CDNS 10-K Analysis: Four Surfaces of Earnings Reveal Hidden Operating Leverage

Cadence Design Systems, the EDA duopoly partner that makes the software behind nearly every advanced semiconductor, reported a 93 basis point operating margin decline in its FY2025 10-K. But strip away a $128.5 million DOJ settlement and $29.2 million in restructuring charges, and adjusted margins actually expanded more than 200 basis points to 31.1% — a fact no major earnings summary mentioned.

That's two readings of the same company. There are two more. Management's non-GAAP operating margin — excluding stock-based compensation, amortization, and acquisition costs on top of the settlement — reached 44.6%. That creates a 16.4 percentage point gap between GAAP and non-GAAP earnings, the widest in Cadence's history and 3 points wider than FY2024. Meanwhile, the cash earnings picture has its own distortion: a $151 million legislative tax windfall inflated reported free cash flow growth from approximately 28% to 42%. Revenue grew 14.1% to $5,297 million, and $7.8 billion in contracted backlog covers 1.5 years of revenue at current run-rate. The question isn't whether Cadence is executing — it is. The question is which Cadence investors are actually buying: the 28% business, the 31% business, or the 44% business.

Four Surfaces of Earnings

Every quarter, Cadence reports a single operating margin number. Every quarter, that number understates — or overstates — the business depending on which costs you count. FY2025 turned this ambiguity into a chasm. GAAP operating income came in at $1,492 million, a 28.2% margin that declined 93 basis points year-over-year. The decline looks like deceleration. It isn't.

The 10-K buries $177.8 million in charges that have nothing to do with the software business: a $128.5 million BIS/DOJ settlement for export control violations, $29.2 million in restructuring under a September 2025 workforce reduction plan, and $20.1 million in discretionary charitable contributions to the Cadence Giving Foundation. Remove those items, and adjusted operating income reaches $1,670 million — a 31.5% margin that expanded more than 200 basis points from FY2024's 29.1%. The underlying business is generating real operating leverage, with total employee costs declining 130 basis points to 48.8% of revenue as headcount efficiency outpaced compensation growth.

But management's preferred non-GAAP view adds another layer. Excluding SBC ($455 million, 8.6% of revenue), amortization of acquired intangibles, and acquisition costs, non-GAAP operating margin hit 44.6% — 16.4 percentage points above GAAP. That gap widened 3 points from FY2024's 13.4 percentage points, and it's growing because every category of adjustment is increasing.

The bridge between GAAP and non-GAAP reveals where the gap is widening. SBC crept up from 8.4% to 8.6% of revenue, partly driven by market-based equity awards granted to senior management that accelerate expense recognition when the stock outperforms. Acquisition and integration costs rose from 2.1% to 2.5% as Cadence absorbed deal costs from its system-simulation buildout. The settlement added 2.4 percentage points that should not recur. Whether the gap narrows in FY2026 depends on whether SBC and acquisition costs continue climbing — and with Hexagon closing, acquisition costs almost certainly will.

Cadence's GAAP-to-non-GAAP operating margin gap widened to 16.4 percentage points in FY2025 — up from 13.4 points a year earlier — because rising stock-based compensation, acquisition costs, and a $128.5 million settlement are inflating the distance between reported and management-preferred earnings. An investor using GAAP sees a 76× P/E. An investor using non-GAAP sees 44×. Both are looking at the same company.

The FCF Mirage

Two of Cadence's strongest FY2025 headlines — 42% free cash flow growth and 19% China revenue growth — are partially manufactured by forces that have nothing to do with the software business.

Start with cash flow. Reported operating cash flow reached $1,729 million, up $469 million from FY2024. Free cash flow hit $1,587 million, a 42% surge that looks like a breakout. But the 10-K discloses a critical detail buried in the tax discussion: the One Big Beautiful Bill Act, enacted in July 2025, restored immediate expensing of U.S. research and development costs.

