Silicon Motion designs the chips that run flash memory
Controllers
the chip that runs flash storage
More than 6 billion
shipped
Phones and servers
use its chips
Raw flash memory wears out and returns errors. A controller chip spreads the writing, corrects errors and tracks where data lives. Silicon Motion designs them and sells to the memory makers.
The company · 2 · What exactly does it sell?
Controllers for drives and phones are most of it
Each $1 of sales, 2025
45%40%0%
SSD controllerseMMC and UFS controllersSSD solutions
In 2026 it also resold finished drives, flash bought and packaged, which earn less than controllers.
The company · 3 · How does it make money?
Paid per chip, and now per finished drive
29.0 centsoperating profit per sales dollar
38%of sales from one unnamed customer
$870.9Mof stock, from $673.0M
Controllers earn a design fee in each chip sold. Finished drives carry the flash itself, so they pay off in a shortage and hurt when flash prices fall.
Is it a good business? · 4 · What gives it an edge?
One software library tuned to every flash maker
More than 6 billion
controllers shipped
$262.7M
research a year
Its phone-storage controller sales roughly doubled in 2026 while phone production fell. It does not lock in flash makers, which can design their own.
The biggest flash makers stepped back by choice and could take the work back.
Is it a good business? · 6 · What does its record show?
Fixed costs cut both ways
Sales, $M
922.1
945.9
639.1
803.6
885.6
Dec 2021Dec 2022Dec 2023Dec 2024Dec 2025
Up11.2% → 22.4%operating profit per $1, 2026
Down-32%sales in 2023
Low6.2%profit per $1 that year
Profit per sales dollar doubled in a year as costs rose far slower than sales. In 2023 the same costs squeezed profit when sales fell.
Is it a good business? · 7 · How much room is there to grow?
Nobody counts the controller market
More than 40%of flash demand from servers
15% to 20%fall in 2026 phone production
2large enterprise customers in production, five more due
Nothing is under contract; customers buy on purchase orders.
Is it a good business? · 8 · Who is in charge?
The founder still runs it
1995Wallace Kou founded it
1.4%owned by Kou
$2.71Mtotal leaders' pay, 2025
Leaders' shares vest with time, not results. It is still fighting MaxLinear in arbitration.
The story right now · 9 · Why did flash prices soar?
AI servers soaked up flash, and prices jumped
More than 40%of flash demand from servers
Late 2027supply expected to balance
+94%the shares on the year
AI data centres store vast amounts of data on flash drives. Demand outran supply, prices rose, and anyone holding flash to resell earned more.
The story right now · 10 · Which part of the growth lasts?
Half is controllers, half is flash at shortage prices
Controller share risesReal and lasting.
Drives resold with flashAt shortage prices.
Stock piles upBought at rising prices.
Flash prices fallDrive profit fades.
We value it on a normal year's profit per dollar, not today's peak. We are wrong if profit holds once flash prices fall.
The story right now · 11 · What changed lately?
Sales rose 131% as stock and cash outflow grew
$560.2Mquarterly sales
29.0 centsoperating profit per sales dollar
$870.9Mstock, from $673.0M
-$256.1Mcash from operations
Mixed quarter
The growth is real, but stock piled up and profit did not turn into cash. The shares fell 18% in 20 sessions after results.
The story right now · 12 · What is everyone arguing about?
How much of the growth lasts?
Many investors think
Lasting share won, sales climbing to $3.36B by 2028
We think
Half is flash at shortage prices, fading before 2028
Finished drives earn less than controllers, yet profit per dollar rose as drives grew. We are wrong if it keeps 48 cents per $1 after flash prices fall.
The price and our view · 13 · What is a share worth?
The price sits about a tenth above our value
$163what we think a share is worth · the last close is of that
$163our value
Attractive below $131
Neutral
Unattractive above $196
$60.55gloomy: growth halves and profit per $1 falls
$163our estimate: a normal year's profit, not the peak
$260sunny: growth runs a quarter faster
Value per share in three futures. None is a prediction of the share price.
