AEHR

Current Thesis

The pivot is VALIDATED, the tape is BLOWOFF. AEHR rerated from a SiC-wafer-test one-trick (ON Semi air-pocket) into the package-level burn-in (PLBI) picks-and-shovels supplier for hyperscale custom AI ASICs — and the market priced it in one violent leg: ~$20 to start 2026, ~$30 end of Q1, all-time high $113.20 on 2026-06-02 (+20.88% off a $99.65 open on the William Blair day). +400–465% YTD, the bulk inside ~2 weeks. Thesis right, entry wrong. The two prior DEFERs ($40→$113) were the documented alpha leak — but the lesson is "buy the clean breakout," not "buy the parabola's exhaustion candle." Theme ai-chip = ACCELERATING; THIS expression = mania/SATURATED. Operator call: no chase at [entry redacted] Probe-only. Want a higher-low retest of the $80–90 breakout shelf into the ~July 7 Q4 print before sizing up.

Bull Case

• 2026-04-16: record $41M follow-on production order (largest in company history) from lead hyperscale AI customer for PLBI of custom AI ASICs (Sonoma systems + turnkey BIMs + sockets); FY2027 deliveries (FY27 starts ~2026-06-27). It is a follow-on to the 2026-02 initial Sonoma order → repeat-buyer = cleanest signal in the file.
• 2026-04 (Q3 call): H2 FY26 bookings >$92M, past the high side of the raised $60–80M guide; Q3 book-to-bill ~3.6x. The pipeline is real, not narrative.
• Guidance: FY26 revenue at the high side of $45–50M, plus a return to non-GAAP profitability in Q4 FY26 — the first operating-leverage inflection of the pivot.
• The $60M ATM (filed 2026-04-08, William Blair / Craig-Hallum) the prior thesis feared is now largely defused: at $113 vs ~$30–40 filing-era price, the remaining capacity costs a fraction of the shares — minimal dilution per dollar, funds the Sonoma FY27 ramp.
• Cluster confirmation: the AI-test/photonics complex is ripping together — AEHR +15.13% leading AAOI, AXTI, AOSL higher on 2026-06-02; first US photonics ETF (EUV) launched 2026-05-13 with 15 triple-digit holdings. AEHR is one of only ~5 Russell 3000 names up >200% in 2026.
• Hyperscaler custom-silicon capex (Google TPU, Amazon Trainium, Meta MTIA, MS Maia) historically yields multiple follow-ons within ~18 months of first production order — $41M is unlikely to be the last from this customer; a 2nd DISTINCT hyperscaler order ≥$20M is the unfired HIGH-conviction trigger.

Bear Case

• Blowoff signature. +400–465% YTD, bulk <2 weeks, topped by +20.88% on 2026-06-02 — the exact William Blair presentation day (catalyst now PASSED). Buying 06-03/04 = buying the exhaustion of a parabola, not a setup.
• Insider selling into the rip (Yahoo flagged "powerful momentum AND insider selling"; Form 4 beneficial-ownership change). Management is not treating $113 as cheap.
• Valuation is extreme: P/S ~65, P/B ~21 on ~$59M trailing revenue. There is no fundamental floor near price — the floor is technical/sentiment.
• Saturation tells firing: mainstream "how-much-you'd-have-made" (2026-05-12) + "5 small caps up 400%" / "Russell 2000 biggest winners" coverage (2026-06-02/03). When the move IS the headline, the easy money is already gone.
• Whipsaw = distribution risk: recent single-session drops of −15% (to $84.62) and −12.5% (to $87.05) preceded rebounds. 12–21% candles are normal on a ~30M float — a stop too tight gets wicked out, too loose and you eat a 20% gap.
• Customer concentration: the $41M + $92M bookings still trace largely to ONE lead hyperscaler. A single push-out/cancellation re-rates the whole multiple. The diversification binary is unfired.

Setup & Price Structure

• ATH $113.20 (2026-06-02 close, +20.88%). The move's pivot levels: breakout shelf / gap base ~$84, prior whipsaw lows $84.62 and $87.05. Below those the rising 50-DMA is the mean-reversion magnet — still well under spot after a parabolic run.
• Structure is parabolic-extended, NOT basing. There is no clean higher-low yet — the name needs to either (a) build a flag above ~$84–90 and hold it for a week, or (b) flush to the 50-DMA and form a higher low vs the pre-breakout base before there's a low-risk entry.
• RSI is almost certainly >80 into the 06-02 spike (blowoff). For a picks-and-shovels a2 this is not an auto-trim, but on THIS tape treat it like a6: no adds, tight risk, no averaging down.
• Entry discipline: do NOT chase [entry redacted] The trade is the retest — a daily close back above ~$90 after a pullback that holds $84, ideally into the July print. A break-and-hold of a fresh high on volume AFTER a multi-day consolidation is the alternative re-entry.

Catalyst Calendar (next 30 days)

• ~2026-06-13 to ~2026-06-27: Q4/FY26 earnings blackout begins (~mid-June) — ATM issuance pauses (reduces paper-supply cap) but also removes a sentiment driver.
• ~2026-06-27: FY2027 begins — first FY with Sonoma $41M deliveries scheduled; framing/PR risk-on.
• No hard binary catalyst INSIDE the 30-day window (next print ~2026-07-07 Benzinga / ~2026-07-21 TipRanks is just OUTSIDE 30 days) → catalyst_date = null. The tape between now and the print is pure flow/sentiment — exactly the condition where chasing a parabola is most punished.
• Watch for (unscheduled): a SECOND distinct hyperscaler order PR (the real upgrade trigger) and any 13D/Form 4 insider-sell follow-ups.

What Would Change Our Mind

• Re-entry (bull): a pullback that holds a daily close ≥[entry redacted] then reclaims ~[entry redacted] and bases → probe; add on a 2nd hyperscaler order ≥$20M (HIGH-conviction upgrade) or a clean Q4 beat (rev >$18M + non-GAAP profit confirmed).
• Invalidation (bear): daily close below $84 = parabola broken, 50-DMA mean-reversion in play, no chase. The thesis itself breaks if the Q4 FY26 print (~2026-07-07) shows revenue <$18M or no return to non-GAAP profit, signaling the $92M H2 bookings are not converting — that flips this from "wait for entry" to "thesis impaired."
• Saturation flip: if the ai-chip/photonics theme rolls to SATURATED (peers AAOI/AXTI/AOSL break their own structure together) while AEHR is still extended, skip entirely — the cluster that confirmed the move also confirms its top.

Correlation Notes

• Trades as a high-beta AI-test/photonics basket member: AAOI, AXTI, AOSL traded up together 2026-06-02; the EUV photonics ETF (launched 2026-05-13) is the thematic proxy. Cluster strength = confirmation; cluster rollover = exit signal.
• Second-order to hyperscaler custom-silicon capex (TPU/Trainium/MTIA/Maia) — same demand driver as MRVL/AVGO ASIC names and MU/memory; AEHR is the burn-in/test toll on that buildout.
• Single-customer concentration means AEHR can DECOUPLE violently from the basket on company-specific order news (up on a 2nd order, down on a push-out) — do not treat the basket as a hedge.
• Macro: rate-sensitive, no-earnings-floor small cap; a RISK-OFF regime hits this 2–3x harder than the index. Size accordingly.