SIMO · Silicon Motion Technology Corporation

Current Thesis

A fabless NAND-controller pure-play levered to the worst memory shortage in roughly 15 years. SIMO doesn't make flash it makes the controller silicon inside nearly every SSD, eMMC and UFS module, so it collects a toll on the NAND buildout regardless of which memory maker wins the wafer. The narrative is ACCELERATING and cluster-confirmed: Q1 2026 (reported 2026-04-30) was a record $342.1M, +105% YoY, EPS $1.58 vs $1.28 consensus, and the stock gapped +31% on the print. TrendForce now pegs Q2 NAND contract prices +70–75% QoQ versus the +55–60% management flagged on the call, so the cycle is still steepening. Sell-side is chasing the move up rather than catching it BofA went $320→$450 on 2026-05-27. RSI cooled from a parabolic 89 to ~71 while price held its base: digestion, not exhaustion. The next binary (Q2 print) isn't until ~2026-07-29, so there is no near-term earnings blackout on a fresh entry.

Bull Case

• Q1 2026 (2026-04-30): record revenue $342.1M, +23% QoQ, +105% YoY; non-GAAP EPS $1.58 vs $1.28 consensus (~23% beat); operating margin above guide; stock +31% on the print.
• Q2 2026 guide (issued 2026-04-30): revenue $393–411M (+15–20% QoQ), gross margin 48.5–49.5%, operating margin 21–22%; management reiterated a ~50% GM target and record annual revenue for 2026.
• Memory supercycle intensifying: TrendForce Q2 2026 NAND contract prices +70–75% QoQ, DRAM +58–63%; NAND supply deficit 4.2%, the tightest since 2011. Controller content per drive rises with PCIe Gen5 and higher channel counts.
• MonTitan enterprise leg: SM8366/SM8388 Gen5 controllers (>14 GB/s seq read, 3.3M IOPS) showcased in the NVIDIA AI ecosystem at GTC (2026-03-16); guided to 5–10% of 2026 revenue net-new AI-datacenter TAM on top of the consumer base.
• Sell-side chasing up: BofA $320→$450 Buy (2026-05-27), B. Riley $250→$312 (May 2026), above the prior cluster (JPM $260 / Susquehanna $275 / Wedbush $230). Strong Buy consensus, 10 buy / 0 sell / 1 hold.
• Cluster confirmation: the whole memory complex (MU, SK Hynix, Samsung, Sandisk/WDC) is breaking out on the same shortage; SIMO is the controller toll-booth, agnostic to the wafer winner.

Bear Case

• The shortage is double-edged. SIMO controller revenue is unit/content-driven, not NAND-$-driven, and IDC projects smartphone shipments -12.9% and PC -11.3% in 2026 as memory inflation throttles consumer demand. A 2H volume air-pocket is the live risk.
• Supply dependency: SIMO "secured supply via NAND partnerships," meaning it is hostage to Samsung/SK Hynix/Micron allocation. The same HBM wafer reallocation driving the shortage can starve the client SSD/mobile controllers SIMO sells into.
• Extension/valuation: ~$277 sits above the consensus average PT (~$257) and above every May target except BofA $450 and B. Riley $312. The dispersion (high $450 / low $145) is real disagreement on cycle durability.
• Cycle-peak risk: memory is brutally cyclical. The metric the mania runs on sequential NAND contract pricing rolling over would re-rate the entire complex fast.
• Sentiment ran hot: RSI 89 parabolic on the post-Q1 gap. Cooled to ~71, but a lot of good news is already in the tape ahead of the 2026-07-29 print.

Setup & Price Structure

• ~$277 currently, versus $263.85 close on 2026-05-15 (Cramer mention) and a +7.2% session on 2026-05-19.
• Post-Q1 breakout: the +31% gap on 2026-04-30 carved a base near $230; price has held above it and grinds higher into open air above the prior shelf.
• RSI path: ~89 parabolic (mid-May) → ~71. Momentum is unwinding the overbought condition without losing structure.
• Trading above all May analyst PTs except BofA $450 effectively in price discovery, which in an ACCELERATING + cluster-confirmed regime reads as a continuation rather than a fade.

Catalyst Calendar (next 30 days)

• ~Monthly (around the 10th): TrendForce / contract-pricing prints on NAND and DRAM the high-frequency confirm/deny on the supercycle thesis. Watch for any sequential deceleration.
• 2026-07-29 (est., just outside the 30-day window): Q2 2026 earnings, after close. The binary guide was $393–411M; beat-and-raise extends the leg, a guide-down on consumer unit weakness breaks it. No blackout inside the next 30 days.
• Ongoing: MonTitan / enterprise SSD design-win headlines on the NVIDIA-ecosystem cadence following GTC (2026-03-16) and Embedded World 2026.

What Would Change Our Mind

• Daily close back below the post-Q1 breakout base (~$230), forfeiting the April gap.
• Q2 revenue (2026-07-29) below the $393M guidance floor, or gross-margin guide below ~48%.
• TrendForce NAND/DRAM contract prices turn sequentially negative the supercycle metric rolling over.
• The memory complex (MU / SK Hynix) breaks down while SIMO fails to hold its base cluster failure.
• An explicit NAND-maker allocation cut that starves SIMO's client controller volume.

Correlation Notes

• High-beta proxy on the memory complex: MU, SK Hynix (000660.KS), Samsung, Sandisk/WDC, and direct controller competitor Phison (8299.TW) Phison prints are the cleanest read-through.
• 2nd-order AI: the demand chain runs AI datacenter capex → HBM wafer reallocation → NAND/DRAM shortage → enterprise SSD + content gains, so it co-moves with the AI-capex complex on the storage leg.
• Inverse exposure to consumer hardware units (smartphone/PC) a structural offset against the memory-price tailwind.
• Taiwan ADR: carries TWD and cross-strait geopolitical tail risk distinct from US-domiciled memory peers.

Notes

This refresh upgrades the read from the prior cautious stance: RSI has cooled off the 84–89 parabolic zone to ~71, a full dossier now exists, and BofA's $450 (vs the May $230–275 cluster) plus the intensifying contract-price data re-confirm acceleration rather than a late-cycle chase.