Samik Chatterjee (JPMorgan): Is shipment variability and the Q4 guide driven by supply; any change in Tier‑1 customers' cluster plans?
    Response: Variability is largely supply-driven—component lead times (typically 38–52 weeks) limit shipments despite strong customer build plans; company is increasing purchase commitments accordingly.

  • Question from Amit Daryanani (Evercore ISI): What's driving the implied slowdown from high‑20s growth toward low‑20s in FY26—why would growth decelerate?
    Response: Management rejects 'deceleration' label—timing/acceptance clauses and variability in when large AI deployments land create pacing differences; underlying demand remains strong but timing is uncertain.

  • Question from David Vogt (UBS): With AI and campus driving targets, what should we expect for the core business outside those areas in 2026?
    Response: AI and campus are expected to drive most growth; the core business may be flattish or single‑digit growth, and guidance conservatively assumes not all targets will fully materialize immediately.

  • Question from Benjamin Reitzes (Melius Research): Confidence in participating in builds affiliated with cloud titans and private model providers?
    Response: Confident—Arista is a key participant today in scale‑out and scale‑across builds and expects to participate more meaningfully in scale‑up use cases over time (more material toward 2027).

  • Question from Timothy Long (Barclays): How should we think about blue‑box economics and where blue‑box deployments will be most successful?
    Response: Blue‑box often evolves to hybrid deployments that incorporate EOS; economics are broadly similar to current cloud/AI business, with scale‑up volumes potentially improving margins—Arista will manage mix to protect overall margins.

  • Question from Meta Marshall (Morgan Stanley): How are front‑end network upgrades progressing versus expectations as inference use cases ramp?
    Response: Ethernet is increasingly replacing InfiniBand; back‑end upgrades (400G→800G→1.6T) are pressuring front‑end upgrades and driving convergence—Arista is being selected for both front and back.

  • Question from Karl Ackerman (BNP Paribas): How to think about market opportunity for disaggregated scheduled fabrics (DSF) versus nonscheduled fabrics?
    Response: Arista supports both DSF and nonscheduled fabrics at scale, offers customer choice, and runs the same software across architectures—neither approach is favored universally.

  • Question from Simon Leopold (Raymond James): What blue‑box assumptions are baked into 2026 guidance and expected breadth of adoption?
    Response: Blue‑box adoption in 2026 is expected to be limited to a small set of advanced customers (~10–20); the lower‑margin mix from blue‑box scenarios is already factored into the 2026 guidance.

  • Question from James Fish (Piper Sandler): Over 3–5 years what mix could blue‑box represent and does front/back convergence erode Arista's advantage?
    Response: Convergence is advantageous to Arista—ability to address both front and back is a differentiator; blue‑box may be high in units but lower in dollars, and overall it should be incremental and beneficial.

  • Question from Antoine Chkaiban (New Street Research): Progress and milestones for the Ultra Ethernet Consortium (UEC)?
    Response: UEC v1.0 was published (June 2025); Arista is fully UEC‑capable and will continue adding compliance/features; UEC/ESUN standards are expected to broaden adoption and enable scale‑up Ethernet deployments.

  • Question from George Notter (Wolfe Research): What are you seeing with neo‑cloud customers and momentum there?
    Response: Neo‑cloud traction is accelerating—many providers now invite Arista for product and network design support, especially those with colo/power assets, and are becoming new customers.

  • Question from Sebastien Cyrus Naji (William Blair): Where are you investing in enterprise GTM and what should drive progress into 2026?
    Response: Investing in campus (VeloCloud integration), geographic expansion (notably Asia), channel development and new‑logo acquisition to capture enterprise white space and cross‑sell into the large TAM.

  • Question from John Jeffrey Hopson (Needham): How to think about opportunity from new AI network builds versus refreshes/upgrades?
    Response: Primary opportunity is new AI data‑center build‑outs (gigawatt scale); refreshes exist but new builds are the main demand driver today.

