
What happened
Madison Investments said Arista Networks led its Mid Cap Fund in Q2 2026 on AI data center networking demand and a growing backlog, then reduced the position. Arista guides for at least US$3.5b to US$3.6b of AI fabrics revenue in 2026, within about US$12.6b total.
Why it matters
At least one large institutional holder views Arista's share price as already reflecting its strong demand and backlog, so the trim looks more like valuation-driven profit taking than a shift in the near-term catalyst, per the report.
What to watch
The larger swing factor remains customer concentration — a pause in AI data center buildouts by one or two major buyers would matter more than a single fund selling. Watch whether large cloud providers keep favoring Ethernet-based solutions over proprietary alternatives.
WHO IT HITSThis lands on institutional portfolio managers holding AI infrastructure names, who now have a live example of a large fund trimming a leader on valuation rather than on demand. It also matters to Arista's investor-relations and corporate-strategy teams, since customer concentration and the Ethernet-versus-proprietary choice are the variables the story says matter most.
Summaries like this, in your inbox every morning.
The story turns on who is selling and why. Madison Investments told investors that Arista Networks was a leading performance contributor for its Mid Cap Fund in Q2 2026, yet the same fund reduced its position. Read together, those two facts point to a valuation judgment rather than a change of heart about the business: the demand and backlog that drove the quarter's gains are, in this holder's view, already in the price.
The article frames the debate around what has to keep going right. Arista's 2026 AI fabrics guidance of at least US$3.5b to US$3.6b, inside roughly US$12.6b of total revenue, rests on large cloud providers continuing to favor Ethernet-based solutions over proprietary alternatives, and on Arista broadening beyond a handful of hyperscalers into enterprise and campus networking. If either side underdelivers, the high margin profile and current valuation expectations could come under pressure, the article notes. Customer concentration is the swing factor, and the more cautious analysts cited already flag reliance on a few hyperscalers as a possible cap on the stock's appeal.
The distance between the camps is wide. One bullish view pencilled in about US$24.4b of revenue and US$8.7b of earnings by 2029, while the narrative in the article projects US$21.4b of revenue and US$8.2b of earnings by 2029, requiring roughly a doubling of earnings from about US$4.0b today. For holders, the outcome appears to hinge on whether hyperscaler AI data center buildouts keep pace and on whether Arista can widen its customer base beyond those few large buyers.
For example, today's edition would include:
AI-summarized, only the topics you pick: one digest a day via Email, LINE, or Slack.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
Google DeepMind chief Koray Kavukcuoglu said Gemini 4 has entered post-training and Google intends to ship an…

T-Mobile US is rolling out AutoPilot inside its Self-Organizing Network, which can automatically reconfigure n…

CrowdStrike Holdings (NasdaqGS:CRWD) became a founding member of the Blueprint Alliance, joined by AWS, Google…

S&P Global Ratings revised its outlook on Western Digital to positive from stable and affirmed its 'BBB-' issu…

On Sept. 18, Disney's Horacio Gutierrez told legal staff the department will get smaller, citing automation an…

At Dreamforce 2026, Salesforce presented an agentic health enterprise, citing UCLA Health's 75,000+ patient in…
