🧩

SIMO

💡
Thesis. Revenue up 127% at a 50.2% gross margin, because NAND makers are outsourcing controller work to Silicon Motion so they can point their own engineering at high-bandwidth memory. The memory boom pulls business toward this company rather than away from it.
Reviewed 11 September 2026, after Q2 results. Next review after Q3 earnings on 29 October 2026.

What it does

Silicon Motion designs controller chips for flash storage. It's fabless — it designs, someone else manufactures.
A controller is the small processor inside every SSD or embedded storage module. It decides where data gets written, handles errors, and spreads wear evenly so the flash lasts. Without one, raw flash memory is unusable.
Four product lines: SSD controllers, eMMC and UFS embedded controllers for phones and IoT, Ferri boot-drive storage for automotive and enterprise, and MonTitan enterprise SSD controllers. CEO is Chia-Chang Kou.
The interesting part isn't the product, it's who's buying it and why. The memory makers used to build their own controllers. Now they'd rather spend that engineering on HBM.

Bull case

  • Third consecutive record quarter. Q2 2026, reported 30 July: revenue $451.0M, up 127% on the year and 32% on the quarter
  • Margins came in above guidance. Non-GAAP gross margin 50.2% against guided 48.5–49.5%. Non-GAAP operating margin 23.1%. GAAP net income $136.1M, or $3.99 per ADS, against $16.3M and $0.49 a year earlier
  • Guidance implies it continues. Q3 $519–541M, up 114% to 124% year over year, with GAAP operating margin of 24.4–25.7%
  • Management is guiding to a doubling of annual revenue in 2026 and operating margins above 30% by year end
  • Every line is moving. SSD controllers up 50–55%, eMMC and UFS up 95–100%, and Ferri boot drive up more than sixteen-fold to nearly a third of total sales
  • The structural driver is the good bit. NAND makers are handing controller work to Silicon Motion so they can focus their own engineering on high-bandwidth memory. The HBM boom is a tailwind here, not a threat
  • A hyperscaler product is coming. A 4nm PCIe Gen 6 controller taped out in August, aimed at cloud providers for 2028
  • Real diversification from consumer electronics toward enterprise, automotive and AI infrastructure. Pays a $2.00 annual dividend per ADS

Bear case

  • Cash is draining into the ramp. Down to $181.8M from $210.9M last quarter and $282.3M a year ago, with $59.2M of new bank loans taken on. Working capital is funding the growth
  • Insiders have been net sellers recently
  • The end markets are shrinking. Smartphone and PC unit volumes are declining, so growth depends on winning more content per device and taking share — not on the market growing
  • Memory pricing cuts both ways. Higher NAND prices raise module-maker costs and could eventually suppress SSD unit demand, which is the volume base
  • The hyperscaler story is 2028. The 4nm PCIe Gen 6 opportunity is years from revenue, whatever the tape-out timing suggests
  • Fabless means dependent on foundry capacity in a market where capacity is tight
  • Expectations are high after a very large run

Major customers

  • NAND flash manufacturers — increasingly outsourcing controller design and production to Silicon Motion. This is the key relationship class and includes the major memory makers
  • SSD module makers and OEMs — buying merchant controllers off the shelf
  • Smartphone and IoT device makers — eMMC and UFS embedded controllers
  • Automotive OEMs and tier-one suppliers — Ferri boot-drive storage, the fastest-growing line by some distance
  • Enterprise and hyperscale customers — MonTitan enterprise controllers today, with the 4nm PCIe Gen 6 product aimed squarely at hyperscalers for 2028
🔗
Read-through. The relationship with
🧠
MU,
🧠
SNDK and the other memory makers runs the opposite way to what a simple "NAND buyer" framing suggests.
Because those makers are diverting engineering capacity to HBM, they are handing controller work to Silicon Motion. A tighter memory market is currently a tailwind here, not a headwind. That is the single most useful cross-read on this page, and it is easy to get backwards.
The automotive Ferri growth is unrelated to AI altogether and provides genuine diversification away from the rest of this file.

What would change the view

  • Whether Q3 lands inside the guided $519–541M — a third straight triple-digit growth quarter would be strong confirmation
  • Operating margin progressing toward the stated 30%-plus target
  • Cash and bank loan balances — the working capital drain is the clearest weak point
  • Ferri and Boot Drive as a share of sales, currently around a third
  • 4nm PCIe Gen 6 design wins with hyperscalers
  • Insider transaction filings

Update log

  • 11 Sep 2026 — Reviewed. No change to the thesis or the verdict.
  • 14 Aug 2026 — Full review after Q2 2026 results. Corrected an earlier read in this file that characterised Silicon Motion as inversely exposed to NAND pricing with only a thin AI link. That was wrong: the outsourcing dynamic means tight memory is a tailwind. The thesis was rewritten around it.
Research and education only — not investment advice.