Microchip and Diodes also sell timing parts to the same buyers.
Is it a good business? · 6 · What does its record show?
Sales tripled from the low
Sales, $M
218.8
283.6
144.0
202.7
326.7
683.0
Dec 2021Dec 2022Dec 2023Dec 2024Dec 2025last 12 mo
Sales$219M → $683Msales, 2021 to the past year
2023−49%sales fall in 2023
Profit15% → 0.9%operating profit per $1
Costs per sales dollar have roughly halved since 2023, but 2023 showed the reverse when sales fell and costs stayed.
Is it a good business? · 7 · How much room is there to grow?
Optical modules for AI are the fast market
93%optical module market growth in 2024
65%forecast growth in 2026
$450MSiTime's own estimate of its 2027 slice
The slice is the seller's estimate; nobody counts timing parts per AI module.
Is it a good business? · 8 · Who is in charge?
A long-serving chief who owns little
2007Rajesh Vashist took charge
About 1%owned by insiders
Renesasnow a big owner
Two big fund managers, an early backer and now Renesas each own far more than management.
The story right now · 9 · Why does AI need more clocks?
AI links need precise clocks
3.6 timessales against a year earlier
65%optical module growth expected in 2026
139.9%share rise in a year
AI data centres link servers with fast optical modules, and each needs a precise clock. Faster links need better clocks, which suits silicon.
The story right now · 10 · Could the orders be stockpiling?
Buyers ordered far ahead once before
A shortage hitsBuyers order ahead.
Orders are cancellableAnd placed far out.
Stock piles upAs in 2021.
Sales fall−49% in 2023.
Own-product sales rising every quarter through 2027, with profit per dollar holding, would prove us wrong.
The story right now · 11 · What changed lately?
Sales beat; deal write-downs ate the profit
$298.7Msales, 3.6 times a year earlier
$129.5Madjusted operating profit
$8.3Mprofit under standard accounting
$331M to $341MDecember-quarter sales forecast
Strong quarter
$61.8M of write-downs of bought assets cut profit, and the release does not split SiTime's parts from the Renesas line.
The story right now · 12 · What is everyone arguing about?
A scarce AI winner, or stockpiled orders?
Many investors think
A scarce AI winner replacing quartz
We think
A real franchise whose orders may be stockpiling
In 2021 buyers ordered far ahead in a shortage, and in 2023 sales fell 49%.
The price and our view · 13 · What is a share worth?
The price sits above our best case
$225what we think a share is worth · the last close is of that
$225our value
Attractive below $180
Neutral
Unattractive above $271
$16.33gloomy: orders unwind as in 2023
$225our estimate: experts' path, then slowing
$410sunny: 2027 as forecast, 2028 up 35%
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 · How could we be wrong?
Stockpiling and a single wafer maker
Stockpiling, againBiggest
Cancellable orders.
Owners' share shrinksShowing now
New shares and deal costs.
One wafer makerWatch
Bosch renewal.
Parts per AI moduleUnknown
Nobody counts them.
Few distributors and the bought factories add to the risk.
The price and our view · 15 · How will we know?
February shows whether orders are real demand
Bosch agreement renewedGood sign
Own sales rise every quarterGood sign
Orders are cancelledBad sign
FEB 202728
Fourth-quarter 2026 results, 28 February. Companies report every three months; each report is a test of the story.
Our view also changes on price alone: if the shares fall to within a fifth of our estimate, we move to Neutral.
The price and our view · 16 · So, what do we think?
The silicon timing leader, priced beyond our best case
$225our value
Attractive below $180
Neutral
Unattractive above $271
UnattractiveMedium conviction
SiTime's timing chips are hard to replace once chosen, and sales more than doubled on AI networking. But orders are cancellable and placed far ahead, one wafer maker makes most parts and new shares shrink owners' slice.
↑ ↓ to move · C for contents
What it does
More about the company
Who it is
Every chip needs a clock: a part that ticks at an exact rate so data moves in step. For decades that part was a sliver of quartz crystal. SiTime makes it from silicon instead, using tiny vibrating structures called MEMS, which tolerate heat, shaking and small spaces better.
It designs the chips and pays others to make them: Bosch and Teledyne make the silicon parts, TSMC and UMC the circuits. Buyers are makers of optical network modules, phones, cars and factory gear, mostly through distributors.
How it makes money
SiTime sells parts by the piece. A buyer designs a SiTime clock into a product, then orders it for every unit built, for years. Most orders go through three distributors, which took 66% of sales in the June quarter, and buyers can cancel them with little notice.
