Silicon Motion just printed +105% revenue growth and guided Q2 up another 15–20%. In two trading days, three of the four stale price targets on the name have refiled at $250 or above — including the one with a perfect track record. The screener-trap is inverting in real time.

Pick: SIMO at $218.78. The old earnings model broke and the analyst stack is repricing it. Logged at the close of Thursday, April 30, 2026.

On Tuesday after the close, Roth MKM had a $140 price target on Silicon Motion. By Wednesday morning, that target was $250. A 79% raise in a single day from a five-star analyst is not an upgrade — it's a model rebuild. It's an analyst looking at a print so far ahead of his estimates that the only honest response is to throw the spreadsheet out and start over.

SIMO is not cheap after a 46% two-day move, and that is not the point. The point is that the old model just broke. The company printed $342.1 million of Q1 revenue against roughly $300 million expected, grew revenue 105% year-over-year, beat EPS by more than 20%, and then guided Q2 to roughly $402 million at the midpoint. The stock reacted immediately. The analyst stack is now catching up in waves, and what was a $180 average target on Tuesday is a $233 average target on Thursday. That gap between price discovery and consensus cleanup is the trade.

Q1 2026 came in at $342.1 million in revenue against a $299.6 million consensus — a 14% top-line beat. Year-over-year revenue grew 105%, sequentially 23%. Non-GAAP EPS was $1.58 against $1.28 consensus, a 23.4% earnings surprise. Both sides cleared the threshold I care about, and by a wide margin.

The mix is what matters more than the headline. Embedded controllers — the eMMC and UFS chips that go into mobile devices, automotive, and AI inference cards — were up over 30% sequentially and over 140% year-over-year. SSD controllers were down 5–10% quarter-over-quarter on normal seasonality, but up 45% year-over-year, with richer mix from PCIe Gen5 and early MonTitan enterprise contribution. Ferri and Boot Drive solutions — the small product line that goes into AI accelerator platforms — was up something like 205–210% sequentially and 755–760% year-over-year. That is not a typo, and it is the line item I would want to see again next quarter.

Worth pausing on the SSD controller line. A 5–10% sequential dip looks like a soft spot in a strong quarter, but in context it's actually the shallowest Q1 seasonal pullback this business has posted in years. Q1 2025 ran 10–15% down. Q1 2023 ran 20–25% down. The dip is getting milder every cycle — that's market-share gains plus the PCIe Gen5 mix shift doing real work inside what reads as a negative line item.

The operating leverage is the part the headline numbers undersell. GAAP net income grew from $19.5 million in Q1 2025 to $66.8 million in Q1 2026 — a 243% year-over-year jump on 105% revenue growth. Non-GAAP net income grew 165%. That is the part that should catch my eye as an accountant: revenue doubled, but income grew much faster. That tells me the model is not just growing — it is scaling. A controller-IP business with this gross margin profile and this kind of OpEx leverage is structurally a higher-multiple business than the old consensus model was building.

Then management guided Q2 revenue to $393–411 million, GAAP gross margin 48.5–49.5%, GAAP operating margin 19.8–21.1% (Non-GAAP: 21–22%, excluding $3.6–4.6M of SBC and dispute-related expenses). Another 15–20% sequential step-up on top of an already record quarter, with margins expanding into the back half.

The market's response was immediate. The stock closed Tuesday at $149.18, gapped open near $196 on Wednesday, and finished the regular session at $217.50, up 45.8% on the day. Thursday added another fractional gain to close at $218.78.

What makes this more than a beat-and-raise: pull up SIMO on any consensus-target screener immediately after the print and the math read as bearish.

The pre-print analyst stack had targets ranging from $95 to $250, with an average around $180 and a median around $167. The stock gapped to $217.50. The screener subtracts the stock price from the average target, divides, and prints a negative number. SIMO's "average analyst upside" became an average analyst downside.

A retail screener does not know that those targets were set against a $149 stock and an old earnings model. It just shows red. And there is a portion of the retail audience that will look at that red, decide the smart money is leaning bearish, and either trim or stay out.

That gap — between what stale targets show and what the new earnings power actually supports — is the inefficiency. The reason it persists is not complicated. A big earnings surprise hits the price in an afternoon. The sell-side model does not. Analysts have to rebuild revenue, margins, out-years, valuation framework, and committee-approved language. That takes time. Consensus moves in batches.

You can watch the batches forming on this name right now.

By Thursday's close, three of the four stale targets I was watching had refiled. The cluster is forming faster than I expected:

That is one fresh initiation, five raises, and one reiteration in three trading sessions. Hosseini at $275 is now the highest target on the screen. The new average price target is $233.33, up from roughly $180 on Tuesday morning. Every move was bullish. None was against.

Craig Ellis at B. Riley deserves his own paragraph. He has covered SIMO with a near-perfect track record over the past two years — every published rating in profit, an average return per call comfortably above the stock's own move. That kind of number deserves an asterisk — a perfect record on a single name is partly a function of how the windows are timed against a stock that has trended up — but Ellis has been the most-followed institutional voice on this name, and his model has been close enough to right that the buy-side reads his notes when they print.

He filed today. $167 → $250, BUY reiterated.