  • Question from Benjamin Bollin (Cleveland Research): How have engineering/delivery lead times evolved and confidence in meeting cloud customers' needs over next 12–18 months?
    Response: Management relies on early customer previews and has increased purchase commitments; component lead times remain long but procurement ramp and efforts to accelerate campus lead times support delivery confidence.

Contradiction Point 1

Demand and Supply Chain Constraints

It involves the company's ability to meet demand due to supply chain constraints, which directly impacts revenue and customer expectations.

What caused the slowdown in sequential growth, and is future growth at risk? - Tal Liani(BofA Securities)

2025Q3: There is no concern about demand, as shipments and revenue follow supply constraints. Variability in demand is greater than Arista's ability to ship, driven by long lead times for components like memory and merchant silicon. - Jayshree Ullal(CEO)

What is driving the upside in cloud results: front-end upgrades or back-end demand? - Meta A. Marshall(Morgan Stanley)

2025Q2: We came into Q2 with a new product introduction, which included Campus and ZTNA. And as our product sets expand, we've now added 20% more to our already rich product portfolio. And these 20% new products are designed for multi-use case applications. - Jayshree Ullal(CEO)

Contradiction Point 2

Growth Rate Expectations

It involves the company's growth rate expectations, which are crucial for investor expectations and strategic planning.

What factors are causing the growth slowdown, and why isn't acceleration expected to continue? - Amit Daryanani(Evercore ISI)

2025Q3: The revenue growth deceleration is due to demand greater than supply capabilities. AI use cases have acceptance clauses that are not seasonally predictable, leading to some lumpiness in quarterly performance. - Jayshree Ullal(CEO)

What is driving the increase from 17% to 25% annual growth rate? - Amit Jawaharlalaz Daryanani(Evercore ISI Institutional Equities)

2025Q2: It is clear that the growth of 17% to 25% limits the gross margin in the short run, and we will look for continued improvement in the long run. The demand for AI, Campus, and cloud are all growing, and they are growing fast. - Jayshree Ullal(CEO)

Contradiction Point 3

AI Backend Sales Target and Performance

It involves expectations and actual performance related to AI backend sales, which are crucial for understanding Arista's strategic focus and growth opportunities.

What caused the slowdown in sequential growth rates, and should we worry about future growth? - Tal Liani (BofA Securities)

2025Q3: Due to supply constraints, we shipped and recognized $720 million in the quarter. This is down from our previous guidance of $750 million. - Jayshree Ullal(CEO)

How does the EOS software layer's value in the network backend compare to white box solutions? - Samik Chatterjee (JPMorgan)

2024Q4: We met our revenue target. We achieved $1.02 billion in Q4, up 47% year-over-year. - Chantelle Breithaupt(CFO)

Contradiction Point 4

Gross Margin Trends

It reflects changing expectations in financial performance, specifically gross margins, which are important for assessing profitability and market positioning.

What are the key gross margin drivers in this quarter's guidance, especially the product margin expected below 60%? - Aaron Rakers (Wells Fargo Securities)

2025Q3: Gross margins for Q3 were 72.4%, and we expect Q4 to be below 60%. - Jayshree Ullal(CEO)

Will Q4 Blackwell revenue be additive, and what is the expected gross margin exit rate? - Stacy Rasgon (Bernstein Research)

2024Q4: Gross margins for Q3 are expected around 75%, with full-year guidance in the mid-70s. - Chantelle Breithaupt(CFO)

Contradiction Point 5

AI Growth and Demand

It involves the company's perspective on AI growth and demand, which is crucial for understanding Arista's strategic focus and market positioning.

Who is driving network hardware decisions for large language model providers, and what is Arista's share in these announcements? - Atif Malik (Citi)

2025Q3: Arista's share is stable, with no significant change from bundling strategies with NVIDIA and white boxes. - Jayshree Ullal(CEO)

Can you explain the value of the EOS software layer to the network backend and compare it to white box solutions? - Samik Chatterjee (JPMorgan)

2024Q4: Upside comes from deployment of 3 customers with 100,000 GPUs each and anticipated momentum in AI, cloud, and enterprises. - Jayshree Ullal(CEO)

Comments



Add a public comment...
No comments

No comments yet