A company that beats its own revenue guidance by about $360 million and grows 58.5% year over year should not be trading about 20% below its peak. That tension is exactly why Arista Networks (ANET) is worth your attention this morning.
On November 3, Arista reported non-GAAP EPS of $1.20, beating estimates by $0.12, with revenue surging 58.5% year over year to about $3.66 billion and exceeding its prior revenue guidance of $3.3 billion by roughly $360 million. That 58.5% year-over-year revenue growth is the highest since the company’s IPO. And yet, the stock had plunged 8.34% on October 30 to $168.86 amid a broader selloff in technology and AI-related stocks. Shares recovered roughly 3.5% on the November 3 session, but the discount from the October 9 high remains meaningful.
The Business Behind the Number
Arista sells the networking switches that connect GPU clusters inside hyperscale data centers. When Microsoft or Meta builds a new AI training facility, they need Arista’s high-speed Ethernet fabric to move data between thousands of accelerators at once. CEO Jayshree Ullal has emphasized Arista’s strategic focus on AI investments as hyperscalers transition toward open Ethernet standards, supported by the Etherlink family of switches and the Extensible Operating System (EOS).
When Arista began its AI journey, management spoke of four or five AI customers with approximately 50,000 to 100,000 Nvidia GPUs. In 2026, it has already surpassed 100 Etherlink customers. That broadening of the customer base is real. But the revenue is still highly concentrated. Management expects at least three, and potentially four, customers to account for 10% or more of revenue, subject partly to product availability and shipment timing.
What Changed This Quarter
Management raised full-year 2026 revenue guidance for the fourth time this year, to $13.3 billion. Then it went further. Arista provided an initial 2027 revenue outlook of approximately $18.5 billion, implying 38% to 40% growth, with AI fabrics expected to double to about $7 billion. The company posted an adjusted operating margin of 49.1% and an adjusted gross margin of 64.0%, both slightly exceeding analyst estimates, and projects Q4 revenue between $3.9 billion and $4.0 billion.
The 2027 AI fabrics target is worth pausing on. CFO Chantelle Breithaupt confirmed that the 2027 guide is based on the $13.3 billion 2026 outlook, within which AI is about $3.5 billion and campus about $1.25 billion. Doubling AI fabrics revenue from $3.5 billion to $7 billion in a single year is not a conservative assumption, but Arista has now beaten its own guidance by double digits for multiple consecutive quarters, so the credibility is there.
The Bull and Bear Case
Bulls point to a simple observation: the orders are still being placed. Arista is increasing inventory and supplier commitments to prepare for customer demand and secure components for AI deployments and new products. Scale-up AI networking is not expected to contribute materially until 2028, with trials beginning in 2027, meaning the product cycle that would juice revenue further hasn’t even started.
Bears focus on the concentration. A limited number of customers account for a substantial portion of revenue, creating a structural risk if capital spending from major cloud hyperscalers decelerates. Arista’s two largest customers represented approximately 26% and 16% of 2025 revenue, respectively. Any pause in Microsoft’s or Meta’s data center build plans lands directly on Arista’s income statement.
What to Watch Next
Three things matter from here. First, whether Q4 revenue lands within or above the $3.9 to $4.0 billion guided range. Second, whether any of the four large customers signals a slowdown in capex, that would move the stock more than any earnings beat. Third, early 2027 order momentum: management expects scale-across to represent at least one-third of AI revenue in 2027, up from roughly 30% in 2026, and that mix shift will tell investors whether the AI buildout is maturing or still accelerating.
At roughly 20% off its high and with AI fabrics targeting a double in one year, the market is asking Arista to prove the concentration risk is worth carrying. Monday’s quarter suggests the answer is still yes.