The most-trusted analyst on this name just told the institutional desks the new earnings power supports a $250 stock. That's the signal that matters most. It just landed.

Two stale targets remain:

Both are sitting roughly $70–90 below current price on a stock where seven other analysts now agree the right number is somewhere between $180 and $275. The longer those two sit, the harder they are to defend. Either revision filing is a likely tightening of the cluster.

Silicon Motion designs the controller chips — the silicon that sits between the host system and the NAND flash — for SSDs and embedded storage devices. Every SSD has a controller. Every embedded UFS module in a phone, car, or AI inference card has a controller. SIMO is the largest merchant supplier in that category outside the captive controllers built by Samsung, SK Hynix, and Kioxia.

The AI capex cycle has been good for HBM (Micron, SK Hynix), good for the GPU itself (Nvidia), and good for advanced packaging (TSMC). It is now starting to be good for storage. AI training and inference workloads need enormous quantities of high-performance NVMe SSDs in the data center, and the growth in cloud AI deployments has pushed NAND prices up 55–60% sequentially in Q1 — the kind of move that signals a real supply-demand inflection rather than a normal cyclical rebound.

The hyperscaler capex prints this week sharpened that picture considerably. Alphabet raised 2026 capex to $180–190 billion and signaled 2027 would "significantly increase" again. Microsoft set 2026 capex at $190 billion — roughly $55 billion above the prior consensus — and said it expects to remain capacity-constrained through the year. Meta raised 2026 to $125–145 billion and explicitly called out higher memory component costs as a driver of the raise. Amazon held its $200 billion figure. Combined 2026 capex across the four largest hyperscalers is now tracking to roughly $725 billion, up about 77% from $410 billion in 2025. Meta's memory-pricing language in particular is the customer side of SIMO's pricing observation — the people writing the checks confirming what the supplier just reported.

SIMO's positioning into that wave shows up in the disclosed wins. MonTitan, the company's enterprise SSD controller, is in early volume production and ramping at five tier-one cloud service providers in 2H26. Boot drive solutions for a leading AI GPU maker are tracking to roughly $50 million this year, with a larger contribution in 2027. Engagement with a leading telecom infrastructure maker is expected to ramp late 2026, and a major search engine company is sampling for next-generation architecture. Management is using customer descriptions rather than names, so I would not overfit the tea leaves. But the categories matter: cloud service providers, AI accelerator platforms, telecom infrastructure, and search.

SIMO is not a primary AI beneficiary. It is a downstream beneficiary of the same capex cycle, sitting one or two layers deeper in the bill of materials, with the kind of margin profile (gross margins guided to 48.5–49.5%) that a controller IP business should command.

The same lag that hits price targets hits forward P/E screens too. Pull SIMO up on most stock-data sites today and the forward P/E reads in the high 30s — Yahoo currently shows around 38x. That looks expensive for a semiconductor name. The number is also wrong, and for the same reason the consensus PT screener is wrong: it's dividing the new stock price by the old EPS estimate.

The pre-print FY2026 EPS consensus was around $5.78. With Q1 actual at $1.58 and the Q2 guide implying roughly $2.00–2.10 in non-GAAP EPS, the company is already at about $3.65 in the first half before the back-half acceleration management is calling for. A reasonable post-print FY2026 EPS lands somewhere in the $7.50 to $9.00 range. That puts the real forward multiple at 24–29x, against SIMO's own 5-year historical mean of around 22x.

After a 46% gap, SIMO is now trading at a modest premium to its historical multiple, on a business that grew EPS 165% year-over-year and is re-rating into a higher-margin product mix. That is not a cheap stock. It is also not the 38x screener trap a casual buyer would walk away from. SIMO no longer trades on its old multiple, and it shouldn't — it isn't the same business. A buyer at $218 is paying for at least mid-2027 to look right.

This is not a name to chase as a normal breakout. The stock already repriced violently. The thesis is that the new earnings power supports a higher level than the chart suggests, not that the chart itself is clean after a one-day 46% move.

That means position size matters more than usual. A starter position makes more sense than pretending the risk reset did not happen. If JPM and BofA file upward in the next two weeks and the stock holds the gap, the setup gets stronger and you can add. If the remaining revisions arrive but the stock fades them, the market may have already pulled forward too much of the move and the trade is over before the cluster finishes.

Mean-reversion sellers will press the name across the next two weeks regardless of fundamentals, and a 5–10% pullback inside the holding window is not just possible but probable. Plan for it.

Entry: SIMO at $218.78, April 30, 2026.

Setup: The print broke the old model. The analyst stack is repricing in waves. The screeners haven't caught up.

Kill switch: The cluster stalls and the stock fades back toward $190 over the next three to four weeks.

Picks in Classical Investor are timestamped at the close of the publication day and tracked publicly against the S&P 500 from that entry. SIMO is logged at $218.78, close of April 30, 2026. The methodology prioritizes 30-day earnings and revenue revisions, revision breadth, estimate dispersion, and narrative-revision alignment over analyst price targets. Past picks and full track record are linked in the archive.

This is not investment advice. I am an accountant, not a registered advisor. Do your own work and size positions accordingly.