Data-centre and network parts were $101M of June's $157.4M. Each sales dollar left about 67 cents after the cost of making the part, on the company's adjusted measure.
Growth and profit
What growth costs
Growth is making SiTime more profitable again: since the 2023 slump, sales have tripled while research and selling costs grew far less. But those costs are still a bigger slice of each sales dollar than in 2021, and a sales drop would reverse the gain, as 2023 showed.
Its record
The numbers
Is it growing?Fast
Sales grew 143% over the twelve months to Sep 2026, faster than the 61% in the year to Dec 2025.
Is it profitable?Thin
Operating margin is 0.9%, up from −74% in the year to Dec 2023.
Does growth pay off?Pays off
Running costs take 55¢ of each sales dollar, up from 49¢ in the year to Dec 2021.
Does it earn good returns?Low
Each dollar put into the business earns 0.2% a year after tax, down from 8.8% in the year to Dec 2021.
Is it making cash?Solid
It turned 6.1% of sales into free cash, $41.4M over the last twelve months.
How steady is it?Swingy
Yearly sales growth has ranged from −49% to +88% over the last 5 years, with sales falling in 1 of them.
Are shares being diluted?Diluting
The share count rose 47% in 4.8 years, and pay in shares costs 18% of sales.
Why it wins
The evidence
Choosing a clock takes a buyer six months to three years of design and testing. Once chosen, it stays for the product's life, often ten years, because swapping it means testing the whole board again. Car makers must formally requalify a part whenever its supplier changes it.
So SiTime wins share when new products are designed, not by cutting prices. The proof is that its prices rose: the cut it keeps from each sales dollar went from 58 cents to 67 cents in a year.
Who could beat it
The rivals in full
Quartz makers in Taiwan and Japan, such as TXC and Seiko Epson, still sell most of the world's timing parts. The newer threat is Texas Instruments, which sells a non-quartz clock using a different vibrating technology aimed at the same high-speed sockets. Skyworks, Diodes and Microchip also sell timing, though for each it is a small side line.
Each rival
TXC, Seiko Epson and Nihon Dempa Kogyo quartz oscillators3042.TW6724.T6779.T
the reference oscillator in an optical module, switch or server board, where quartz is the incumbent part
the low-jitter oscillator in a networking board, a silicon alternative to MEMS from a far larger chipmaker
Skyworks timing (the former Silicon Labs clock and oscillator line)SWKS
the clock generators and jitter cleaners on a data-centre board, the acquired business's main sockets
Diodes and Microchip crystals, oscillators and timing chipsDIODMCHP
standard oscillators and clock buffers in industrial and car boards, sold on price
How big it could get
The market and who has signed
The market
The fastest-growing market is optical modules, the parts that carry data between AI servers as light. An industry research firm, LightCounting, measured that market up 93% in 2024 and forecasts 65% growth in 2026.
Who has signed
Almost nothing is signed. Apple, 17% of 2025 sales, buys under an agreement that sets terms but promises no volume. Everyone else orders piece by piece on orders they can cancel, and a deal to put SiTime's parts into Renesas's chips is still only being explored.
Who runs it
Leaders, pay and the cash
Who runs it
Rajesh Vashist has run SiTime since 2007 and Elizabeth Howe its finances since 2023. Insiders own only about 1%; two big fund managers, an early backer and now Renesas each own far more. Most of the yearly bonus rides on sales growth and the performance shares on beating other chip stocks, while nothing pays on profit or on how much each share earns.
What they did with the money
Management bought Renesas's timing business for about $4.0B, paying $1.50B in cash and 3.56M new shares. To fund it, SiTime borrowed $1.35B through bonds that pay no interest but can turn into shares above $1,040 each.
The last report
The latest report
SiTime's first quarter with the Renesas business beat its forecast
What it did to the thesis Strong quarter
It passed the sales and margin parts of our test and the December outlook is higher still. Whether orders are real demand shows in February.
The report
In the September 2026 quarter, the first with Renesas's timing business, sales were $298.7M, above the $285M to $295M forecast and 3.6 times a year earlier. SiTime's own parts and the acquired line are not split out in the release.
It kept about 71 cents of each sales dollar after making the part, against 67 in June, and adjusted operating profit was $129.5M. Profit under standard accounting was only $8.3M, because of $61.8M of write-downs of acquired assets. On the call, SiTime guided December sales to $331M to $341M with margin near 70%, and said the Renesas timing line grew 67% on the year.
