Arista Networks at Rosenblatt AI summit: AI networking broadens

Arista Networks, at Rosenblatt’s 6th Annual Technology Summit: The Age of AI (Part II) on 18 August 2026, presented a broadening AI networking story with strong demand across customer groups and a fast-growing Etherlink portfolio. Management discussed near-term supply constraints, tougher competition, and an early scale-up market, while keeping margin targets intact.

Arista did not report quarterly earnings during the call but provided a detailed update on growth and margins. Management emphasized that the margin outlook reflects mix rather than pricing power, with a 63% gross profit margin and a 31% return on equity in the last twelve months. With a market capitalization of $242.55 billion, Arista is scaling rapidly at 33% revenue growth over the past year.

Chantelle Breithaupt, Chief Financial Officer, Arista Networks, highlighted that Arista was "almost born for this AI moment" with the Etherlink portfolio in June 2024, connecting to the cloud and AI. Breithaupt explained that the broader customer mix demonstrates the company's ability to serve various AI use cases, moving from large customers to a diverse customer base.

Brendan Gibbs, Vice President of Product Line Management, added that the company is in the scale-up phase, with AI networking still early but with significant future market potential.

Arista outlined three networking layers: scale-out for large backend training clusters, scale-across for data center and AI deployments, and scale-up for an emerging market. Gibbs stated that AI customers require a stable software base, as they are in a race to innovate and expand in the cloud and GPU as a service. Arista's open MSA (Multi-Scale Architecture) has more than 100 companies signed on, with EOS (Ethernet Operating System) as a central part of the company's value proposition.

Management discussed supply constraints and the steady but important growth of enterprise and campus networking. Breithaupt emphasized that the company is focusing on prioritizing and not distracted by other things, maintaining a strong position across AI networking use cases. The company sees AI networking as still early, with near-term focus on scale-out and scale-across, while scale-up could become a larger business later in the decade.