SiTime SITM

At a glance

One-line thesis. SiTime is the MEMS silicon-timing specialist whose precision clocks are riding the AI data-center build-out — revenue grew 61% in FY25, Q1 CY26 was up 88% YoY, and the Street pencils +84% for CY26 — but the stock trades at ~41× trailing EV/sales on a business that is still GAAP-unprofitable (stock-comp is ~29% of revenue), the CEO sold ~$43M of stock in June, the chart has broken below its 50-DMA, and the full-year estimate requires a second-half ramp roughly 50% above the quarter just reported — so this is a Watch: a name we want, at a price (~$430–450) that pays us for the execution risk.

1. What it is

SiTime Corporation (Nasdaq: SITM) develops, manufactures, and distributes silicon-MEMS-based precision timing solutions — resonators, clock ICs, and a broad family of oscillators — that replace legacy quartz timing devices. End markets span communications and enterprise/data-center infrastructure, automotive, industrial, IoT, mobile, consumer electronics, and aerospace & defense; product is sold through a channel-partner network of distributors and resellers. Founded 2003, IPO 2019-11-20, headquartered in Santa Clara, CA; CEO Rajesh Vashist; just 395 full-time employees — a remarkably lean, fabless model (~$380M TTM revenue ≈ $960K per employee).

Segment & geography honesty check: FMP reports a single "Reportable Segment" ($326.7M FY25, $202.7M FY24) with no product-line or geographic revenue split in the data pull (seg_geo is empty). The company profile notes operations spanning Taiwan, Hong Kong, and the United States — a supply-chain and shipment footprint concentrated in Asia that we cannot quantify from this dataset, so we flag it rather than guess.

The strategic story: precision timing is a small-dollar, high-criticality socket — and AI data centers multiply the number of clocks per system. That is what the quarterly acceleration (§5) is showing.

2. The expert thesis (traceable)

No expert-panel coverage — this note is fundamentals-driven. A search of the Synthos knowledge base returns zero claims on SITM (0 voices, 0 claims). This name entered via the quant momentum screen (+194% 12-mo vs SPY +21%), not via conviction-track expert voices, and the honest house standard applies: there is no high-skill thinker underwriting this thesis, no countervailing short case in the KB, and conviction is therefore Low by construction. Everything below is built from filed financials, live analyst estimates, and technicals in the FMP data pull.

What the Street says (context, not conviction): 9 Buy / 0 Hold / 0 Sell, consensus target $664.17 (median $650, high $900, low $450). A unanimous-Buy tape on a momentum name deserves skepticism, not comfort — nobody is paid to be cautious here. FMP's own quant rating is a C (overall 2/5), with P/E, P/B, ROE, and ROA all scoring 1/5 — the machine sees exactly the expensive-and-unprofitable profile we score below.

3. Synthos scores & the Bull / Base / Bear cases

The one-glance judgment — three scores, 0–10, each anchored to real metrics:

The three cases (our own scenario model — assumptions labeled; each target a ~12–18-month fair value; no false-precision probability blend). All EPS figures are the consensus non-GAAP EPS line (see the §5 data caveat).

Synthos fair value = the base case, ~$640 (+4%), full range $430–$900. Our base sits just below the Street's $664 consensus — anchored on the same estimate tape, with a haircut for its thinness (2–5 analysts per year). The honest signal is the shape: only ~4% base-case upside against ~30% bear-case downside — asymmetry pointing the wrong way at today's price, which is exactly why this is a Watch and the trigger sits at ~$430–450. This is a tracked call — the Forecaster Scorecard grades it once it matures.

4. Exponential Potential

Synthos separates compounders from exponentials (accelerating, multi-baggers-from-here). SITM has the exponential shape — with a cyclicality asterisk few exponentials carry:

• Forward growth: revenue CY26E→CY28E ~32% CAGR ($601.9M → $796.0M → $1,052.5M); consensus non-GAAP EPS $7.84 → $10.81 → $14.68 (~37% CAGR).
• Acceleration (the 2nd derivative): positive into 2026 — +40.8% (FY24) → +61.2% (FY25) → +84.3% (CY26E) — then the price action decelerates to ~+32.3% (CY27E) and ~+32.2% (CY28E). Quarterly revenue: $60.3M → $69.5M → $83.6M → $113.3M → $113.6M — note Q1 CY26 was flat sequentially; the acceleration must resume in H2 for the year to work.
• Room to run: at $16.2B, SITM is ~1/13th the size of the MRVL-class AI-silicon names; if precision timing content per AI rack keeps multiplying, a doubling is not capped by market size. The binding constraint is demand cyclicality, not TAM.
• The asterisk: this exact business printed $283.6M (FY22) → $144.0M (FY23), a −49% collapse, when the post-COVID inventory cycle turned. The current acceleration is real; so is the historical proof that it can violently reverse.

