Arista Networks (ANET) Q4 FY25 earnings call: guidance and Q&A

What are the factors that would determine if Arista gains 1 or 2 additional 10% customers in 2026?

Jayshree Ullal explained that variables such as deferred revenue acceptance criteria and timing, ongoing demand, and the ability to ship given supply chain constraints will influence whether these customers reach the 10% threshold. She expressed confidence in the demand but acknowledged the variables.

“Some of it may be sitting in deferred. So there's an acceptance criteria that we have to meet. And there's also timing associated with meeting the acceptance criteria. Some of it is demand that is still underway.”

asked by Meta Marshall · answered by Jayshree Ullal

Given Q1 FY26 guidance of 30% growth, why is the full-year FY26 guidance only 25%?

Jayshree Ullal clarified that the 25% full-year guidance reflects what Arista can realistically *ship* given new products, early stages of AI adoption by customers, and the law of large numbers on a growing base, rather than a lack of demand. She emphasized providing a 'high dose of reality' early in the year.

“I'm giving you the greatest visibility I can fairly early in the year on the reality of what we can ship, not what the demand might be. It might be a multiyear demand that ships over multiple years.”

asked by Samik Chatterjee · answered by Jayshree Ullal

Can management quantify the revenue and gross margin impact from memory dynamics and constraints, and what could be the upside in a perfect world?

Jayshree Ullal stated that memory prices have worsened significantly in 2026, becoming 'horrendous' and 'exponentially higher,' leading to potential onetime price increases on selected SKUs. Chantelle Breithaupt confirmed the 62-64% gross margin guide is comfortable despite these pressures, indicating they have a path to mitigate the impact.

“in 2026, the situation has worsened significantly. We're having to smile and take it just about at any price we can get and the prices are horrendous. They're an order of magnitude exponentially higher.”

asked by David Vogt · answered by Jayshree Ullal

How much of the $3.25 billion AI guide factors in scale-up networking, and what is the growth outlook for non-AI and non-campus segments?

Jayshree Ullal explained that significant scale-up production is expected in 2027, tied to the ESUN specification and 1.6T technology, with trials in Q4 this year. She noted that when AI and campus segments grow very fast, other segments may grow less, without providing specific growth rates for non-AI/non-campus business.

“the real production level will be in 2027, primarily centered around not just 800 gig, but 1.6T.”

asked by Aaron Rakers · answered by Jayshree Ullal

How does Arista participate in multibillion-dollar model builder clusters (e.g., Anthropic) and leverage these relationships to scale with larger cloud companies?

Jayshree Ullal confirmed that network infrastructure is critical for model builders, who are increasingly diverse (e.g., Gemini, xAI, Anthropic, OpenAI). She noted that these models are distributed across multiple locations and cloud titans, creating opportunities for Arista to work with both AI specialty providers and cloud titans to integrate cloud and AI.

“I think the biggest issue is not only the model builders, but they're no more in silos in one data center, and you're going to see them across multiple colos and multiple locations and multiple partnerships with our cloud titan customers that we've historically not worked with this.”

asked by Amit Daryanani · answered by Jayshree Ullal

How will product deferred revenue come off the balance sheet, and how much is included in the full-year revenue guidance?

Chantelle Breithaupt reiterated that deferred revenue is driven by new products, customers, and use cases like AI, with acceptance criteria varying from 6 to 18 months. She stated that Arista does not guide deferred revenue due to its volatility and the difficulty in predicting when deployments will be accepted and released, making it hard to quantify its contribution to full-year guidance.

“What goes into deferred is new product, new customers, new use cases. The great new use case is AI. The acceptance criteria for that for the larger deployments is 12 to 18 months. Some can be as short as 6 months.”

asked by George Notter · answered by Chantelle Breithaupt

What is the materiality of Neocloud momentum, and is AMD's new chips a catalyst for Arista's networking business?

Jayshree Ullal described Neoclouds as a mix of financially strong and more speculative players, with AI driving this segment. She noted that while some Neoclouds are plentiful in quantity, they are not yet significant in dollar numbers. Regarding AMD, she stated that 20-25% of current deployments use AMD as the preferred accelerator, where Arista is favored for its open standards approach over vertical stacks.

“Today, when we look at our deployments, we see about 20%, maybe a little more, 20% to 25% where AMD is becoming the preferred accelerator of choice.”

asked by Benjamin Reitzes · answered by Jayshree Ullal

Can management elaborate on Arista's telemetry capabilities and how they provide competitive differentiation, particularly for use cases like Agentic AI?

Jayshree Ullal highlighted Arista's real-time streaming telemetry, which is central to EOS and CloudVision. Ken Duda, President and CTO, added that the EOS architecture's state orientation captures network state and streams it to CloudVision. For AI, this extends to in-network data sources (flow control, RDMA, buffering) and host-level information, providing a unified view that aids customers in building and debugging complex AI solutions.

“we pull those -- that information all together in CloudVision and give the operator a unified view of what's happening in the network and what's happening in the host.”

asked by Ryan Koontz · answered by Kenneth Duda