June is when Micron has to do the work.

Why June stops being a sell-side story

Through May, the move was driven by analysts catching up to math that was already on the page. DRAM contract prices had moved 90–95% QoQ in Q1 2026 (per TrendForce contract pricing data). Hyperscaler capex guides were the largest in the history of the industry. Micron itself had already printed 75% gross margin in FQ2. The estimates had to come up. Analysts came around to it in sequence. The stock came around to it in compressed time.

That’s a story that runs out. There are only so many banks left to capitulate. Goldman is still the lone NEUTRAL on the name — though they’ve quietly walked their PT from $235 in December all the way to $400, while keeping the rating unchanged. Morgan Stanley’s $450 OVERWEIGHT was set back in February and called MU the firm’s top US semi pick — they haven’t needed to chase. When everyone else above $1,000 becomes consensus, the next leg has to come from somewhere other than sell-side mathematics.

It has to come from company numbers. From the actual income statement. From forward guidance. From management saying out loud what the spreadsheets have been implying.

The cleanest test arrives on Wednesday June 24, after the close. Micron reports fiscal Q3 2026 (the quarter ending May). It’s the first earnings print since the whole memory complex re-priced. Every number management gives is going to be measured against the run-up — not against last quarter, not against last year, but against the $1 trillion the market has now decided the company is worth.

The bar Micron has to clear

Here’s what management guided when they reported FQ2 on March 18, 2026 (per Micron’s FY2026 Q2 8-K). Coming off a quarter that printed $23.86B in revenue at 75% non-GAAP gross margin and $12.20 in EPS, the company told the Street to expect this:

Numbers like those are difficult to internalize. A quarterly revenue figure that, alone, exceeds Micron’s entire fiscal 2023 revenue of $15.5B (per the Micron FY2023 10-K). A non-GAAP gross margin profile that puts a memory company in the same neighborhood as software. EPS of $19 in a single quarter — the stock didn’t close above $63 across all of 2018, the previous cycle peak.

What anchors the stock at $923 isn’t that bar by itself. It’s the question of whether management is sandbagging again. The last two quarters they did:

Chart 2 — Revenue, gross margin, and the beat pattern

The last verified quarter beat its guide by 28% on revenue and 7pp on margin. June 24 tells us if that pattern holds.

Bars are actual revenue; the dashed line in the FQ2 bar shows where management originally guided when reporting FQ1. The 7pp GM beat on FQ2 is what produced the EPS surprise of $12.20 vs $8.42 guide. Source: Micron 8-K filings (FY25 Q4, FY26 Q1 and Q2 press releases), company guidance.

Two quarters of evidence is a pattern, not a guarantee. But it’s also the strongest possible signal we have about management’s posture: in this cycle, they’ve been conservative every single time. The right question isn’t whether $33.5B/$19.15 prints. It’s how far above those numbers it prints, and whether that’s already in the stock at $923.

What I’m actually watching on the print

The headline beat is almost ceremonial at this point. Three out of the last four quarters Micron has beaten the top of the range. The pricing data through May suggests they should again. The interesting reads aren’t whether they beat — it’s how, and what comes after.

One. Gross margin trajectory. 81% is a number that doesn’t sit comfortably in a memory company’s P&L. If they print 82–83%, the bull thesis (memory has structurally rerated, AI is permanent demand) gets actual fundamental support. If they print 79–80%, the bear thesis (this is peak; mix and ASP are already rolling) gets its first echo. The range matters more than the point.

Two. Forward guide for FQ4 and color on FY27. The Street is modeling FQ4 above the FQ3 guide. Anything that anchors FY27 EPS at or above $80 (against the $100 we’ve been using) keeps the trim ladder running. Anything that walks back FY27 commentary — even tonally — is the cycle-end signal we’ve been calibrated for since April.

Three. HBM3E ASP commentary. TrendForce flagged this in April: any 5%+ sequential decline in HBM3E contract pricing would compress margins meaningfully. If management strikes a defensive tone on HBM pricing for the second half — that’s the first place a cycle starts to roll. I will be re-reading the prepared remarks specifically for this line.

Four. Capex. Samsung’s $73B 2026 capex is the largest single-year semi spend in history. Micron’s response is the cleanest read on whether management thinks the pricing environment is durable. A capex guide that ramps Idaho aggressively into FY28 reads as confidence. A capex guide that throttles back reads as discipline — or fear.

The scenario tree (not a point prediction)

The honest answer to “what do you think prints” is that the range of defensible outcomes is wide enough that pretending I know the number is a tell. Here’s what I actually think the matrix looks like, given verified contract pricing data, the FQ2 beat pattern, and what management was hinting at on the last call:

The reason I’m not putting a point estimate on this: I genuinely don’t know which of those four lands. The probabilities below are illustrative subjective priors, not implied options or sell-side odds. The base case is most likely (~50%), the big beat is second-most-likely (~25%), guide-in-line is third (~20%), and the negative surprise is the tail (~5%). Those priors come from pattern recognition, not proprietary channel checks.

