Arista Networks (ANET) investor relations material

Supply chain management and revenue outlook

• Demonstrated strong supply chain management, outperforming peers during COVID and current industry-wide constraints, especially for advanced semiconductor components and memory, with demand outpacing supply.

• Maintains confidence in delivering a 20%+ CAGR revenue target over three years, despite ongoing supply challenges and long lead times for key components.

• Gross margin guidance remains stable at 62%-64% for the year, with only one price increase implemented so far, aiming to balance customer partnership and value.

• Aggressively securing supply to meet demand, especially for memory, with some lag in passing cost increases to customers, but maintaining long-term relationships over short-term margin gains.

• Deferred revenue reflects strong shipment growth, clarifying that it represents invoiced and billed shipments, not just order growth.

Technology and product innovation

• Rapid evolution in customer network architectures, with accelerated cycles for switches and accelerators, now as short as 18 months, and increasing adoption of high-radix Ethernet-based designs.

• Ethernet is increasingly validated for AI scale-out architectures, with projections that AI will comprise 30% of business by 2026, up from near zero a few years ago.

• High-radix switches and modular chassis provide competitive differentiation, with co-engineering partnerships enabling advanced, pathfinding products not yet broadly available.

• Scale across solutions are driving significant AI revenue growth, expected to reach $1 billion, as customers adopt distributed and modular data center architectures to address power and space constraints.

• Deep engagement with frontier AI labs and neoclouds mirrors hyperscaler partnerships, focusing on co-design and engineering extension to support rapid innovation.

Market trends and customer engagement

• Customers are rapidly iterating architectures, with some deploying multiple competing designs in parallel and shifting quickly based on operational needs.

• Agentic AI is anticipated to drive future enterprise data center and campus modernization, with current activity focused on vertical-specific use cases and forward-looking planning.

• Network observability and automation, through platforms like EOS and CloudVision, are positioned as key differentiators for managing complex, less visible AI-driven environments.

• Rack-scale networking and integrated solutions are becoming more important as inference workloads and edge deployments grow, leveraging existing high-density networking expertise.

• Token economics in AI are highly sensitive to network performance, making low-latency, low-jitter solutions critical for cost efficiency.