We estimate what a share is worth from the cash the business should earn over many years. Our rule: buy only when the price is at least a fifth below our estimate, as a cushion for being wrong.
The price and our view · 14 · What if it all works?
As a bet on the flash shortage, it is fragile
Speculationthe kind of bet this is
2.2×if it works
up to 29%kept if it fails
Multiples of the company's value when the bet was last struck, 16 Nov 2026.
The price already pays for much of the shortage lasting. If supply catches up, the controller business is left with stock written down.
The price and our view · 15 · How could we be wrong?
The shortage ends with stock bought at the top
The shortage endsBiggest
Supply balances in late 2027.
Stock bought at the topShowing now
Flash bought at rising prices.
A bond repayment in 2029Watch
$1.15B could be demanded back.
Who the big customer isUnknown
The buyer of 38% of sales.
China took 64% of billings, and the company is based in Taiwan.
The price and our view · 16 · How will we know?
February shows whether stock stops rising
Sales of $602M or more, 53.8 cents kept per $1Good sign
Stock no higher than $870.9M, cash turning positiveGood sign
48 cents kept per $1 after flash prices fallWe turn positive
FEB 20272
Fourth-quarter 2026 results, 2 February. Companies report every three months; each report is a test of the story.
Our view also changes on price alone: we turn positive if the shares fall a fifth below our estimate, and negative if they rise a fifth above it.
The price and our view · 17 · So, what do we think?
Real controller gains plus flash resold at shortage prices
$163our value
Attractive below $131
Neutral
Unattractive above $196
NeutralMedium conviction
Silicon Motion designs the chips that run flash memory, and sales more than doubled. But about half the growth is finished drives resold at shortage prices, and stock has piled up. Neutral.
↑ ↓ to move · C for contents
What it does
More about the company
Who it is
Raw flash memory is unreliable: cells wear out and return errors. A controller is the chip that hides this, spreading writes evenly, correcting errors and keeping track of where data lives. Silicon Motion, founded in Taiwan, designs them and has shipped more than 6 billion.
It sells to module makers that build phone storage and PC drives, and to flash makers such as Kioxia and Micron. Their alternative is to design their own, which the biggest flash makers already do for many products.
How it makes money
It is paid per chip or per drive shipped. In 2025 almost all sales were controllers: about half for PC and server drives, the rest for phone storage. By Q2 2026 near 30% came from finished drives, where most of the price is the flash inside. One customer, named only PHISEMI, bought 38% of everything in the first half of 2026, and China took 64% of billings.
Growth and profit
What growth costs
Growth has made it more profitable: research and office costs rose far more slowly than sales, so operating profit per sales dollar doubled in a year. The same fixed costs cut the other way in 2023, when sales fell 32% and profit nearly vanished.
Its record
The numbers
Is it growing?Steady
Sales grew 10% in the year to Dec 2025, slower than the 26% in the year before.
Is it profitable?Healthy
Operating margin is 11%, down from 23% in the year to Dec 2022.
Does growth pay off?Pays off
Running costs take 38¢ of each sales dollar, up from 23¢ in the year to Dec 2021.
Does it earn good returns?Decent
Each dollar put into the business earns 10% a year after tax, down from 33% in the year to Dec 2021.
Is it making cash?Thin
It turned 0.7% of sales into free cash, $6.29M over the year.
How steady is it?Swingy
Yearly sales growth has ranged from −32% to +71% over the last 5 years, with sales falling in 1 of them.
Are shares being diluted?Stable
The share count fell 3.8% in 4.0 years, and pay in shares costs 3.0% of sales.
Why it wins
The evidence
Its edge is scale in firmware: one library tuned to every major flash maker's memory, built over billions of chips and $262.7M of research a year. A module maker mixing flash from several makers can ship a working drive without its own controller team.
That edge showed in 2026, when phone-storage controller sales roughly doubled while phone production fell. But it does not lock in the flash makers, who design their own and stepped back by choice.
Who could beat it
The rivals in full
Phison, a Taiwanese rival, sells the same controllers to the same buyers. The bigger threat is the customers: Sandisk designs most of its controllers in-house and Micron its data-centre ones. Marvell and Microchip sell storage controllers too, as small parts of much larger companies.