What I’m more confident about: the forward commentary on FQ4 and FY27 matters more than the print itself. UBS’s $1,625 PT requires roughly $100+ in FY27 EPS. DA Davidson’s $1,500 is similar. If management gives any tonal signal that walks those numbers down, the stock cracks regardless of how good the in-quarter beat looks. If management leans into FY27 the way they leaned into FQ3 last quarter, the rally has another leg.

The stock reaction is mostly a function of how stretched current consensus already is. Sell-side estimates for FY27 EPS have moved meaningfully higher across the May cascade. If the call validates $100+, those PTs hold. If it validates $80–85, the higher PTs start looking aspirational.

The June calendar (in one screen)

For context, the macro calendar this month is dense and the FOMC at the end of it is the next real test. For the broader macro setup — PCE, labor, the rotation within the AI complex — see the May 31 weekly recap. Here are the discrete events that move this position:

What jumps out is that MU’s print lands a week after FOMC. By the time we hear from Mehrotra, the rates picture for the rest of the year is set. The market will know whether we’re in “dovish hold” or “hawkish pause” before the most important AI-cycle data point of Q2 hits the tape. That sequencing is unusually clean.

Where the sell-side actually stands today

Eight notable changes since the May 28 piece. DA Davidson raised PT from $1,000 to $1,500 (May 28) — sits exactly on our bull case. Barclays raised from $675 to $1,175 (May 27). Mizuho raised a third time, from $800 to $1,150, after two prior raises in May. Susquehanna raised from $600 to $1,750 (May 29) — the new Street high above UBS, a $125 premium and a 192% increase in one note. Raymond James doubled from $530 to $1,100 (June 1). And then, through the Jun 4–5 selloff, two more capitulations landed on Monday Jun 8: Wells Fargo raised PT from $550 to $1,220 (+120%; Buy reiterated) and Cantor Fitzgerald raised from $700 to $1,500 (Overweight; analyst C.J. Muse called MU a “top pick” with memory undersupplied through 2028). Also: Goldman strategist Ben Snider hiked the S&P 500 target to 8,000 on May 27 (per Bloomberg coverage of the strategy note), naming Micron and NVIDIA as roughly a third of 2026 S&P EPS growth — while Goldman’s single-stock desk kept MU at NEUTRAL.

The pattern of those last two raises matters. Both Wells Fargo and Cantor raised after Broadcom’s Q3 total revenue guide came in slightly light of the most aggressive Street estimates and after the worst single session on the semiconductor index since April 2025. If the sell-side read either of those as cycle-end signals, you would not see another +120% PT raise on the same name three sessions later. You would see PT cuts. We got the opposite.

Chart 3 — Sell-side PT distribution, updated through June 8

Wells Fargo + Cantor raised through the drawdown. Eleven of eighteen actions at or above our base PT.

Spot price line is the June 8 intraday level around $948 — about 12% below the Jun 3 closing all-time high of $1,079.57. Goldman and Morgan Stanley still below current. Wells Fargo (Jun 8, +120% from $550) and Cantor (Jun 8, “top pick”) raised through the drawdown — the cleanest signal that the sell-side reads the pullback as positioning, not fundamentals. Source: bank notes via Bloomberg, Investing.com, MarketBeat, TipRanks, Gurufocus, Finbold, firm disclosures.

The cross-confirmation read — Goldman’s strategy desk endorsing the MU earnings thesis while Goldman’s single-stock desk stays NEUTRAL at $400 — is exactly the kind of internal dissonance that resolves loudly when someone capitulates. I’d watch for a Goldman PT raise sometime in the next four weeks. Morgan Stanley at $450 OVERWEIGHT is the other obvious candidate — their February PT was set when MU was below $200; they’ve had MU as their top US semi pick the whole way up.

Position update and what fires next

I trimmed one-third on Friday May 29 at an average fill of $919, moving the position from roughly 6% of book down to 4%. Cost basis on the remaining two-thirds is unchanged. The trim ladder continues:

The framework keeps doing its job. The first trim was mechanical — the trigger fired, I sold, no decision required. That’s the point of pre-committing. The hardest part of this trade was always going to be selling into strength while the narrative still felt durable. The ladder makes that decision in advance, so the in-moment version of me doesn’t have to be strong-willed about it.

Three things would change the plan before the next rung:

A clean miss on FQ3 revenue or guide. Stock back to $800–850 range. I hold the remaining position; the thesis is intact at lower multiples and the cycle just got a little longer. No additional buying — we already own enough.

HBM3E pricing softness in the call. This is the kill-shot I’ve been calibrated for. If management language goes defensive on HBM ASPs for second-half FY26, I exit a second tranche immediately on the call regardless of stock price.

Hyperscaler capex revision downward. Watch the FY27 capex guides as MSFT/GOOG/META report in July. Any meaningful cut from the current $745B aggregate is the leading indicator that supply > demand by 2028 starts pulling forward.

None of those have happened. Most of them probably won’t. The base case is that June is a confirmation month, the print is good, and the discipline plays out the way it’s designed to. But the framework only works if I’m as willing to fire it on bad news as I was on good.

I’ll write the follow-up the morning of June 25.

One-page summary