Arista Networks (ANET) — 2Q26 Earnings Preview

Timing correction: ANET is scheduled to report today, Tuesday, August 4, 2026, after the U.S. market close—not tomorrow. The earnings call is for 2Q26.

Bottom line

ANET enters earnings with a favorable fundamental setup—AI Ethernet demand, cloud spending, and scale-across networking are all accelerating—but the debate has shifted from whether demand is strong to how much of that demand can be supplied, recognized as revenue, and converted into margins.

The company’s own 2Q guide is for approximately $2.80 billion of revenue, $0.88 non-GAAP EPS, and 46–47% non-GAAP operating margin. Street expectations are only modestly higher at roughly $2.828 billion and $0.89 EPS. That means the headline hurdle is manageable; the stock’s reaction is likely to hinge more on the magnitude and credibility of a raise to the full-year outlook, AI revenue commentary, gross-margin trajectory, and evidence that component shortages are not becoming the binding constraint on growth.

My framing is constructive, but with a high qualitative bar: a simple revenue/EPS beat may not be enough if management does not reaffirm or raise its $11.5 billion 2026 revenue target, $3.5 billion AI-fabric goal, and margin framework.


The setup: what ANET has already promised

In 1Q26, Arista delivered:

Metric 1Q26 actual 2Q26 guide
Revenue $2.709B ~$2.80B
Revenue growth 35.1% YoY Street: ~28.3% YoY
Non-GAAP EPS $0.87 ~$0.88
Non-GAAP gross margin 62.4% 62–63%
Non-GAAP operating margin 47.8% 46–47%

Management raised its 2026 outlook following 1Q to:

Importantly, 1Q revenue beat Arista’s own $2.6 billion guide by about 4.2%, showing the company began the year with meaningful momentum. The question for 2Q is whether this outperformance can continue despite increasingly tight supply of advanced silicon, memory, optics, and other components.


What matters most this quarter

1. Full-year revenue guidance—and whether supply caps the upside

The most important question is whether ANET can raise its $11.5 billion revenue target again.

Management said in May that demand was the strongest it had seen during CEO Jayshree Ullal’s tenure, but also stressed that supply was lagging demand. It described constraints as a potential one- to two-year issue, particularly around leading-edge chips, wafers, and memory.

A raise would signal that procurement actions, multiyear supply agreements, and customer deployment schedules are permitting more shipments than previously assumed. Conversely, a maintained guide accompanied by more forceful supply-constraint language would probably be viewed less favorably—even if reported 2Q numbers beat consensus.

Investor read-through: - Bullish: Revenue beats and the FY26 outlook rises above $11.5B. - Neutral: Beat, but FY guide stays unchanged because supply remains limiting. - Bearish: Strong demand commentary but constrained shipments, weaker project timing, or a guide that implies a muted second half.


2. AI revenue: $3.5B target, scale-across mix, and customer diversification

ANET’s AI networking opportunity is evolving from a conventional scale-out Ethernet story into three distinct markets:

  1. Scale-out: connecting large AI clusters.
  2. Scale-across: interconnecting distributed AI clusters and data centers.
  3. Scale-up: rack-level high-speed connectivity, which management sees as more of a 2027–2028 opportunity.

For 2026, scale-out and scale-across are the earnings-relevant pieces. Management said scale-across should account for at least one-third of AI revenue this year. This matters because scale-across leans on Arista’s high-value routing platforms, traffic engineering, software, and operational tooling—potentially supporting differentiated content and market positioning.

Investors should listen for:

The key risk is that AI hardware can ship before installations, cabling, GPU availability, and customer acceptance are complete. That can delay revenue recognition even when end-market demand is strong.


3. Deferred revenue: a positive demand signal, but not a clean near-term catalyst

Arista ended 1Q with approximately $6.2 billion of total deferred revenue, up from about $5.37 billion at year-end. Most of the balance was product-related.

That is encouraging in the sense that it reflects large deployments and committed customer activity. But it should not be treated as straightforward backlog that will automatically convert next quarter. Management has emphasized that AI product qualification and customer acceptance can take six to eight quarters, versus historically shorter cycles.

For the quarter, investors should focus on management’s explanation of:

A large deferred-revenue build is not inherently negative, but it would heighten the focus on conversion timing and execution risk.