"On July 4, 2025, the One Big Beautiful Bill Act ('OBBBA') was enacted in the United States. The OBBBA included the restoration of the immediate expensing of United States research and development costs starting in fiscal 2025. This legislation decreased our fiscal 2025 cash tax payments by approximately $151 million."

That $151 million represents 8.7% of operating cash flow. Without it, FCF would have been approximately $1,436 million — still strong, still growing, but at 28% rather than 42%. The OBBBA benefit was unusually large in its first year because it captured previously capitalized R&D from prior periods. Whether it persists depends on legislative stability, but its magnitude will likely diminish in future years as the catch-up effect fades.

The China story has a similar structure. Revenue from China grew 18.7% to $680.0 million — a headline that suggests resilience despite export controls. But the three-year trajectory tells a different story: FY2023 revenue was $679.5 million, meaning FY2025 simply recovered to the level that prevailed before a 15.7% collapse in FY2024 during the BIS investigation. Over two years, China revenue grew by $0.5 million. This is a V-shaped recovery, not a growth trajectory. And BIS demonstrated its willingness to intervene directly: a 40-day export freeze from May 23 to July 2, 2025 temporarily banned all EDA software exports to China before being rescinded.

Cadence's reported 42% free cash flow surge to $1.59 billion was inflated by a $151 million legislative tax windfall from the OBBBA, reducing underlying FCF growth to approximately 28% and the durable FCF margin to 27.1%. Investors using FY2025 FCF or China growth as valuation inputs are building on a partially inflated base — the durable run-rate is lower and the geographic risk is unchanged.

The Capital Allocation Squeeze

Cadence is stretching capital across four simultaneous priorities while incremental returns on invested capital have collapsed to their lowest level in recent history. The question is whether the €2.7 billion Hexagon D&E acquisition — the largest in company history — will reverse the trend or accelerate it.

In FY2025, Cadence repurchased $925 million in stock (58% of reported FCF), paid $116.5 million in debt service on $2.5 billion in Senior Notes, spent $141.9 million in capital expenditures, and invested $1,712 million (33% of revenue) in R&D. Now it faces the Hexagon closing, which requires approximately €1.89 billion in cash — roughly $2.05 billion at current exchange rates.

Post-Hexagon, Cadence's balance sheet enters unfamiliar territory. Current Senior Notes total $2.5 billion ($500 million due 2027, $1.0 billion due 2029, $1.0 billion due 2034). Add an estimated $1.05 billion revolver draw to fund the acquisition, and pro-forma debt reaches approximately $3.55 billion against $1,720 million in EBITDA — a 2.1× debt/EBITDA ratio. That's well below the 3.5× covenant limit, leaving roughly $2.5 billion in additional borrowing capacity before tripping covenants. But it's the highest leverage Cadence has ever operated at, and it arrives alongside a separate constraint.

"Failure to comply with the terms of the plea agreement, the administrative settlement agreement or our probation could result in further criminal, civil, or administrative proceedings or denial of export privileges. ... our ongoing obligations under these agreements will generally apply to any new business entities we acquire, which could limit our ability to acquire new businesses that may be strategically important to us."

The BIS/DOJ criminal plea didn't just cost $140.6 million in cash. It imposed a 3-year probation whose compliance obligations extend to every entity Cadence acquires — including Hexagon's operations. Any violation during probation could result in denial of export privileges, threatening the 13% of revenue from China. This creates a compliance burden on future M&A that no peer faces.

The deeper concern is capital efficiency. Cadence's return on incremental invested capital — a measure of how much additional operating income each dollar of new capital generates — fell to 6.8% in FY2025, down from a three-year average of 28.7%. This metric comes from the MetricDuck pipeline rather than direct filing calculation, but the filing data corroborates the trend: goodwill grew from approximately $1.6 billion in FY2022 to $2.75 billion while operating income growth decelerated relative to invested capital growth. Purchase obligations doubled from $78.3 million to $162.1 million, with $99.6 million due in 2026 alone — signaling accelerating infrastructure commitments.

