Management guided free cash flow down 25-35% and announced a $500M buyback in the same breath.
Thesis: Qnity holds a real content-per-wafer position in CMP, advanced packaging and thermal materials, and the ICS 175bps margin expansion is volume-driven and disclosed. But 2026 upside is explicitly gated by customer capacity online at TSMC and the OSATs, and management anchored the midpoint to external MSI data at 6.5% while guiding free cash flow down 25-35% off a base flattered by one-time spin working capital release. Beat-and-raise mechanics are intact; the cash conversion story is not.
Verdict: HOLD — Conviction: MEDIUM
Catalyst: Customer advanced packaging and advanced node capacity coming online in the back half. Management: high end of range 'is really tied to capacity expansions at our customers and their ability to scale new node transitions and bring that capacity online.'
Key Risk: Memory demand destruction. Management named it: low end is 'more constraints really coming from the memory market and significant demand destruction.' Memory is 20% of semi portfolio.
The Tell: On the low end of guidance, management volunteered the exact downside scenario unprompted: 'more constraints really coming from the memory market and significant demand destruction.' They then immediately framed memory as 20% of semi and unit-driven with pricing out of the equation. Naming your own worst case in the same breath you discount it is the tell.
Friction Level: MODERICY_FRICTION — Bull reads the low guide midpoint as sandbagging against 10% organic growth and POR wins; bear reads it as management naming their own ceiling, gated by TSMC and OSAT capacity expansion they do not control.
2025Q4 · Feb 26, 2026
3 of 5 readings agreed3 HOLD · 1 AVOID · 1 LONGRead the full call
Q1 Growth Mid-Teens, Full Year Guide 6.5%, Cash Flow Down 30%
Qnity holds a real content-per-wafer position in CMP, advanced packaging and thermal materials, and the ICS 175bps margin expansion is volume-driven and disclosed. But 2026 upside is explicitly gated by customer capacity online at TSMC and the OSATs, and management anchored the midpoint to external MSI data at 6.5% while guiding free cash flow down 25-35% off a base flattered by one-time spin working capital release. Beat-and-raise mechanics are intact; the cash conversion story is not.
On the low end of guidance, management volunteered the exact downside scenario unprompted: 'more constraints really coming from the memory market and significant demand destruction.' They then immediately framed memory as 20% of semi and unit-driven with pricing out of the equation. Naming your own worst case in the same breath you discount it is the tell.
Adjusted EPS guided up 6-18% while adjusted free cash flow guides from $706M (15% of sales) down to $450-550M, a 25-35% decline. Management simultaneously declared a dividend and a $500M buyback authorization against that cash guide-down. Buyback language is 'opportunistic depending on market conditions.'
Q1 implies mid-teens YoY growth but full-year midpoint is 6.5%, requiring deceleration every quarter after Q1. Management could not name the driver beyond 'wait and see' on memory and said the high end is tied to customer capacity expansions they do not control.
Advanced logic moving high-70s to low-mid 80s, DRAM mid-80s to high-80s is presented as a tailwind, but those levels cap consumables pull-through until incremental customer capacity comes online, which management conceded is outside their control.
Seventh consecutive quarter of strong organic growth, FY organic sales +10% to $4.75B, adjusted pro forma EPS +12% to $3.35. Q4 beat holds even after $40M of sales shifted out into Q3, so the underlying beat is larger than headline.
POR wins across every single line of business in 2025, each a design-in scaling to production over 2-3 years. Emblem CMP pad platform targets N3, N2 and HBM3/4 with customer feedback described as outstanding.
ICS advanced packaging, interconnect and thermal each grew 20%+ with 175bps of segment margin expansion to just over 25%, driven by volume and favorable mix, disclosed as such. Management expects similar dynamics in 2026.
CapEx elevated to 9% of sales in 2026 versus 6% normalized, funding node transition ramps and local-for-local capacity, behind a guide the street may read as conservative. $100M EBITDA transformation program with $140M cost to achieve.
-25
CAUTIOUS"Management anchored guidance to external MSI/PCB indicators 'right down the fairway towards where the market estimates are,' used 'wait-and-see' on memory, and answered the high-end range with 'really tied to capacity expansions at our customers.' Interim CFO on first standalone call."
Full transcript with specific utilization rates, segment margins, FCF walk, memory exposure math, and named capacity constraints. Interim CFO and hedged memory commentary are the visible gaps.
Customer advanced packaging and advanced node capacity coming online in the back half. Management: high end of range 'is really tied to capacity expansions at our customers and their ability to scale new node transitions and bring that capacity online.'
Memory demand destruction. Management named it: low end is 'more constraints really coming from the memory market and significant demand destruction.' Memory is 20% of semi portfolio.
“We had a strong finish to the year with fourth quarter net sales of $1.2 billion, up 8% year-over-year as we continue to capitalize on key growth drivers namely advanced nodes, advanced packaging and interconnect as well as thermal management solutions.”
Figures are extracted from the transcripts and checked in code: every quote appears word for word. Descriptive only: none of this feeds the verdict.
Seventh straight beat, 10% organic growth, EPS +12%, and the Q4 print absorbed a $40M pull-forward into Q3. ICS advanced packaging, interconnect and thermal each grew 20%+ with 175bps of real margin expansion to 25%+. POR wins across every line of business. CapEx at 9% of sales versus 6% normalized is management backing the guide with capital. Guidance midpoint at 6.5% is anchored to external MSI/PCB indicators while 2025 delivered 10% with content outperformance on top. The street is pricing a memory demand-destruction event into a business where memory is 20% of semi and unit-driven with premium weighting.
Free cash flow guides from $706M to $450-550M, down 25-35%, while EPS guides up. The 2025 FCF base benefited from favorable working capital following the spin, a one-time release that flattered the year they used to anchor cash conversion. They announced a dividend and $500M buyback against that guide-down, with language hedged to 'opportunistic.' Q1 mid-teens YoY versus 6.5% full-year midpoint requires deceleration every quarter and management could not name the driver. High end of range is gated by TSMC and OSAT capacity online, not their own demand. Consolidated margin expanded only 40bps while faster-growing ICS at 25% dilutes from semi's mid-30s.
Say/Do Gap
EPS guide up 6-18% against a 25-35% FCF guide-down, plus dividend and $500M buyback announced in the same release. Buyback is 'opportunistic depending on market conditions,' an authorization without commitment.
Capacity Ceiling
Full-year midpoint 6.5% versus implied Q1 mid-teens requires deceleration each quarter after Q1. High end explicitly tied to customer capacity online at TSMC and OSATs, a variable management conceded they do not control.
High Utilization as Ceiling
Advanced logic at low-mid 80s and DRAM at high-80s utilization cap consumables pull-through per wafer until incremental customer capacity arrives, which is the exact gating variable management named.
HOLD. The structural content-per-wafer thesis is real and ICS margin expansion is disclosed and volume-driven, but the tape is asking two questions I cannot size: how much of the 2025 FCF was one-time spin working capital, and whether customer capacity lands in H2. Asymmetry is capped on the upside by a 6.5% midpoint anchored to MSI and gated by TSMC, and unquantifiable on the downside until memory resolves. Wait for the H1 print to confirm the Q1-to-full-year deceleration math. If FCF tracking misses $450M or Q2 comes in flat versus Q1, the re-rate happens first on cash conversion.