What's happening now
The background
The shares rose 139.9% in the year to 2 October, and more than half of that came from buyers paying more for each dollar of sales rather than from sales themselves. They have since fallen 28% with no news of the company.
SiTime reported its first quarter with the Renesas business on 4 November; the call guided December sales to $331M to $341M, and the shares have since risen to $472.56. Each share becomes three on 7 December. In February 2027 its wafer agreement with Bosch, the only maker of its silicon parts, comes up for renewal.
Where we disagree
Both sides in full
The market
SiTime is a scarce AI winner, replacing quartz in every network link, with Renesas making it the whole timing supplier.
We believe
The franchise is real and the price now allows for it. We do not yet know that the year-ahead orders are all real demand.
The argument
In 2021 buyers also ordered far ahead during a chip shortage. The shares then fell 73%, and in 2023 sales fell 49% as customers used up their stockpiles. Today's orders are again long and cancellable.
The evidence against us: own-product sales that keep rising every quarter through 2027 with the margin holding would show the demand is end use.
What it's worth
How we value it
A timing-chip maker doubling on AI orders and the Renesas line is valued on analysts' 2026 sales, then 45% and management's 30% a year, with a GAAP margin after stock pay and the deal's amortisation, slowing toward 4% by year ten, each year closing 30% of the gap.
The three cases
Bear case$16.33
The cancellable orders unwind as in 2023, when sales fell 49% (C37, C14): 2027 grows only 10% and 2028 is flat, the margin settles at 20% under the amortisation, and new capital earns only its 9% cost.
Our value$225
2026 lands on analysts' $875M (C30), SiTime's own sales more than doubling and the Renesas line adding about $85M a quarter from July (C07); 2027 grows 45%, short of analysts' $1,353M, and 2028 30%, management's yearly target (C28); the GAAP operating margin, after stock pay and about $180M a year of amortisation (C08), rises from 0.9% to 30%.
Bull case$410
2027 reaches analysts' $1,353M (C30) and 2028 grows 35%, above the 30% a year management targets for five to eight years (C28), and the GAAP margin settles at 38%, near today's non-GAAP 34% (C03) once the amortisation fades.
What could go wrong
Every risk
The risks
Stockpiling, again
Buyers order 12 to 18 months ahead on cancellable orders; if supply catches up they could stop ordering, as in 2023.
One wafer maker
Bosch makes most of its silicon parts under an agreement ending February 2027; a worse renewal would hit costs or supply.
Owners' share shrinks
Shares outstanding rose 27% in 18 months, and adjusted profit leaves out stock pay and about $230M a year of deal costs.
Few distributors
Three distributors took 66% of June-quarter sales, so one changing its stock levels moves the quarter.
The acquired factories
Renesas still makes the acquired parts for several quarters and the business is short of supply.
Insiders selling
The chief executive sold 12,000 shares outside a trading plan in September, and no insider bought in a year.
What we still do not know
How many timing parts sit in each AI network module or server rack, and at what price; nobody outside the company publishes it.
Whether Bosch renews the wafer agreement on the same terms in February 2027.
What the Renesas business earns on its own once it leaves Renesas's factories and shared costs.
How much of the year-ahead orders is real demand rather than buyers securing scarce parts.
What changes our view
Every route to a new view
Every route to a new view
→ Neutral
The share price falls to about $271, 20% above what we think the business is worth.
→ Attractive
The share price falls to about $180, 20% below what we think the business is worth.
The next test
It supports the case if
Revenue at or above $331M, non-GAAP gross margin at or above 67%, and inventory below $180M - the order book was end demand and the acquired business is holding.
It pressures the case if
Revenue below $298.7M, non-GAAP gross margin below 67%, or inventories rising above $180M while revenue falls.
We would also be wrong if
Sales keep growing above 30% a year through 2029 with the margin above 38%.
Texas Instruments' rival wins no large socket by the end of 2027.
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
2 of 4
Potential
2 of 4
Risk profile
3 of 4
Valuation
1 of 4
Evidence quality
3 of 4
How we weighed it
ValueNegative
The price sits above our value and above the bull case, even on a very steep path.
BusinessMixed
A mixed business: sticky chips, but new shares and deal costs dilute owners.
GrowthPositive
Sales more than doubled and the next outlook is higher.
RisksNegative
Buyers may be stockpiling, one wafer maker makes most parts and the share count rose.
RivalsMixed
No filed rival compares directly, so the price is judged on its own.
The groupMixed
AI chip names move together on spending news.
Latest newsPositive
The last report passed the sales and margin tests.
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.