Exponential Potential: Very High (8/10) — genuine acceleration at a small cap, scored one notch below what the raw growth would suggest because the estimates themselves flag post-2026 deceleration and the 2023 drawdown proves the downside case is not hypothetical.

5. Financials (real numbers — FMP annual/quarterly)

• Revenue: FY25 (CY) $326.7M, +61.2% (FY24 $202.7M, +40.8%; FY23 $144.0M, −49.2%; FY22 $283.6M). The five-year arc is boom (FY21 +88%) → bust (FY23) → bigger boom — this is a deeply cyclical grower, now at a new high.
• Quarterly trajectory: Q1'25 $60.3M → Q2 $69.5M → Q3 $83.6M → Q4 $113.3M → Q1'26 $113.6M (+88.3% YoY, ~flat QoQ). Guidance-implied next step: Street models Q2'26 at ~$146.5M — and the full-year $601.9M consensus then requires H2 ≈ $342M (~$171M/qtr average), a ~50% step-up from the quarter just printed. That embedded ramp is the thesis risk.
• Margins: gross 55.7% TTM (FY25 53.6%, up from FY24's 51.6%). But GAAP operating margin is still negative: FY25 operating income −$60.4M (−18.5%), Q1'26 −$4.7M — R&D alone is 32.0% of TTM revenue and SG&A 33.8%.
• Earnings quality — read carefully. FY25 GAAP net loss −$42.9M (EPS −$1.72); TTM net margin −6.4%. Yet the Street's tracked (non-GAAP) EPS prints positive and beating: Q1'26 $1.44 actual vs $1.16 est; the last five prints all beat ($0.26/$0.47/$0.87/$1.53/$1.44 vs $0.11/$0.29/$0.71/$1.20/$1.16). The bridge is mostly stock-comp ($103.5M FY25 — 31.7% of revenue) plus $40.2M of D&A. FMP's income-quality metric (−4.3) is distorted by the small GAAP loss; the honest statement is: cash flow is real but modest, GAAP profits don't exist yet, and the non-GAAP story requires ignoring very large stock pay.
• Data caveat (estimates): FMP's 2026–2028 estimate rows are internally inconsistent — negative EBITDA/EBIT averages alongside strongly positive net income/EPS. The EPS line ($7.84 / $10.81 / $14.68) reconciles with the non-GAAP EPS the beat history tracks, so we use EPS and revenue and discard the estimate EBITDA/EBIT lines as an aggregation artifact. Coverage is thin: 2–5 analysts per year.
• Cash flow: FY25 operating CF $87.2M, capex −$52.5M (13.0% of TTM revenue — heavy for "fabless," reflecting capacity build), FCF $34.7M → FCF yield ~0.3% at this cap.
• Balance sheet: cash + short-term investments $808.4M vs total debt $4.7M (finance leases) → net cash ~$804M (~5% of the market cap); equity $1.16B; goodwill + intangibles $234.5M (18% of assets); current ratio 12.5. Funded, however, by a $451.6M common-stock issuance in 2025 — weighted shares grew from 23.1M (FY24) to 25.0M (FY25) to 26.3M (Q1'26), ~14% dilution in five quarters.

6. Valuation — priced in or room?

There is no trailing-earnings anchor: TTM GAAP EPS is negative, so the trailing P/E is meaningless (FMP prints it as −615×). What you actually pay: EV/sales 41.5× trailing (~26× CY26E), price/sales 42.8×, P/B 14.0×, P/FCF ~301×, FCF yield 0.33%. On the consensus non-GAAP EPS tape the forward P/E runs ~79× (CY26E $7.84) → ~57× (CY27E $10.81) → ~42× (CY28E $14.68) — the multiple roughly halves in two years at a flat price if estimates land. FMP's letter rating is a blunt C (2/5; P/E, P/B, ROE, ROA all 1/5). Street targets (context): consensus $664.17, median $650, high $900, low $450 — a 2.0× high-to-low band on a unanimous 9-Buy tape. The reverse read: $615 today requires the ~50% H2 revenue step-up to land and the market to keep paying 40×+ sales-multiple-equivalent for a decelerating-in-2027 growth curve. Not a value entry by any test we run; the price that changes the math is the ~$430–450 zone (~44× a bear-shaded CY27E EPS, coincident with the Street-low target and the rising 200-DMA).