Each rival
Phison Electronics, a Taiwanese controller and module designer the 20-F names first among rivals8299.TWO
the module or flash maker buys its controller, or a whole drive design, from Phison instead of Silicon Motion
The flash makers' own controllers: Samsung, SK hynix and Solidigm, Kioxia, Micron and Sandisk005930.KS000660.KS285A.TMUSNDK
the flash maker designs the controller for its own SSDs and phone storage instead of buying one, as Sandisk and Micron already do for most products
a server-drive maker buys its enterprise SSD controller from Microchip instead of MonTitan
The flash makers' own finished drives and boot-drive suppliers to GPU and DPU makersSNDKMU005930.KS
a server or car maker buys a finished boot drive or automotive drive from a flash maker or one of the two or three other qualified suppliers
How big it could get
The market and who has signed
The market
Nobody publishes how many controllers are sold or at what price, so the market cannot be counted directly. What is measured is the flash they sit beside: servers now take more than 40% of flash demand, phones and notebooks nearly 40%, and phone production falls 15% to 20% in 2026. TrendForce expects flash supply to stay short through 2026 and balance in late 2027.
Who has signed
Nothing is under contract. Customers buy on purchase orders. Its enterprise controllers are in production with two large customers, with five more due, and its server boot drives are one of two or three suppliers for each chip maker's programme. Neither comes with a disclosed volume.
Who runs it
Leaders, pay and the cash
Who runs it
Wallace Kou founded Silicon Motion in 1995 and still runs it; he owns 1.4% of the company and the chairman 1.2%, so the managers own a stake but not control. Jason Tsai has been chief financial officer since 2024. The company discloses only total pay for its leaders, $2.71M in 2025, and their shares vest with time, not with results. It is still fighting MaxLinear in arbitration for a $160M fee after that merger collapsed in 2023.
What they did with the money
In August it borrowed $1.15B at no interest through bonds that turn into shares only if the price passes about $380. The chief executive said the money secures the components for its growth, meaning flash. It also pays a dividend and buys back a few shares.
The last report
The latest report
Silicon Motion's sales rose 131% to $560.2M while inventory and cash outflow grew
What it did to the thesis Mixed quarter
Sales and margin beat our lines, so the growth is real. But stock piled up and the profit did not turn into cash, so the open question, whether the margin outlasts the flash shortage, is not settled and the view stays neutral.
The report
But inventory rose to $870.9M from $673.0M and cash from operations was minus $256.1M. Management guided the next quarter to $588M to $616M.
What's happening now
The background
The shares nearly doubled in a year, and they swing hard on any sign the flash shortage is easing. They fell from a June high when TrendForce said price rises were slowing, then jumped after the July results.
The September-quarter results came on 29 October and the shares fell 18% in the 20 sessions to 11 November, though they are still up 94% on the year. The next test is the December-quarter results in early February: whether inventory stops rising and cash from operations turns positive. TrendForce still sees supply catching up in the second half of 2027.
Where we disagree
Both sides in full
The market
Silicon Motion has won lasting share from flash makers, and sales will climb from $1.96B this year to $3.36B by 2028.
We believe
Half the growth is lasting share; half is flash resold at shortage prices, which fades when supply catches up, likely before 2028.
The argument
The company's own annual report says finished drives earn less than controllers, because most of their price is flash it cannot mark up. Yet margins rose as drives grew to near 30% of sales. That only adds up if the drives are earning shortage prices on flash bought earlier.
We would be wrong if margins stay at 48% or more after flash prices start falling.
What it's worth
How we value it
Valued on the analysts' revenue path for three years, each later year closing 30% of the gap to 4%, on official operating profit with stock pay a cost and tax paid in full; the thesis's own evidence is the check on that path.
What the price assumes
Sales growth in the first year, then slowing as ours does, at the same profit per sale.
The price assumed
49.5%
Our estimate
60%
Measured at the close of
6 Nov 2026
The three cases
Bear case$60.55
Growth runs at half the analysts' path, 30%, 20% and 11%, the margin settles a fifth lower at 12.4%, and new capital earns only the 8% rate: the downturn test.
