CEO says their biggest customer is 'not exactly known for being gentle' and guides to further pricing cuts while gross margin still expands.
Thesis: Cirrus is a clean cash machine with elite unit economics trapped inside an extreme concentration structure. The $50M beat and 49% FCF margin are real, but 94% single-customer exposure with acknowledged zero pricing power means the quarter's quality cannot be capitalized as a run rate. The PC franchise has concrete leading indicators (SDCA transition, design win funnel) but generates ~$40M against $2B revenue. The asymmetry favors waiting for PC content to reach materiality rather than buying the peak quarter.
Verdict: HOLD — Conviction: MEDIUM
Catalyst: PC revenue roughly doubling again in FY27 off mainstream platform ramps, with SDCA adoption reaching 50% and the AI voice codec sampling translating to design wins with 2x ASP. Management explicitly deferred a FY27 number.
Key Risk: When the flagship cycle rolls and 94% of revenue normalizes downward, the Q4 guide of $410-470M already embeds 20-29% sequential decline. CFO admitted 'we didn't call the peak.' Next peak miss could be to the downside.
The Tell: CFO Woolard told Tore Svanberg: 'we just were further away from the peak than we thought when we gave guidance last quarter... we just didn't call the peak.' Unguarded admission that management cannot accurately forecast where the flagship peak sits. Cuts both ways: this quarter was upside, the miss could be downside.
Friction Level: MODERATE_FRICTION — Both sides agree on the numbers: elite cash generation, clean balance sheet, 94% concentration, zero pricing power. Disagreement is whether SDCA and PC design wins (19 to 60+ programs) constitute a real re-rating catalyst or remain immaterial against a $2B base.
2026Q3 · Feb 3, 2026
3 of 5 readings agreed3 HOLD · 2 AVOIDRead the full call
Beat By $50M, 94% Single Customer, Zero Pricing Power
Cirrus is a clean cash machine with elite unit economics trapped inside an extreme concentration structure. The $50M beat and 49% FCF margin are real, but 94% single-customer exposure with acknowledged zero pricing power means the quarter's quality cannot be capitalized as a run rate. The PC franchise has concrete leading indicators (SDCA transition, design win funnel) but generates ~$40M against $2B revenue. The asymmetry favors waiting for PC content to reach materiality rather than buying the peak quarter.
CFO Woolard told Tore Svanberg: 'we just were further away from the peak than we thought when we gave guidance last quarter... we just didn't call the peak.' Unguarded admission that management cannot accurately forecast where the flagship peak sits. Cuts both ways: this quarter was upside, the miss could be downside.
Revenue of $580.6M came in $50M above top of guidance. Analyst Tore Svanberg explicitly stated 'those are pretty, pretty big beats' and noted the buy-side was not positioned for this magnitude of upside surprise.
Gross margin hit 53.1% despite year-over-year pricing reductions that management confirms are 'greater than' the 50bps compression reported. Cost reductions offset most pricing headwind, showing operational leverage against a structurally negative pricing environment.
60bps sequential margin improvement came primarily from 'reduction in inventory reserves' per CFO Woolard, an accounting benefit rather than pricing or mix improvement. Masks ongoing YoY pricing erosion from the flagship customer.
CEO Forsyth stated larger customers are 'not exactly known for being gentle in price negotiations' and explicitly guided to 'further pricing adjustments to our products' ahead. Confirms structural inability to defend ASP against customer leverage.
94% of revenue from one customer in a peak quarter, with CFO admitting 'we just didn't call the peak.' A peak-quarter beat being capitalized as a run rate creates expectation risk when the flagship cycle rolls.
Management addressed headwinds directly with specifics: Android decline quantified as biggest YoY contributor to general market weakness, EOL unwinding explained by ten-year-old parts on defunct nodes, PC metrics given with SDCA percentages and program counts.
