Why Arista Networks Is A Buy Despite Rising Competition

Arista Networks delivered record Q2 revenue of +37.7% YoY and outperformed expectations, while raising 2026 revenue growth guidance from 28% to 40% YoY on accelerating momentum. Gross margin fell 220 bps due to customer mix, but operating margin still expanded 110 bps YoY, reflecting operating leverage and expense discipline. The analyst reiteration as a “buy” aligns with strong top-line trajectory and upgraded growth outlook.

Analysis

ANET’s strength matters less as a one-quarter beat and more as evidence that AI networking spend is still moving from pilot to repeatable rollout. If hyperscalers are pulling forward switch purchases, the second-order winners are the component vendors with the tightest attach rates to high-speed data-center builds: Broadcom for custom networking silicon, Credo and Marvell for connectivity, and optical names like Coherent/Lumentum if the buildout remains bandwidth-heavy. The losers are legacy networking incumbents whose enterprise-heavy mix cannot match that growth vector, with Cisco most exposed on relative growth and multiple compression rather than absolute revenue loss.

The margin mix hit is the key tell. It implies the current demand mix is probably more cloud/customer-concentration heavy and less benign than the headline growth suggests, so the market should not extrapolate peak gross margin. That said, operating leverage holding up despite mix pressure signals pricing power and operating discipline; if this persists for 2-3 quarters, consensus 2026 numbers likely still have room to rise, which supports a premium multiple rather than just EPS growth.

Contrarian risk: the market may be overconfident that this is a clean, broad-based cycle, when it may actually be a handful of hyperscaler programs that can decelerate quickly if capex budgets get re-phased. The falsifier is any sign that guide credibility weakens next quarter: flat-to-down order growth, another gross margin step-down, or commentary that the 2026 growth path is front-loaded. Time horizon is mostly 1-3 months for estimate revisions, with 6-18 months for whether ANET sustains share gains versus Cisco and other Ethernet incumbents.

Key Decisions for Investors

• Buy ANET on any post-earnings consolidation; prefer adding only if the stock resets 5-8% below the peak reaction, since the medium-term catalyst is estimate revisions rather than immediate re-rating.
• Pair trade: long ANET / short CSCO for 1-3 months to express share gain in AI/data-center switching versus slower legacy enterprise exposure; thesis breaks if CSCO shows accelerating campus/data-center bookings or ANET gross margin keeps sliding.
• Basket expression on the AI networking supply chain: overweight ANET, BRCM, and CRDO into the next quarterly print cycle; highest upside if hyperscaler capex commentary stays constructive.
• Use a watch alert on ANET gross margin and 2026 guidance cadence; if gross margin falls another ~150-200 bps while revenue still accelerates, trim because the market may start treating growth as lower-quality.