Cadence's return on incremental invested capital collapsed to 6.8% in FY2025 from a three-year average of 28.7% because acquisition-driven goodwill and intangibles now consume $3.47 billion of the balance sheet without generating proportional operating income. Hexagon will add significantly more invested capital to the denominator. If it generates returns above the 6.8% threshold, the silicon-to-system strategy is working. If it doesn't, Cadence is paying for TAM expansion with shareholder capital dilution.

The $7.8B Floor

Despite all the noise — the settlement, the tax windfall, the China regulatory overhang — Cadence's competitive position is strengthening underneath. The proof is in the backlog.

"Year-end backlog was $7.8 billion and revenue expected to be recognized in the next 12 months from remaining performance obligations was $3.8 billion."

At $7.8 billion, Cadence's year-end backlog covers 1.5 times its annual revenue — the strongest forward visibility signal in the filing. The $3.8 billion in remaining performance obligations expected within twelve months means 72% of FY2025's revenue is already contracted before a single new deal closes. Backlog grew $800 million sequentially from Q3's $7.0 billion, indicating bookings accelerated into year-end. For a company trading at 76× earnings, this level of revenue certainty is the single strongest argument against the bear case.

The operating leverage is real, not just accounting. Gross margin expanded 30 basis points to 86.4% despite cost of product and maintenance rising 19% year-over-year — faster than the 14.1% revenue growth. Hardware emulation costs are rising as Cadence installs more verification and prototyping systems, but services cost efficiency (down 3%) offset the pressure. Employee costs — salary, benefits, and SBC combined — fell from 50.1% to 48.8% of revenue. That 130 basis point improvement in labor intensity on a $5.3 billion revenue base translates to roughly $69 million in operating leverage.

The duopoly with Synopsys creates pricing power that no peer in this comparison enjoys. Cadence and Synopsys together control roughly 74% of the $15 billion+ EDA market, and customers cannot switch without requalifying their entire design flow with foundries like TSMC — a process that can take years and cost tens of millions. No single customer accounts for more than 10% of revenue, eliminating concentration risk. Other Americas revenue surged 81% to $168 million as Cadence diversified its geographic base, though from a small base (3% of total).

The complication is that recurring revenue declined from 83% to 80% as hardware and up-front product sales grew faster than subscription licenses. Revenue recognized over time also declined, from 80% to 76%. This isn't churn — it's a mix shift toward hardware emulation and prototyping systems that carry higher margins but less recurring predictability. If recurring revenue stabilizes at 80%, the business model remains intact. If it drifts toward 75%, the valuation premium deserves scrutiny.

Cadence's $7.8 billion year-end backlog covers 1.5 times its annual revenue, with $3.8 billion in remaining performance obligations providing 72% revenue visibility for the next twelve months — the strongest forward contract coverage in its peer group.

"Entity List restrictions and other trade restrictions may also encourage customers to seek substitute products from our competitors, including a growing class of foreign competitors and open source alternatives, that are not subject to these restrictions or to develop their own solutions, thereby decreasing our long-term competitiveness."

The filing explicitly acknowledges a competitive threat that the backlog numbers don't capture: Chinese domestic EDA competitors and open-source alternatives are gaining ground as U.S. export restrictions push customers to seek non-restricted substitutes. The backlog provides a multi-year buffer, but it cannot immunize against a structural shift in China's EDA ecosystem.

What the Price Demands

At $312.58 per share, Cadence trades at 76.4× GAAP earnings — a multiple that would look absurd for most software companies. But Cadence isn't most software companies, and the filing provides unusually strong evidence both for and against the premium.

The ROIC paradox is the cleanest framing. Cadence's return on invested capital of 16.2% is identical to Qualcomm's 16.2%. Yet Cadence commands 2.3× the P/E multiple — 76.4× versus 33.5×. The market is pricing Cadence's duopoly lock-in, 80% recurring revenue, and $7.8 billion backlog at a 130% premium over a semiconductor company with comparable capital efficiency. On non-GAAP earnings ($7.14 EPS), the P/E compresses to 43.8× — expensive but within range for a high-quality compounder with 14%+ growth and 30% FCF margins.

