AI Forecast Tracker
← Back to scoreboard
SoftwareANET

Arista Networks

Q1 FY2026 (reported May 5): revenue $2.71B (+35.1% YoY, +8.9% QoQ), beating estimates. Management raised full-year 2026 guidance to ~$11.5B (+27.7% YoY) and lifted the AI-fabric revenue target to $3.5B, anchored by Meta and Microsoft — Microsoft alone was ~26% of FY2025 revenue with a second hyperscaler customer (~16%) pushing two-customer concentration to ~42%. The stock still fell ~12-14% after hours: management flagged wafer, memory, and optics supply constraints that will pressure gross margins for 1-2 years even as demand outpaces supply. The bigger overhang is competitive — IDC data shows NVIDIA's Spectrum-X platform overtook Arista for the #1 spot in datacenter Ethernet switching (~21.5% share, up from under 4% two years ago), winning Meta and Google as new Ethernet-gear customers and directly contesting the "open Ethernet beats proprietary lock-in" pitch Arista built its business on. Forward P/E ~44-45x on a ~$210-235B market cap leaves little room for a guidance miss. EOS/CloudVision software plus Broadcom Tomahawk5/Jericho3-AI merchant silicon still underpin real differentiation, and Arista remains a Gartner Leader in enterprise wired/wireless LAN — but the moat looks less durable than it did a year ago.

AI Impact Score
6.7/10
↑ Positive
Scoring Breakdown
Sector Base
9
AI Revenue Exposure
8
Moat Durability
6
Disruption Risk (lower=better)
6
AI Adoption Maturity
8

Scenarios

Bull Case

Ethernet is winning the broader AI-fabric war — hyperscalers still prefer multi-vendor optionality over InfiniBand lock-in, and the $3.5B AI revenue guide could prove conservative if demand keeps outpacing supply. Enterprise/campus LAN leadership (Gartner Magic Quadrant Leader) diversifies revenue beyond the hyperscaler AI capex cycle.

Bear Case

NVIDIA's vertical GPU-plus-networking stack is structurally threatening: Spectrum-X already overtook Arista in datacenter Ethernet switching and captured Meta and Google as customers — using the same open-Ethernet argument Arista built its pitch on. Two-customer concentration (~42% of revenue) plus supply-driven margin pressure through 2027 leave a premium ~44x forward P/E multiple exposed to any stumble.

Key Factors to Watch

  • ●Q1 FY2026 revenue $2.71B (+35.1% YoY); FY26 guidance raised to ~$11.5B (+27.7%), AI-fabric revenue guide raised to $3.5B
  • ●Stock fell ~12-14% after the Q1 print despite the beat-and-raise — wafer/memory/optics supply constraints flagged to pressure gross margins through 2026-2027
  • ●Customer concentration — Microsoft ~26% and a second hyperscaler (~16%) of FY2025 revenue — ~42% combined from two customers
  • ●NVIDIA Spectrum-X overtook Arista for #1 in datacenter Ethernet switching (IDC: ~21.5% share, up from <4% two years ago), winning Meta and Google as new Ethernet customers
  • ●EOS/CloudVision single-software-image plus Broadcom Tomahawk5/Jericho3-AI merchant silicon partnership; Gartner Leader in enterprise wired/wireless LAN diversifies beyond hyperscaler AI

Score History

DateScoreDirectionNote
2026-09-056.7PositiveW36 refresh — score holds 6.65. Q2 2026 (Aug 4) revenue $3.036B (+37.7% YoY), a first $3B+ quarter, non-GAAP EPS $1.02 (+39.7%), FY26 guidance raised a third time to about $12.6B (+40%). Held deliberately: Arista still does not report AI-fabric revenue as an actual line item, only a forward target held unchanged at 'at least $3.5B', so the guidance raise is a bundled-demand figure and not an AI breakout. Gross margin fell to 63.4% from 65.6% on wafer and memory cost pressure — the supply-constraint risk already encoded at disruption_risk 6, now materializing as forecast. Deferred revenue backlog grew about $643M to $3.63B; Etherlink AI-fabric customers crossed 100 cumulative, a customer count rather than a revenue disclosure.
2026-07-196.7PositiveBaseline assessment triggered by Bud Finance inbox signal (2026-07-17) on Q1 beat-and-raise, wafer/memory/optics margin-pressure warning, and NVIDIA Spectrum-X competitive threat.

Software Peers

Last researched: 2026-09-05

This is research and analysis, not financial advice. Scores reflect AI impact potential, not investment recommendations.