This weekly review checks the prior week's published stock posts against subsequent market action and company-specific developments. The review window runs Friday May 22 through Wednesday May 28, 2026, Europe/Dublin time. Price moves use the latest available market close (May 28, 2026) at the time of writing.
Summary
Two AI infrastructure names published this week: Arista Networks (ANET) and Micron Technology (MU). ANET is flat-to-up on no company-specific catalysts. MU pulled back modestly from publication price but gained a second major analyst re-rating (Susquehanna, $1,750 target) confirming the structural memory-as-AI-infrastructure thesis.
— Published May 25. Bullish at $154.03. Latest $155.27. Move: +0.81%. Current read: tracking.
— Published May 27. Watchlist add-on-pullbacks at ~$946. Latest $923.52. Move: -2.38%. Current read: tracking, pullback toward the add zone.
Arista Networks (ANET)
1. What happened
ANET closed at $155.27 on May 28, +0.81% from the $154.03 publication reference. The stock traded in a tight range ($149–$158) across the three trading days (May 26–28), with no sharp moves either direction. Volume was elevated on Tuesday May 26 (11.8M vs ~7M average), which coincides with the Bechtolsheim insider sale.
The notable development: Andreas Bechtolsheim, Arista co-founder and 10% owner, sold approximately $34.5 million in stock on May 26. This is worth flagging but requires context — Bechtolsheim is a regular seller under pre-arranged plans, and one sale does not invalidate the AI switching thesis. However, at a time when the stock is 15% below its April highs, any insider selling carries more signal than a sale at an all-time high would.
Broader sector context: Cisco has rallied 32% in the past month while Arista has fallen 10%, suggesting the AI networking trade is fragmenting rather than lifting all boats. Arista's open-standards Ethernet thesis is a long-term play and does not need to win every month.
2. Thesis check
Nothing in the three trading days since publication contradicts the core thesis: Arista switches attach to GPU clusters at near-linear revenue ratios, hyperscaler capex is still accelerating, and Ethernet continues to gain share over InfiniBand. The Bechtolsheim sale is a yellow flag — it merits attention at Q2 earnings but does not break the bull case on its own. The Cisco outperformance is noise at this timescale; Arista's competitive position in the hyperscaler switching fabric has not changed.
3. Action now
The original buy thesis at $154 remains valid. The stock has barely moved (+0.81%), so the entry window is essentially unchanged. The $130-140 add zone from the original post is still the place to add on weakness. The Bechtolsheim sale adds a data point to monitor but is not an actionable sell signal given the pattern of pre-arranged insider sales at Arista.
4. Next checkpoint
Q2 2026 earnings, expected late July/early August 2026. Watch for: (1) revenue growth staying above 25% YoY, (2) gross margin stabilization above 60%, (3) any Spectrum-X competitive commentary from management, and (4) whether Bechtolsheim files additional Form 4 sales in the weeks ahead. If gross margins dip below 58% or NVIDIA Spectrum-X wins a design at a top-3 hyperscaler, the thesis weakens.
Micron Technology (MU)
1. What happened
MU closed at $923.52 on May 28, -2.38% from the ~$946 pre-market reference at publication on May 27. The stock experienced significant intraday volatility on publication day (May 27): it opened at $955.66, traded as high as $956.16, then sold off to $888.15 before recovering to close at $928.41. On May 28, it opened at $929.83 and closed at $923.52, continuing the modest drift lower.
The major development: Susquehanna raised its price target on MU to $1,750 from $600 on May 29 (after the review window but highly relevant), maintaining a Positive rating. This is the second massive target increase in a week — UBS went from $535 to $1,625 on May 26. Two top-tier banks independently arriving at targets near $1,600-1,750 in the same week is a strong structural validation of the memory-as-AI-infrastructure thesis, even if the stock is digesting the UBS day gains.
The semiconductor sector broadly rose on May 29 (AMD, NVDA, MU all mentioned as gainers), suggesting the May 27-28 pullback was profit-taking after the UBS day spike rather than thesis deterioration.
2. Thesis check
The Susquehanna upgrade is a material confirmation of the structural re-rating thesis. Two major banks (UBS and Susquehanna) have now independently repriced Micron from cyclical-memory multiples to AI-infrastructure multiples within a single week. Both cite the same drivers: HBM sold out through 2026, long-term supply agreements reducing cyclicality, and DRAM pricing strength driven by AI capacity allocation. The stock's inability to hold above $946 despite the UBS upgrade is a near-term cautionary signal, but the fundamental narrative is stronger today than it was at publication.
The original post's framework — buy on pullbacks to $700-750, wait for a better entry if not already owned — looks prescient. The stock's post-spike drift confirms that buying at the all-time high after a 19% surge was not the optimal entry.
3. Action now
The original Watchlist stance with a $700-750 add zone remains the right posture. The Susquehanna upgrade adds conviction to the structural thesis, but conviction alone does not fix an elevated entry. At $923, MU is still above the recommended add zone. The original $3-5% combined portfolio sizing and June 24 earnings catalyst checkpoint remain appropriate. If the stock continues drifting toward $800-850 on no fundamental deterioration, that would begin to look attractive for a first entry.
4. Next checkpoint
Q3 FY2026 earnings, expected ~June 24, 2026. Revenue threshold: $37B+ validates the UBS/Susquehanna re-rating. Below $33.7B (low end of the estimate range) would be a problem. Watch for: (1) gross margin trajectory staying above 70%, (2) HBM pricing commentary and capacity allocation, (3) any additional major analyst target increases following the UBS/Susquehanna pattern. If the stock reaches $700-750 before earnings on no fundamental news, that is the add zone. Below $500, the cyclical thesis has reasserted and the position should be reduced.