What does the stock price require? At an enterprise value of approximately $81.8 billion and underlying FCF of $1,436 million (ex-OBBBA), sustaining a 25× terminal EV/FCF multiple with a 10% required return implies Cadence needs to generate roughly $5.3 billion in annual free cash flow within five years. That's a 29.7% annual FCF CAGR — achievable only under three conditions.

First, revenue must compound at 14-16% annually. The $7.8 billion backlog supports this in the near term, and FY2026 guidance of $7.4 billion from the 8-K implies 40% growth off the current base (likely including Hexagon). Second, adjusted operating margins must expand toward 35%. The direction is right — 29.1% to 31.1% in one year — and employee cost leverage suggests room to continue. Third, Hexagon must contribute meaningfully without proportionally increasing the cost structure. Management expects 28 cents of EPS dilution from Hexagon in FY2026, but the acquisition targets a ~$50 billion engineering simulation market that could multiply Cadence's addressable TAM.

The filing supports conditions one (backlog visibility) and two (labor leverage, margin trajectory). Condition three is the swing factor — and the 6.8% ROIIC on recent acquisitions is the key warning sign. If Hexagon generates returns above that threshold, the valuation is achievable. If the ROIIC decline continues, Cadence is paying for a TAM story with shareholder capital.

Cadence trades at 76.4 times GAAP earnings — 2.3 times the P/E of Qualcomm despite identical 16.2% returns on invested capital — a premium that requires approximately 30% annual free cash flow growth for five years to justify.

What to Watch

Five metrics will test the thesis over the next twelve months:

1. FY2026 ROIIC — if it rebounds above 15% after Hexagon integration, the capital allocation strategy is vindicated. Below 5% and the acquisition-driven model is destroying value.

2. GAAP/non-GAAP gap — a narrowing below 14 percentage points in FY2026 (with the settlement gone) would signal SBC and acquisition cost deceleration. A widening above 15 pp would confirm structural adjustment creep.

3. China revenue — above $750 million annualized in any quarter signals genuine demand growth that refutes the V-recovery thesis. Below $600 million signals regulatory damage.

4. Recurring revenue mix — stabilization at 80% is fine. A decline toward 75% suggests the hardware shift is structurally altering the business model's predictability premium.

5. Pro-forma debt/EBITDA — the 2.1× post-Hexagon ratio should decline toward 1.5× within two years given $1.6 billion annual FCF. If leverage rises instead, management is deploying cash toward additional acquisitions rather than deleveraging.

At $312.58, the market implies Cadence can sustain near-30% FCF growth for five years while integrating its largest acquisition under criminal probation. The filing supports the revenue visibility ($7.8 billion backlog, 72% forward contracted) and the margin trajectory (31.1% adjusted, employee leverage improving). But it complicates the capital efficiency story (6.8% ROIIC, widening non-GAAP gap) and the geographic narrative (China flat over two years despite a "19% growth" headline). Every investor in Cadence is implicitly choosing an earnings surface — the filing makes the choice explicit.

Frequently Asked Questions

What does Cadence Design Systems do?

Cadence makes the software tools that semiconductor companies use to design chips. It operates in a duopoly with Synopsys, together controlling about 74% of the $15 billion+ electronic design automation (EDA) market. Customers cannot easily switch EDA vendors because their design workflows must be requalified with foundries like TSMC — creating near-total lock-in. Cadence generates 80% recurring revenue from multi-year software licenses, hardware emulation systems, and semiconductor IP blocks, with no single customer exceeding 10% of revenue.

Why did Cadence's operating margin decline in FY2025?

The reported 93 basis point decline to 28.2% is misleading. Cadence absorbed $128.5 million in BIS/DOJ settlement charges, $29.2 million in restructuring costs, and $20.1 million in discretionary charitable contributions. Excluding these items, adjusted operating margin expanded approximately 200 basis points to 31.1%. The underlying business exhibited genuine operating leverage on 14.1% revenue growth, with employee costs declining 130 basis points as a share of revenue from 50.1% to 48.8%.