7. Technicals (from the tech block)

• Trend: cracked in the near term, intact on the long lens. $615.46 is below the 50-DMA ($689) but far above the 200-DMA ($432); the 50 remains well above the 200 (no death cross).
• Location: −31.7% off the 52-week high ($901.48) — that is also the max drawdown from peak — and +223.7% off the 52-week low ($190.16). This correction is already deep; this is not a fresh-highs momentum entry.
• Momentum: RSI(14) 37.7 — weak-neutral, approaching but not yet at the oversold 30 line. MACD −6.4 (negative and confirming the pullback).
• Relative strength (the screen's origin): +194.0% 12-mo vs SPY +21.1% / QQQ +31.2%; +77.3% 3-mo vs SPY +14.6% / QQQ +23.6%. Enormous leadership — which also means an enormous amount of good news is banked.
• Read: technicals are deteriorating within an uptrend — a high-beta leader mid-correction, below its 50-DMA with negative MACD. For a Watch name, that is confirmation to wait: the lower-risk zone is a stabilization near the rising 200-DMA (~$432), which happens to bracket our bear-case fair value ($430) and the Street-low target ($450). Price and value converge there; they have not converged here.

8. Moat & competitive position

SiTime's moat is architectural: silicon-MEMS timing offers programmability, small size, and shock/vibration/thermal resilience that legacy quartz struggles to match, and SITM is the category's dedicated pure play — 395 employees generating ~$380M TTM revenue, with R&D at 32% of revenue compounding the device lead. Timing sockets are small-dollar but design-in sticky: once a clock tree is qualified, it rarely changes mid-generation. Limits of the moat, honestly: timing is a component market with giant adjacent players (broadline analog and quartz incumbents) that can price aggressively at the commodity end; SITM's premium depends on staying at the precision frontier; and negative ROIC (−3.3%) says the moat is not yet producing returns on capital — the profit proof is still pending.

Peer set (FMP-supplied, market cap): Bullish $4.0B, Cirrus Logic $7.5B, Descartes $6.3B, Klaviyo $5.1B, Universal Display $3.8B, Onto Innovation $15.1B, Qorvo $7.7B, Sanmina $11.7B, Semtech $12.9B, SoundHound $3.0B. This is a size-bucketed grab bag, not a comp set — the true timing/frequency-control comparators are not in the supplied list, so judge SITM against precision-timing and AI-infrastructure component names, not this cohort. Within the list shown, SITM carries by far the richest sales multiple.

9. Management, capital allocation & guidance

• Capital allocation: growth-everything — no dividend, no buyback; FY25 saw a $451.6M equity raise, $52.5M of capex, and $103.5M of stock-comp. Raising equity near highs is rational treasury management, but the shareholder experience is ~14% dilution in five quarters while GAAP profits remain negative.
• Insider activity — the yellow flag of this note. June 2026 Form 4s are all sales, zero buys: CEO Rajesh Vashist sold 40,000 shares on 06-03 at $701.13 and another 20,000 on 06-15 at ~$750 (~$43M total); director Raman Chitkara 2,000 @ $727.38; officer Vincent Pangrazio 2,000 @ $725.32; officer Lionel Bonnot 1,500 @ $700–736. Some may be 10b5-1 diversification — the CEO still holds ~388K direct shares — but a synchronized officer/director cluster within ~15% of the all-time high is a tell we do not ignore on a 42×-sales stock.
• Execution track record (the offset): five consecutive quarterly beats on both EPS and revenue (e.g., Q1'26 revenue $113.6M vs $103.5M est). Management under-promises and over-delivers on the prints. No management guidance claims are in our KB (SITM_mgmt does not exist), so we cannot quote or half-weight their forward book — noted as a coverage gap.

10. Catalysts & what to watch

• Next earnings: 2026-08-05 (Q2 CY26; Street EPS $1.93, revenue ~$146.5M, +110% YoY). The key line: does revenue re-accelerate sequentially (Q1 was flat QoQ) and does anything said about H2 support the embedded ~$171M/qtr second-half run-rate?
• AI data-center demand signals: hyperscaler capex commentary and AI-infrastructure order flow — the entire 2026 estimate leans on this end market staying hot.
• The H2 ramp math: CY26 consensus $601.9M minus H1 (~$260M actual+est) requires ~$342M in H2. Track it print by print.
• Insider-filing cadence: another CEO-scale sale cluster would upgrade the yellow flag; an open-market buy would partially retire it.
• Dilution: further equity issuance or SBC running above the ~29%-of-revenue TTM rate.