4. Gross margin is likely the key swing factor

ANET’s near-term margin setup is mixed:

Management guided 2Q non-GAAP gross margin to 62–63%, below the long-running levels investors have come to associate with the business. It has been clear that large hyperscale customers tend to be lower margin, and that component availability has become more important than maximizing price.

The critical issue is not simply whether ANET lands at 62% or 63% in 2Q. It is whether management can defend the 62–64% full-year range while continuing to grow AI revenue rapidly.

What would be encouraging: - Gross margin near the high end of the 2Q range; - Stable or improving outlook for 2H margins; - Evidence that targeted pricing actions, second sourcing, and supply agreements are mitigating cost inflation.

What would concern investors: - Margin below guide; - Explicitly worsening memory/silicon cost commentary; - Higher expedite costs or supply decommitments; - A full-year gross-margin range that moves lower.


Balance-sheet and supply-chain context

Arista is financially well equipped to secure supply:

The strength of the balance sheet gives ANET purchasing power and flexibility to reserve components. At the same time, high inventory and multiyear commitments create a two-sided risk: they support delivery into demand, but could become a margin and working-capital problem if customer projects slip, demand is revised downward, or architectures change.

The central supply-chain question for earnings is therefore: Are elevated commitments a source of competitive advantage, or evidence that Arista is taking increasing inventory risk to preserve growth?


Stock and expectations

ANET closed at $184.92 on August 3, up about 8.6% from its May 5 close before the prior quarterly report. The shares have therefore recovered meaningfully into earnings, while the consensus revenue estimate sits only about 1% above management’s 2Q guide.

That creates an asymmetrical setup:


Earnings scorecard

Area What investors want to hear What could disappoint
2Q revenue / EPS Clear beat versus ~$2.83B / ~$0.89 Merely in line, particularly if tied to project timing
FY26 revenue Raise above $11.5B or strong confidence in additional upside Guide unchanged because supply is constraining shipments
AI fabrics Reaffirm/raise $3.5B target; sustained scale-out and scale-across momentum AI target unchanged with less convincing shipment outlook
Gross margin Near 63%; FY range intact Margin miss, worsening memory/silicon cost outlook
Supply Better allocation, credible multiyear sourcing, manageable lead times Decommits, longer lead times, inability to meet customer demand
Deferred revenue Healthy conversion and manageable acceptance cycles Larger balance due to project delays or testing bottlenecks
Customers New hyperscaler/AI-titan wins and diversified neocloud traction Greater dependence on a few existing large customers
Campus / enterprise Progress toward $1.25B campus goal and VeloCloud cross-sell AI strength masking weakness in enterprise or campus

Key questions for the call

  1. Can Arista raise the $11.5B FY26 revenue outlook, or is supply now the gating factor?
  2. Is the $3.5B AI-fabric target still conservative, achievable, or constrained by component availability?
  3. How much of AI revenue is now scale-across, and is that mix proving accretive or dilutive to gross margin?
  4. Have memory and merchant-silicon costs worsened since May, and can ANET maintain 62–64% FY gross margin?
  5. What is driving deferred-revenue movement: more deployments, longer customer acceptance periods, or both?
  6. Are one or two additional 10%-plus customers still likely in 2026?
  7. What does management now expect for supply availability and customer demand entering 2027?
  8. Is campus progressing toward the $1.25B annual target, or is investment increasingly concentrated in AI?

Investment conclusion

ANET remains one of the more compelling public beneficiaries of AI infrastructure spending because it combines high-speed Ethernet switching, routing, software, and a credible expanding role in scale-across AI architectures. The company’s opportunity is no longer just hyperscale front-end networking; it is becoming a broader AI-fabric and distributed-data-center opportunity.

For this report, however, execution and conversion matter more than the demand narrative. Demand is already well established. The decisive issues are whether Arista can ship enough product, recognize more of its substantial deferred balance, protect margins amid component inflation, and raise the outlook again.

The clean bullish outcome: a revenue/EPS beat, full-year revenue and/or AI-target increase, stable gross-margin framework, and reassurance that supply commitments are enabling—not constraining—2026–27 growth.

The main downside outcome: a modest beat accompanied by unchanged guidance, increasing supply bottlenecks, slower deferred-revenue conversion, or a lower-margin outlook.