How significant is the $151M OBBBA tax windfall?

The One Big Beautiful Bill Act restored immediate R&D expensing in the United States, reducing Cadence's FY2025 cash tax payments by $151 million. This represents 8.7% of operating cash flow and inflated reported FCF growth from approximately 28% to 42%. While R&D expensing may continue under current law, its first-year impact was unusually large because it captured previously capitalized R&D from prior years. Investors should use $1,436 million (ex-OBBBA) as the baseline for growth projections rather than the reported $1,587 million.

Is Cadence's China revenue actually growing?

Not really. The "19% growth" to $680.0 million in FY2025 is a base-effect rebound from $573.1 million in FY2024, which itself was a 15.7% decline from $679.5 million in FY2023. Over two years, China revenue is essentially flat (+$0.5 million). BIS imposed a 40-day export freeze on EDA software to China during mid-2025 that "negatively impacted revenue." Cadence remains under 3-year criminal probation, and the filing warns that the United States "may consider reimposing" export restrictions.

What does the 16.4 percentage point GAAP/non-GAAP gap mean?

Cadence's non-GAAP operating margin (44.6%) exceeds GAAP (28.2%) by 16.4 percentage points — the widest gap in company history, up from 13.4 pp in FY2024. The gap includes SBC (8.6% of revenue), amortization of acquired intangibles (2.0%), acquisition and integration costs (2.5%), restructuring (0.6%), and the BIS/DOJ settlement (2.4%). Non-GAAP EPS of $7.14 is 75.9% higher than GAAP EPS of $4.06. Whether you see Cadence as a 76× business or a 44× business depends entirely on which costs you treat as "real."

Can Cadence afford the Hexagon acquisition?

Yes, but at the highest leverage in company history. The €2.7 billion Hexagon D&E acquisition requires approximately €1.89 billion in cash. Cadence has $3.0 billion in cash (29% offshore) and a $1.25 billion undrawn revolver. Pro-forma total debt of approximately $3.55 billion against $1,720 million EBITDA yields a 2.1× debt/EBITDA ratio — well below the 3.5× covenant limit but Cadence's peak leverage. With $1.6 billion in annual FCF, Cadence can deleverage within two years — but this assumes no additional large acquisitions.

Why did ROIIC collapse from 28.7% to 6.8%?

Return on Incremental Invested Capital measures how efficiently new capital generates operating income. Cadence's acquisition-driven "silicon-to-system" strategy has deployed billions into goodwill ($2.75 billion) and intangibles ($718 million) that haven't yet generated proportional returns. The 6.8% one-year ROIIC versus 28.7% three-year average signals capital dilution — recent acquisitions are adding more capital to the balance sheet than they're adding in operating income. Note that ROIIC is a pipeline-calculated metric; the filing corroborates the trend through goodwill and intangible growth rates. Hexagon will test whether this trajectory reverses.

How does Cadence compare to its peers on valuation?

Among assigned peers, Cadence trades at 76.4× GAAP P/E — above QCOM (33.5×) and APP (67.7×), below SHOP (169.4×). On ROIC, Cadence's 16.2% matches QCOM exactly but commands 2.3× the P/E multiple — a premium the market attributes to Cadence's recurring revenue, backlog visibility, and duopoly positioning. On FCF margin, Cadence's 30.0% exceeds QCOM (29.0%) and SHOP (17.4%) but trails APP's exceptional 64.2%. The premium over QCOM is partly justified by 80% recurring revenue and $7.8 billion backlog. The premium over APP is harder to justify given APP's superior FCF generation and 70% revenue growth.

What is Cadence's $7.8 billion backlog?

As of year-end 2025, Cadence had $7.8 billion in total backlog covering approximately 1.5 years of revenue at current run-rate. Of this, $3.8 billion is expected to be recognized in the next 12 months, representing 72% of FY2025 revenue already contracted. Backlog grew $800 million sequentially from Q3's $7.0 billion, indicating bookings accelerated into year-end. This provides exceptional revenue visibility and partially mitigates the "growth sustainability" risk embedded in the 76× P/E — Cadence doesn't need to win significant new business to deliver 14%+ growth in the near term.