+60
CONFIDENT"Management volunteered specific metrics unprompted: SDCA adoption from 15-20% to 50%, 75% socket win rate, 60+ programs by FY27, inventory days at 63. CFO gave direct seasonality answer without hedging. Voluntary headwind disclosure on Android and EOL products."
Rich transcript with specific unit metrics on PC penetration, design win counts, and program pipeline. Deflected specifics on largest customer relationship by policy but answered pricing and seasonality directly with numbers.
PC revenue roughly doubling again in FY27 off mainstream platform ramps, with SDCA adoption reaching 50% and the AI voice codec sampling translating to design wins with 2x ASP. Management explicitly deferred a FY27 number.
When the flagship cycle rolls and 94% of revenue normalizes downward, the Q4 guide of $410-470M already embeds 20-29% sequential decline. CFO admitted 'we didn't call the peak.' Next peak miss could be to the downside.
“In Q3 fiscal 2026, we delivered revenue of $580.6 million, which was above the top end of our guidance range driven by demand for components shipping into smartphones and a favorable mix of end devices.”
“For Q4 fiscal 2026, we expect revenue in the range of $410 million to $470 million.”
Figures are extracted from the transcripts and checked in code: every quote appears word for word. Descriptive only: none of this feeds the verdict.
Beat Above Buy-Side Whispers quarter: $580.6M revenue $50M over guide, record non-GAAP EPS of $2.97, 49% FCF margin, zero debt, $1.08B cash, inventory at 63 days. Gross margin expanded 60bps sequentially while absorbing YoY pricing cuts, showing the cost engine offsets ASP decline. The PC franchise has hard leading indicators: SDCA adoption going 15-20% to 50% by calendar year end, 75% socket win rate, programs scaling from 19 in FY25 to 60+ in FY27, top six laptop vendors already shipping. The AI voice codec doubles ASP and ships calendar '27/'28. Market prices this as a single-customer audio supplier while management builds a PC business that doubles two years running with rising content per socket.
The beat is entirely a peak-quarter artifact from 94% single-customer concentration, and management admits it: 'we didn't call the peak.' Q4 guide of $410-470M embeds 20-29% sequential decline. The 60bps margin gain came from inventory reserve release, not operational improvement. General market revenue declined both sequentially and YoY on Android exit and EOL unwinding. CEO acknowledges zero pricing power: customers are 'not exactly known for being gentle' and 'further pricing adjustments' are expected. PC revenue at ~$40M against $2B base is immaterial. AI codec ships calendar '27/'28. Automotive has no revenue target. Growth optionality is dated years out while the core faces structural ASP erosion.
Mix Shift Spin
60bps sequential margin gain attributed to 'reduction in inventory reserve' per CFO. Accounting benefit masks that YoY pricing pressure is 'greater than' the 50bps reported contraction. Real pricing erosion exceeds reported margin decline.
Unsustainable Trend Confidence
94% single-customer revenue in peak quarter with CFO admitting forecasting miss on peak location. When flagship rolls, Q4 guide already shows 20-29% sequential decline. Peak-quarter beat risks being capitalized as run rate.
AI Label Without Substance
AI voice codec is 'sampling' with revenue landing calendar '27/'28 against $2B annual base. Automotive SAM of $800M by 2029 has no revenue target or timeframe. AI narrative runs years ahead of financial substance.
Say/Do Gap
CEO commits to 'maintain gross margin' while simultaneously acknowledging customers are 'not exactly known for being gentle' and 'further pricing adjustments' ahead. Ability to offset pricing with cost reductions is the central unproven claim.
HOLD. The quarter is clean: 49% FCF margin, no debt, inventory contracting. But you cannot size a position when 94% of revenue depends on one customer's product cycle and management admits both zero pricing power and an inability to call the peak. The PC metrics are real leading indicators but at ~$40M against $2B they do not move the multiple yet. Wait for PC revenue to prove out in FY27 prints before sizing. This is a watchlist name with a specific trigger: two consecutive quarters of mainstream PC revenue doubling on schedule.