In short: Snipe's final trade (42:21). "Data center networking is still extremely robust." Fits his broader call that AI-infrastructure capex runs from $900 billion this year to $1.6 trillion next.
Arista sells the high-speed switches that connect thousands of servers inside a data center. When companies build AI data centers, they need much more of this networking gear. Snipe's pick rests on AI spending still rising sharply next year, which keeps orders for Arista's equipment strong.
In short: Cited third on revenue growth: "Arista Networks: +32.0%" — though it is eighth on the post's own sort (Kaspi, PDD, Lantheus, Harmony and Halozyme sit between). EPS CAGR 40.7%, ROIC 51.5%, FCF margin 44.9%, net margin 35.2%. Listed only.
In short: Named twice. First inside Brown's NVIDIA argument as the exhibit for how much AI-driven market cap sits in names most people can't identify: "Arista Networks — you probably don't even know what it is — is a $235 billion market cap." Then again on Evercore's top six tech-hardware stocks to own for the second half, where the note says the market is underwriting the durability and length of Arista's growth. Brown owns it off his best-stocks list: "this is another one right off the best stocks in the market. We've written about it… even when they have these dips, they still remain in their longer term uptrends just because of the sheer magnitude of the growth and share price."
In short: The other name Tom Lee added to his core list, and one Terranova already owns: "it's been a core holding. It's participating in AI networking along with Cisco and Ciena. We've had it for quite some time. It's worked incredibly well, and the revenue growth is accelerating." Simpson concurs: "I agree with Tom on Arista." Notable against Terranova's own framing of capital rotating out of semis — networking is the part of the AI complex he is not trimming.
Arista makes the high-speed switches that move data between the thousands of chips inside an AI data centre. It is the plumbing of the AI buildout rather than the processing, and Terranova has held it as a core position for some time.
His argument is the cleanest kind: revenue growth is accelerating. Tom Lee added it to his core ideas list today, and Simpson agrees with the call.
What makes it notable is the contrast with the rest of Terranova's day. He describes money leaving momentum names, memory chips and semiconductors — and yet networking is the part of the AI complex he is keeping, alongside peers Cisco and Ciena.
In short: Named as Cisco's competitor and, like Cisco, "a big beneficiary of AI data centers." Referenced as the peer set, not rated.
15:47EPS was a$122 up 23% versus last year. The company raised both EPS and revenue guidance. Again, not a surprise, but the stock was down after hours as the guidance did not meet the whisper numbers. Moving on, I want to talk about upstarts and established franchises and who wins and who loses and why. One way to frame the last few decades and AI is the transition of all industries from analog to digital.
In short: Supply catches up. Q2 revenue +38% Y/Y to $3.0B (a $210M beat) and adjusted EPS $1.02 ($0.13 beat), with adjusted operating margin reaching 50% against the 46–47% guided. The change is supply, not demand: Arista has secured memory for 2026, expanded manufacturing and distribution capacity and extended visibility into 2027 — management still expects industry shortages to persist through 2028, but better availability is finally letting it convert backlog into revenue. AI demand stays exceptional: Etherlink AI fabrics now have 100+ cumulative customers, up from 4–5 in 2024, and purchase commitments climbed to $9.7B as Arista pre-buys components ahead of deployments. FY26 revenue guidance raised for the third time this year to roughly $12.6B (+40%) from $11.5B, with operating margin guidance lifted to 48–49%; Q3 revenue of $3.3B came in ~$350M above consensus. "The remaining constraint is no longer customer demand, but how quickly Arista can secure enough components to ship into it." A disclosed author holding.
Arista makes the high-speed switches that move data between servers inside data centres — the plumbing of an AI cluster. Sales rose 38% to $3.0 billion, and half of every dollar of revenue became operating profit, better than the 46–47% margin management had promised.
The change this quarter was on the supply side. Arista has now locked in the memory chips it needs for 2026, expanded its manufacturing and distribution, and has visibility into 2027. It still expects the industry-wide shortage to run through 2028 — but with parts secured, it can finally convert orders it already holds into shipped revenue.
Demand is not in question. Its Etherlink AI networking product went from four or five customers in 2024 to more than 100. Arista has also committed $9.7 billion to buying components in advance — a company only pre-pays on that scale when it is confident the orders are real.
Guidance rose for the third time this year, to roughly $12.6 billion (up 40%), with margins raised too. The line that captures it: the constraint is no longer finding customers, it is getting enough parts to serve them. The author owns it; analysis, not a recommendation.
In short: Brown Advisory (Large-Cap Sustainable Growth): the software-defined case for the switching layer. Arista's "software-centric architecture, anchored by its proprietary Extensible Operating System (EOS)," plus open Ethernet, delivers "superior telemetry, automation and network management… better performance, reliability, and total cost of ownership versus competitors." They expect it to be "a structural share gainer in networking," delivering "over 20% revenue growth and peer-leading profitability for years to come," and initiated into "temporary stock weakness during the quarter due to transitory supply chain concerns."
Arista makes the switches that move data between the thousands of chips inside an AI data centre. Brown Advisory's argument is that this is really a software business wearing hardware clothes: one operating system, EOS, runs across the whole range, so customers get better monitoring, automation and reliability, and end up paying less over the life of the equipment than they would with a rival.
They expect Arista to keep taking share as AI pushes networks to be rebuilt more often, with revenue growing more than 20% a year and the best profitability in its industry. They bought during a dip caused by short-term supply problems — the sort of wobble that has nothing to do with the long-run case.
