ANET: The Thesis Works, but $188 Is No Longer an Entry
Arista’s operating thesis still works; the stock’s setup does not. ANET closed at $187.81 on 14 September 2026, 21.9% above the $154.03 price at which the May underwrite said BUY, so today’s decision is to hold the business and stop chasing the multiple.
01ANET — HOLD
Why now. This is a quiet-day re-underwrite, not a new catalyst call. The evidence base is unchanged since the 25 May 2026 memo, but price has moved enough to change the action: ANET’s 14 September 2026 close was $187.81, versus the memo’s $154.03 entry and $188 base target (market data; Q1 filing).
The evidence. Arista reported Q1’26 revenue of $2.709bn, up 35.1% year over year for the three months ended 31 March 2026. Q1’26 gross margin was 61.9%, operating margin 42.7%, operating cash flow $1.694bn, and cash $2.790bn at 31 March 2026, with no debt in the memo’s underwrite (Q1 filing; Arista IR). Those figures still support the core case: high-speed Ethernet demand can compound while EOS protects customer stickiness. They do not justify treating the old target as a fresh entry.
Levels & triggers. Hold at $187.81; accumulate only at $160 or below, roughly 15% beneath the 14 September 2026 close, unless a later filing raises the earnings base enough to support a higher level. Confirmation requires revenue growth to remain at least 25% year over year and gross margin to remain at least 58% in each of the next two reported quarters. The thesis is killed by revenue growth below 25% for two consecutive quarters, gross margin sustainably below 58%, or a disclosed major switch-vendor loss at a top hyperscale customer (Q1 filing; Arista IR).
Horizon. The HOLD is for 12–24 months, with the decision reset at each quarterly filing.
The bear case. NVIDIA’s Spectrum-X can turn Ethernet’s growth into Arista’s margin problem: if customers prefer a tightly integrated GPU-and-network stack, Arista could lose share precisely as lower-margin AI switches become a larger mix. At $187.81 as of 14 September 2026, even a healthy deceleration toward the memo’s 25% floor would leave less room for valuation error than existed at $154.03 (NVIDIA Spectrum-X; market data).
What I’m watching: the next 10-Q’s year-over-year revenue growth and gross margin; a sub-25% growth print or sub-58% margin would move HOLD toward SELL, while a clean reset to $160 or below with those thresholds intact would move it back to ACCUMULATE.
02Sources
Independent equity analysis, for information only, not investment advice.
Originally published in the Axelrod Research newsletter.