They raised the 2026 forecast 50% since November while disclosing that backlog is weighted to Q1 and Q2.
Thesis: SITM owns pricing power in the oscillator bottleneck for 1.6T optical modules. ASPs up tens of percent on the node transition with unit growth layered on top is the textbook Beat-and-Raise Machine setup. Cash flow confirms reported earnings: $87.2M vs $23.2M prior year. The market is pricing concentration risk as if CED growth decelerates next quarter, but the 50% upward revision to 2026 forecasts since November says the opposite. Size it, but respect the short-dated backlog.
Verdict: LONG — Conviction: MEDIUM
Catalyst: Q1 print at or above the $101-104M guide with GM at 62% confirms book-to-bill converts. Any further upward revision to hyperscaler 1.6T forecasts through H1 2026.
Key Risk: Backlog duration is 'well within twelve months' and 'definitely weighted to Q1 and Q2.' If hyperscaler capex pauses in H2 2026, there is no longer-dated cushion. DSO at 36 days on a second consecutive quarter would confirm deterioration.
The Tell: Rajesh volunteered that 'many of them are booking in advance of real demand' then immediately walked it back: 'I don't mean they're ahead of it. I mean, they're on top of it.' The unprompted correction reveals awareness that the sentence reads as double-ordering.
Friction Level: MODERATE_FRICTION — Both sides agree the quarter is strong. Disagreement is whether short-dated backlog weighted to two quarters with 53% CED concentration is a timing quirk or structural fragility.
2025Q4 · Feb 4, 2026
3 of 5 readings agreed3 LONG · 2 HOLDRead the full call
Book-to-Bill 1.5, but Half of It Lands in Two Quarters
SITM owns pricing power in the oscillator bottleneck for 1.6T optical modules. ASPs up tens of percent on the node transition with unit growth layered on top is the textbook Beat-and-Raise Machine setup. Cash flow confirms reported earnings: $87.2M vs $23.2M prior year. The market is pricing concentration risk as if CED growth decelerates next quarter, but the 50% upward revision to 2026 forecasts since November says the opposite. Size it, but respect the short-dated backlog.
Rajesh volunteered that 'many of them are booking in advance of real demand' then immediately walked it back: 'I don't mean they're ahead of it. I mean, they're on top of it.' The unprompted correction reveals awareness that the sentence reads as double-ordering.
Q4 revenue $113.3M beat guidance with EPS tripling to $1.53. Q1 guide of $101-104M at ~70% YoY with GM stepping to 62% extends the streak. Management hit the exact 2025 exit margin target set twelve months prior.
1.6T transition drives ASP increases of 'tens of percent' per Rajesh. 800G demand continues growing while 1.6T layers on top. Customers have no alternative for programmable MEMS oscillators at these clocking frequencies.
GM progression 59.3% FY25, 61.2% Q4, guided 62% Q1 driven by CED mix shift. Renesas acquisition adds $300M at ~70% GM on top. Mix-driven expansion verified against prior guidance.
Beth walked through Q1 guidance with specific margin drivers, opex range, and share count. Rajesh addressed backlog duration directly when Jim Schneider asked. No deflection on concentration questions.
Book-to-bill over 1.5 but CFO disclosed backlog is 'well within twelve months' and 'definitely weighted to Q1 and Q2.' Short-duration, near-term-weighted orders in AI networking are a double-ordering signature. Not a fortress.
Physical AI TAM claims of $20 per robot, $15 per robotaxi, and L4 ADAS content have zero revenue attached. Consumer grew 7%, auto/defense/industrial grew 19% on $24.5M. The speculation is narrative padding.
'Exceptional' three times, 'spectacular' in closing, 'there is no stopping in the AI data center world.' No hedging detected. Management has not begun to see the turn coming.
+80
CONFIDENT"Rajesh volunteered book-to-bill unprompted: 'we wanted to give you a metric of the demand strength.' Beth gave specific numbers on margins, opex, share count, and backlog duration. Zero hedging language across the entire call."