Our value$163
Revenue grows 60%, 41% and 22% over three years, the analysts' path (C56), held to at most 60% a year, then fades toward 4%, closing 30% of the gap each year; the margin settles at its five-year average of 15.5%, not today's cycle peak.
Bull case$260
Growth runs a quarter above the analysts' path, 75%, 51% and 27%, the margin reaches 17.0%, and new capital earns 15%.
The bet
Seen as a bet
Even if it works, the shares look expensive. The best case we can defend is the sales analysts expect in 2028 at the margin the company now guides to. Valued as a chip maker at the top of a cycle, that is only a little above today's value.
If it fails, flash supply catches up, buyers work through their stock as in 2023, and shareholders keep a business earning its 2025 profit plus its stakes in two customers: a small fraction of today's value.
What could go wrong
Every risk
The risks
The shortage ends
TrendForce expects flash supply to balance in late 2027; drive sales and their margins would fall with flash prices.
Inventory bought at the top
It holds $673.0M of inventory, mostly flash bought at rising prices, and borrowed $1.15B to buy more; falling prices would force write-downs.
One unnamed customer
PHISEMI took 38% of first-half 2026 sales, and the filings say nothing more about who it is.
Flash makers take work back
The biggest flash makers design their own controllers and could stop outsourcing once their factories have spare room.
China and Taiwan
China took 64% of billings and the company is based in Taiwan, so trade rules or tension there hit it directly.
A bond put in 2029
Bondholders can demand $1.15B back in August 2029, which could land in a downturn.
What we still do not know
Who PHISEMI is, the customer behind 38% of sales.
How much of drive sales are server boot drives and how much car drives, and what each earns.
How much of the $1.15B goes into flash, and at what prices.
Which buyer is behind a jump in sales billed to Israel, most likely NVIDIA's server chips.
What changes our view
Every route to a new view
Every route to a new view
→ Attractive
The share price falls to about $131, 80% of what we think the business is worth.
→ Unattractive
The share price rises to about $196, 20% above what we think the business is worth.
→ Attractive
Gross margin stays at 48% or more in the first two quarters after flash prices start falling. Sales of phone-storage controllers keep growing on a year earlier in those same quarters.
→ Unattractive
Sales fall from one quarter to the next while inventory stays above $673M. Gross margin falls below 46%.
The next test
It supports the case if
Revenue of $602M or more, GAAP gross margin of 53.8% or more, inventory at or below $870.9M and positive cash from operations.
It pressures the case if
Revenue below $588M, gross margin below 50%, or inventory above $870.9M with cash from operations negative a third quarter.
We would also be wrong if
A flash maker says it is taking controller work back in-house.
Enterprise controllers and boot drives reach a fifth of sales with named buyers.
The verdict
Scores and how we weighed it
The scores
Growth
3 of 4
Business quality
3 of 4
Financial position
3 of 4
Execution record
3 of 4
Per share capture
3 of 4
Potential
2 of 4
Risk profile
3 of 4
Valuation
3 of 4
Evidence quality
3 of 4
How we weighed it
ValueMixed
The price is a little above our value, inside the margin, with the bear case far below.
BusinessMixed
Controller chips for flash memory, but half the growth rides on shortage prices.
GrowthMixed
Growth is strong now and the value assumes much of it lasts.
RisksNegative
The shortage ends, inventory was bought at the top, one unnamed customer dominates and flash makers may take work back.
RivalsMixed
There is no filed cohort to compare, so the price is judged on its own.
The groupNegative
Memory names move together on the shortage; its margin is the group's cycle.
Latest newsMixed
Sales and margin beat, but stock piled up and profit did not turn into cash.
In plain words
Our estimate of value
We forecast how much cash the business can earn over the next ten years and beyond, then ask what that stream is worth today to an investor who wants an 8% yearly return. We do it three times: a gloomy future, an expected one and a sunny one. We call a share Attractive only when its price is at least 20% below the expected estimate, and Unattractive when it is 20% above.