Thesis tripwires (what would change the call): upgrade trigger — price into the $430–450 zone with the growth tape intact, or a Q2 print that de-risks H2 (guide ≥ $160M for Q3); downgrade trigger — a sequential revenue decline, CY26 consensus cuts, gross margin rolling below ~52%, or a second insider-sale cluster.

11. Key risks

• Valuation / de-rating (dominant): ~41× trailing EV/sales, ~79× forward non-GAAP EPS, beta 2.91 — any estimate cut re-rates this violently; the stock is already −32% from its high with momentum negative.
• The embedded H2 ramp: full-year consensus requires H2 revenue ~50% above the Q1 run-rate; a miss converts the "accelerating exponential" into a broken-momentum story overnight.
• Proven cyclicality: 2023 revenue fell 49% in an inventory correction. Component businesses whipsaw; the current AI pull could over-order exactly the same way.
• GAAP unprofitability & SBC: stock-comp ~29% of revenue is the gap between the loss the accountants see and the EPS the Street cheers; that transfer is borne by holders via dilution (~14% in five quarters).
• Insider selling: ~$43M of CEO sales plus three other insiders in June, no buys.
• Concentration & data gaps: a single reportable segment, no geographic split in the data (Asia-centric operations per the profile), thin analyst coverage (2–5 estimates/yr), and a 395-person company with key-person exposure to its long-tenured CEO.
• Competition: precision timing invites well-capitalized analog/quartz incumbents if the socket economics keep improving; negative ROIC says the franchise hasn't yet proven pricing power in the accounting.

12. Verdict, position sizing & monitoring

Watch. SiTime is exactly what the momentum screen is for — a genuine, accelerating exponential (revenue +61% FY25, +88% YoY last quarter, +84% expected CY26) with a fortress net-cash balance sheet ($804M) and a five-quarter beat streak. But the entry math fails today: base-case fair value ~$640 offers +4% against a −30% bear case, the business is still GAAP-unprofitable with stock-comp at ~29% of revenue, the chart has broken its 50-DMA with negative MACD, the CEO just sold ~$43M of stock, and the full-year estimate embeds a second-half ramp ~50% above the current run-rate. We want the business; we are not paid to underwrite that stack of risks at $615.

• Sizing: no position at current levels. If already held, cap at ≤1% of the flagship and honor the §10 tripwires. The action zone is ~$430–450 (bear-case fair value / Street-low target / rising 200-DMA confluence) with the growth tape intact — there, the risk-reward inverts and this becomes a Tactical candidate.
• Monitoring: re-score at the 2026-08-05 print (H2 ramp evidence is the whole ballgame); track insider filings monthly. This verdict is logged as a tracked Synthos call as of 2026-07-06 at $615.46.
• Single biggest risk to the Watch call itself: a blowout Q2 that de-risks H2 could run the stock away from our trigger — the acceptable cost of demanding a margin of safety on a 2.91-beta, 42×-sales name.

Provenance & disclosures

• Traceability: 0 KB claims, 0 expert voices — a \bSITM\b search of the Synthos knowledge base returns nothing, so no expert thesis is cited and none is invented; conviction is Low by construction. This note is built entirely from the FMP data pull (filings-derived financials, live estimates, quotes, insider filings, technicals).
• Data as-of: fundamentals 2026-03-31 (Q1 CY26, filed 2026-05-07) · estimates & prices 2026-07-06 · insider filings through 2026-06-15. Forward figures are analyst consensus (FMP), labeled as estimates.
• Estimate caveat: FMP's 2026–2028 estimate rows show negative EBITDA/EBIT beside positive net income — internally inconsistent; we use the revenue and (non-GAAP-consistent) EPS lines and discard the estimate EBITDA/EBIT as an aggregation artifact. Analyst coverage is thin (2–5 estimates per year; only 2 for 2028 EPS).
• Earnings-quality caveat: TTM GAAP results are a net loss; all forward P/Es quoted use the consensus non-GAAP EPS tape. Stock-comp is ~29% of TTM revenue — the GAAP/non-GAAP bridge is dilution.
• Data gaps stated: no geographic or product-line revenue split in the pull (seg_geo empty; single reportable segment); no management-guidance claims in the KB; the FMP-supplied peer list is not a true timing-industry comp set.
• Not investment advice. Independent research, educational and informational only, never personalized. Hypothetical/forward figures are labeled; the only performance numbers Synthos will headline are the live, real-money Flagship's.
• Version: 2026-07-06. Prior versions available via the deep-dive version dropdown ("based on the info at the time").