Full passage: premium transcript (PDF).
In short: Rossbach (new): best-in-class networking hardware + software (took share from Cisco for years); once hyperscalers clear permits/construction/chips/memory, "their attention will shift to the network" — nobody wants latency. Deepening with Meta and Microsoft, expanding with Anthropic and Oracle; ~28% revenue-growth guidance looks achievable (deferred revenue building, big purchase commitments); 35%+ EPS growth 2026 underpins 41x. Risk: TSMC-centered supply pressure.
Arista makes the network switches and software that shuttle data between chips inside giant data centers. Rossbach's sequencing insight: hyperscalers have been consumed by permits, construction, chips and memory — but once those clear, the constraint moves to the network, because latency wastes everything else you built. Arista already serves Meta and Microsoft and is expanding with Anthropic and Oracle; guidance of ~28% revenue growth is supported by building deferred revenue and purchase commitments. It's expensive (41x) but he argues 35%+ earnings growth justifies it.
In short: Where bulls may rotate instead of the capital-intensive hyperscalers — power generation, semiconductors, and "tech equipment companies like Arista and Cisco."
Arista makes high-speed networking gear that wires together the servers inside AI data centers. Eisman lists it (with Cisco) as the kind of "tech equipment" company AI bulls might rotate into — alongside power generation and semiconductors — instead of the hyperscalers that now have to raise huge amounts of capital. It's the "sell the picks and shovels" side of his supplier-vs-operator logic, with the caveat that if token pricing or power limits bite, "all bets are off."
11:04Communities are starting to push back. I'm getting the impression that the level of AI data center construction is slower than the bulls are hoping for. This risk bears tracking. The AI story has definitely changed because of the capital intensity. That's what's new. That means I believe that for investors who remain bullish, they may shy away from the capital intensive hyperscalers and focus more on power generation, semiconductors, and tech equipment companies like Arista and Cisco.
In short: A networking seller into AI demand — a cyclical "right place, right time" winner he doesn't expect to last 3–5 years.
3:54So they have immense pricing power. We have advanced packaging, chiplets, HBM integration and so on. We have HBM Memory, SK Hynix, Micron, and Samsung. Micron literally today just surpassed a trillion dollar market cap going up 18%. Over the past year, the stock is up like 10x because Micron is selling memory. All of a sudden, everybody wants memory.
In short: The subject of the dive, and a pass on price. "Arista Networks is a great business but we don't like the valuation level the company is currently trading at. We're happy to stay patient and wait for better opportunities in the market." Quality is not in doubt — Total Quality Score 7.9/10, gross margin 63.5%, ROIC 41.0% (5-yr avg 49.1%), ROE 28.8%, net margin 38.3%, FCF/net income 142.9%, net cash, CAPEX/revenue 1.1%, R&D/revenue 13.5%, revenue +26.9% and EPS +34.6% a year over five years, FCF +46.7% a year over ten, 36.7% CAGR since the 2014 IPO (4,089.4% total against SPY's 396.8%). The two failures: SBC at 12.6% of net income, "a red flag"; and 37.7x forward PE against a 32.1x ten-year average, with the reverse DCF demanding 18.9% FCF growth for ten years for a 10% return. Earnings Growth Model: 11.0%. Called an Owner-Operator Stock — insiders hold 17.1%, CEO Jayshree Ullal 2.2% ($4.8bn), co-founder Andreas Bechtolsheim 14.6%. Risks named: 35% of revenue from Microsoft and Meta, technological disruption, and inventory built for long lead times.
Arista makes the switches and routers that move data between the computers inside a data centre. Every AI system and every cloud service depends on that plumbing, and Arista sells it to Microsoft, Meta, Google and Oracle. It is the gold-rush shovel seller: it gets paid whichever AI company wins.
By almost every measure of business quality it is exceptional. It keeps 63.5 cents of gross profit on every dollar of sales, earns a 41% return on the money invested in the business, has more cash than debt, spends barely 1% of revenue on physical equipment, and turns more cash into free cash flow than it reports as profit. Sales have grown about 27% a year for five years and free cash flow nearly 47% a year for ten. Since listing in 2014 the shares are up more than 4,000%, against under 400% for the S&P 500.
Two things stop it being bought. First, staff are paid heavily in shares — 12.6% of profit, and 16% averaged over five years — which quietly transfers value from existing shareholders and is called a red flag. Second, the price. The shares cost 37.7 times next year's expected profits against a ten-year average of 32.1, and working backwards from today's price the company would need to grow its cash generation by 18.9% every year for a decade just to deliver a 10% annual return.
So the conclusion is a wait, not a rejection: "Arista Networks is a great business but we don't like the valuation level the company is currently trading at. We're happy to stay patient and wait for better opportunities."
The risk to watch if it ever does get cheap: 35% of the revenue comes from just Microsoft and Meta, and those two customers are large enough to build or buy alternatives.
In short: Fifth-best YTD performer at +29.4%, and the strongest long-run compounder in either table — 5-yr CAGR +55.3%, 10-yr +46.0%. Performance table only; no view.
In short: The month's best performer in the investable universe at +40.7%. Listed in the performance table only — no thesis, no valuation, no rating.
Nothing matches this filter.
Verbatim excerpts from the public transcripts (auto-pulled at each mention's timestamp, lightly cleaned). Timestamps link into the video; "source page" opens that commentator's full analysis at this row. Click a mention's header line to expand it (one open at a time). For personal study — not investment advice.