Should investors use GAAP or non-GAAP earnings for Cadence?

Neither alone is sufficient. GAAP (28.2% margin, $4.06 EPS) understates profitability by including genuinely one-time items like the $128.5 million settlement. Non-GAAP (44.6% margin, $7.14 EPS) overstates it by excluding $455 million in SBC that represents real economic dilution — 30.5% of operating income expensed through equity rather than cash. The most informative metrics are adjusted operating margin (31.1%), which excludes one-time items but keeps SBC, and underlying FCF margin (27.1% ex-OBBBA), which captures actual cash generation after all costs including the dilution effect of SBC.

What would make the investment thesis wrong?

The thesis depends on Hexagon failing to improve ROIIC and non-GAAP adjustments continuing to widen. It would be wrong if: (1) FY2026 ROIIC rebounds above 15%, proving Hexagon contributes operating income faster than invested capital grows; (2) China revenue exceeds $750 million, proving genuine demand growth rather than base-effect recovery; (3) the GAAP/non-GAAP gap narrows below 14 percentage points, showing SBC and acquisition cost deceleration; or (4) OBBBA is repealed and FCF still grows 20%+, confirming the cash generation engine is self-sustaining without legislative support.

What is the BIS/DOJ settlement and why does it matter going forward?

Cadence pled guilty to criminal conspiracy for export control violations involving a Chinese customer, paying $140.6 million in cash and entering 3-year probation. This isn't just a one-time financial charge — the probation terms apply to any new entity Cadence acquires, including Hexagon, which imposes an ongoing compliance burden and could limit future M&A flexibility. Any violation during probation could result in denial of export privileges, directly threatening 13% of revenue from China. The filing also warns that the United States "may consider reimposing" export restrictions on EDA software that were temporarily imposed for 40 days in mid-2025.

Methodology

Data Sources

This analysis draws from three primary sources: (1) Cadence Design Systems' FY2025 10-K Annual Report, filed February 19, 2026 (CIK: 0000813672); (2) the Q4/FY2025 8-K Earnings Release, filed February 17, 2026; and (3) MetricDuck's automated financial data pipeline, which extracts standardized metrics from XBRL filings for 5,000+ U.S. public companies. Peer comparison data (QCOM, SHOP, APP, INTC) comes from the same pipeline applied to each peer's most recent annual filing.

Limitations

• Pipeline debt classification gap: The MetricDuck pipeline reports $0 total debt for CDNS due to an XBRL concept mismatch. All debt figures in this analysis use filing-sourced data ($2.5B Senior Notes).
• ROIIC is pipeline-derived: The 6.8% ROIIC and 28.7% three-year average come from the metrics processor and cannot be independently verified from 10-K numbers alone. Filing evidence (goodwill and intangible growth rates) corroborates the directional trend.
• Organic vs. inorganic revenue: Cadence does not quantify acquisition-contributed revenue separately. The 14.1% growth rate includes contribution from recent acquisitions.
• Hexagon pro-forma estimates: Post-Hexagon leverage calculations assume EUR/USD 1.085 exchange rate and estimated funding mix ($1.0B cash, $1.05B revolver). Actual numbers will differ.
• Peer set mismatch: The assigned peers (QCOM, SHOP, APP, INTC) are not direct EDA competitors. Synopsys (SNPS) would be the most relevant peer but was not in the assigned set. Comparisons focus on business model analogues rather than direct competitors.
• APP revenue gap: APP (AppLovin) FY2025 revenue was not available in the extracted pipeline data. The approximately $4.7B figure is from filing intelligence, not pipeline-verified.

Disclaimer: This analysis is for informational purposes only and does not constitute investment advice. The author does not hold positions in CDNS, QCOM, SHOP, APP, or INTC. Past performance and current metrics do not guarantee future results. All data is derived from public SEC filings and may contain errors or omissions from the automated extraction process. Investors should conduct their own due diligence before making investment decisions.