Arista in the AI stack

The business and what the filings show

Arista sells high-speed Ethernet switches and the software that runs them, to cloud and enterprise customers. Over its latest four quarters it reported $10.5B of revenue, +33%, at a 63% gross margin and a 43% operating margin, among the highest in the networking layer.

AI clusters need fast back-end networks between accelerators, and the industry’s move toward Ethernet for those networks plays to Arista’s strengths. Its customers are concentrated among the largest cloud buyers, and its 10-K discloses each customer above 10% of revenue.

Arista’s competition comes from Nvidia’s own networking, from Cisco, and from hyperscalers building switches to their own designs with contract manufacturers. Its margins are the sign of whether its software keeps it ahead of that pressure.

How AI shows up in the numbers

AI shows up at Arista as switching for the back-end networks that link accelerators and the front-end networks that connect them to storage and users. AI clusters need far more bandwidth per server than traditional workloads, which is why networking spend rises with every new generation of accelerators. Arista’s results therefore move with the size of the clusters its customers build, not just the number of servers they buy, and its customers’ shift from InfiniBand to Ethernet adds to that.

For scale, Arista’s revenue over the latest four quarters equals 2% of the $566B the five largest buyers spent on capital projects over their latest four quarters. Their capex rose +81% on the year before, while Arista’s revenue grew +33%: slower than the spending, so it is capturing a smaller share of each new capex dollar, or more of its business sits outside the build-out.

How it talks about AI

In its latest earnings release, furnished to the SEC on August 4, 2026, Arista mentioned AI 8 times in 827 words of narrative text, about 9.7 mentions per 1,000 words. Of its 6 sentences about AI, 1 included a number: a revenue figure, a growth rate, a customer count or another measurable result. Across the 8 releases we read, back to November 7, 2024, AI mentions per 1,000 words went from 8.0 to 9.7.

Specific claims are easier to check against the numbers than general enthusiasm, which is why the share of AI sentences that carry a figure is worth watching alongside how often AI is mentioned. The sentences below are quoted from the release as published, after removing safe-harbor text and company boilerplate.

• “Arista introduced the 7060XE7 Series, a next-generation 1.6-terabit portfolio for AI fabrics – The new Etherlink™ platforms deliver up to 100 Tbps of system bandwidth and support Linear Pluggable Optics, cutting interconnect power consumption by approximately 60% compared to traditional pluggable optics.”
• “Introduced 1.6 Tbps AI fabric platforms, including liquid-cooled options optimized for scale-up, scale-out, and scale-across networks.”

Source: earnings release (Form 8-K, Exhibit 99.1), August 4, 2026.

The trend in the numbers

Since the start of 2023, Arista’s largest quarter by revenue was Q2 2026 at $3.0B, and its highest gross margin was 65% in Q2 2025. The latest quarter set the revenue record for the period.

Quarter by quarter, Arista’s revenue grew +27% in Q3 2025, +29% in Q4 2025, +35% in Q1 2026, +38% in Q2 2026 on the same quarter a year earlier. Growth has been accelerating.

Gross margin in the latest quarter was 63%, against 65% a year earlier, so more of each new dollar is passing through to costs. Operating margin moved from 45% to 45% over the same span. Over the latest four quarters it turned 50% of its revenue into operating cash flow and spent 1% of revenue on capital projects, against 1% a year earlier.

Where it sits in the stack

Arista sits in the networking and optics layer of the AI stack. It is paid by the AI clouds and servers layers; its own spending flows to chips, foundry & tools.

Among the 10 companies in this layer on our map, it ranks number 4 by revenue and number 5 of 10 by revenue growth. Its gross margin of 63% compares with 53% for the layer as a whole, a sign that it keeps more of each dollar than its neighbors. The full layer analysis is on the networking page.

Latest four quarters for 10-Q filers; latest fiscal year for 20-F filers. Companies under $100 million of revenue are left out.

Against the 10 US-listed companies in its layer, Arista grew +33% over the latest year versus +25% for the layer combined, and converted 49% of revenue into free cash flow versus 19% for the layer. It is taking share of the layer’s revenue.

What has to go right, and what could go wrong

What has to go right. Ethernet has to keep winning AI back-end networks, and Arista’s software has to keep it ahead of both Nvidia’s networking and hyperscalers’ own switch designs.

What could go wrong. Concentration in a few cloud customers, and competition from white-box switches built by contract manufacturers, could pressure its margins as AI networking matures.

These describe the conditions the business depends on, as the filings show them. They are not a view on the stock.

Through the lens of Investing in AI

Rob May’s book offers six mental models for telling AI moats from traps. The one most useful for Arista’s layer is “AI is a collection of technologies”: Neural nets, Bayesian methods, symbolic systems, and evolutionary algorithms have different cost structures, moats, and risks. Treating AI as one bucket hides the differences that matter. Invest in defensibility, not in a single paradigm.

The question it asks of any company is: What kind of AI is this, and how hard is it to replicate? Applied to Arista, the filings on this page are the place to start: the margin trend shows whether it holds scarce value or passes it through, the customer and financing disclosures show who it depends on, and the commitments show what it has promised before the revenue arrives.

Customers and concentration

US companies must disclose any customer that accounts for 10% or more of revenue, and foreign filers usually describe their largest customers. Concentration matters because a company’s growth is only as durable as the handful of buyers behind it, and in the AI stack those buyers are often the same few companies.

• “Sales to these two customers represented 26% and 16% of our total revenue for the year ended December 31, 2025, respectively.” 10-K, February 17, 2026

Commitments and backlog

• Backlog. Arista doesn’t report a comparable total of remaining performance obligations in its XBRL data.

What to watch

Arista’s next quarterly report will cover the quarter ending around September 29, 2026. Large companies must file a 10-Q within 40 days of a quarter’s end and a 10-K within 60 days of the fiscal year’s end, so the new figures should arrive by about November 28, 2026 at the latest. This page, the layer pages and the next quarterly report all update from that filing.

• Revenue from AI back-end networks and what the company says about them, the core of its AI exposure.
• Customer concentration in the 10-K, since a few cloud buyers account for much of its revenue.
• Gross margin against Nvidia networking and white-box competition, the test of whether its software keeps it ahead.
• Its place in the next quarterly AI Stack Report, which updates every company on the map from the new filings.

Figures from Arista’s SEC filings, latest period ended June 30, 2026. Other companies: all company pages.