Full transcript with specific segment revenue, margin drivers, DSO, inventory, acquisition financing terms, and backlog duration. Nearly every number traceable.
Q1 print at or above the $101-104M guide with GM at 62% confirms book-to-bill converts. Any further upward revision to hyperscaler 1.6T forecasts through H1 2026.
Backlog duration is 'well within twelve months' and 'definitely weighted to Q1 and Q2.' If hyperscaler capex pauses in H2 2026, there is no longer-dated cushion. DSO at 36 days on a second consecutive quarter would confirm deterioration.
“Revenue in Q4 was $113.3 million, up 66% year over year and 36% sequentially.”
“Thus, we project revenue in the range of $101 to $104 million, up roughly 70% year over year at the midpoint.”
Figures are extracted from the transcripts and checked in code: every quote appears word for word. Descriptive only: none of this feeds the verdict.
Q4 revenue $113.3M up 66% YoY with EPS tripling. GM crossed 61.2% against guided target, Q1 guides to 62%. Cash flow from operations $87.2M vs $23.2M confirms earnings quality. Inventory declined to $81.7M while revenue grew 36% sequential: clean sell-through, channel not stuffed. Book-to-bill above 1.5 with customers raising 2026 1.6T forecasts 50% since November. ASPs up tens of percent on 1.6T transition with 800G still growing. Pricing power is structural in programmable MEMS oscillators. Renesas adds $300M at 70% GM with minimal product overlap, cross-sell in both directions. CED at 53% of revenue was a deliberate plan from 12% at IPO, not accidental concentration.
Backlog is short-dated: 'well within twelve months' and 'definitely weighted to Q1 and Q2' per the CFO. Receivables jumped from 24 to 36 days DSO, a 50% sequential increase, dressed up as linearity normalization. CED concentration at 53% with growth entirely hyperscaler-capex dependent. Consumer grew 7%, auto/defense/industrial just $24.5M. The Renesas deal layers $900M of debt on $88M cash with 4.13M new shares on a 27M base, roughly 15% dilution. The acquired clock business grows slower than oscillators per Rajesh. Physical AI TAM claims of $15-20 per robot have no revenue behind them. Short-dated near-term-weighted backlog in AI networking is a textbook double-ordering signature.
DSO jumped 50% sequentially
DSO went from 24 days in Q3 to 36 days in Q4. Receivables at $45M against $113.3M quarterly revenue implies prior quarter AR was roughly $22M. Receivables likely doubled in one quarter while revenue grew 36% sequential. Management called it 'linearity returned to more normal.'
Short-dated near-term-weighted backlog
Beth Howe disclosed book-to-bill is 'well within twelve months' and 'definitely weighted to Q1 and Q2.' Voluntary disclosure in an unusual departure from prior practice. Short-duration AI networking orders are a double-ordering signature.
CED concentration at 53%
CED grew from 12% at IPO to 53% of revenue. Every growth claim traces to hyperscaler capex. Consumer segment grew 7%, auto/defense/industrial just $24.5M growing 19%. Remove AI capex and growth collapses toward 25-30%.
$900M debt on $88M cash
Renesas deal adds ~$900M committed Wells Fargo debt plus 4.13M shares (15% dilution). Management claims under 2x leverage within 24 months, implying pro forma EBITDA of $450M+ against a combined base that appears to be $150-200M.
LONG at MEDIUM conviction. The operating data is too clean to fade: margins expanding from real mix shift, cash flow confirming earnings, inventory declining into revenue acceleration, and a 50% upward forecast revision since November. The asymmetry favors long because the market is pricing concentration risk while the leading indicator (1.6T customer forecasts) moved sharply higher. Size for the short-dated backlog risk: add on the Q1 print, trim if DSO holds above 34 days for a second quarter or CED decelerates below 80% YoY.