CDNS — Cadence Design Systems, Inc.

The tape, not the fundamentals, is driving sentiment as of 2026-09-13: CDNS fell roughly 15-18% over a 6-to-7 session losing streak in September 2026, sits -10.45% over one month and -24.83% over three months, with RSI(14) at 32.75 and price 7.87% / 13.48% / 11.45% below its 20-, 50- and 200-day SMAs respectively. This happened despite a clean Q2 2026 beat (revenue $1.584B, +24% YoY; non-GAAP EPS $2.11) and a guidance raise to $6.26-6.34B revenue / $8.10 non-GAAP EPS, so the selling is multiple compression from elevated rates and rotation out of expensive software, not an earnings disappointment. Analysts remain constructive (consensus Buy; 26 Buy / 4 Hold / 1 Sell; average target ~$401.60), which is exactly the disconnect — the sell-side sees upside while momentum funds keep selling.

The trend inflected upward, not downward. YoY revenue growth ran 18.8% (Q3 2024), 26.9% (Q4 2024), 23.1% (Q1 2025), 20.2% (Q2 2025), then decelerated hard to 10.1% (Q3 2025) and 6.2% (Q4 2025) — before re-accelerating to 18.7% (Q1 2026) and 24.2% (Q2 2026). Full-year growth confirms the re-acceleration: FY2024 +13.5% to $4.64B, FY2025 +14.1% to $5.30B, FY2026 guided to ~+19% at the $6.30B midpoint. Operating margins are broadly stable in the high-20s to low-30s (28.4% in Q2 2026 vs 29.0% in Q2 2025; 32.3% in Q4 2025), with R&D at 33.5% of revenue in Q2 2026 down from ~35% two years earlier. Free cash flow is improving on a trailing basis — Q2 2026 FCF of $582.3M was the highest single quarter in this three-year dataset, up from $333.5M in Q2 2025. Two caveats worth naming: Q2 2025 GAAP EPS of $0.59 was depressed by a below-the-line charge (net income $160.1M on $370.4M of operating income), so Q2 2026's $1.33 EPS overstates underlying momentum; and Q1 2026 gross profit of $1.413B on $1.474B of revenue (95.8%) looks like an accounting/reclass artifact rather than a real margin step-change. Net: this is a genuine re-acceleration supported by the record $8.1B backlog, not a one-quarter blip.

After the September 2026 drawdown the valuation is fair-to-attractive on a relative basis, though not cheap in absolute terms. As of 2026-09-13, at $289.37: TTM P/E of 57.19x versus a 5-year average of 69.78x (~18% discount); TTM P/S of 13.65x versus a 5-year average of 16.16x (~15% discount); PEG (TTM) of 1.59; forward P/E of ~35.6x on the $8.139 next-FY consensus EPS; EV/EBITDA of 36.9x; P/B of 11.5x; FCF yield of ~2.1% on TTM FCF of $1.679B. A 35.6x forward multiple against ~19% revenue growth and expanding free cash flow is a defensible PEG near 1.9-2.0 for a wide-moat EDA duopolist with 88.5% gross margins and 30.8% operating margins. The 5-year average P/E is also flattered by GAAP earnings depressed by acquisition amortization, so the discount versus history should not be taken entirely at face value. Verdict: Fairly valued, with a real de-rating cushion that did not exist six months ago.

Price and estimates have decoupled. As of 2026-09-13 the company has beaten and raised (FY26 revenue guidance lifted to $6.26-6.34B; non-GAAP EPS to $8.10) while the stock fell 24.83% over three months — meaning the market is de-rating the multiple faster than it is marking down the earnings. Two readings: (a) the market is pre-emptively pricing an AI-capex digestion cycle or a growth normalization from ~19% toward low-double-digits, or (b) this is a rates/rotation artifact in a name that has simply become too expensive to hold in a high-rate tape. The tell will be Q3 2026 in late October/early November 2026: if the record $8.1B backlog converts and guidance is raised again, the estimates grind the multiple back up and the drawdown looks like an entry; if bookings decelerate, the de-rating was prescient. Do not read Q2 2026's +125% GAAP EPS growth as signal — it is mostly the depressed $0.59 Q2 2025 base, not new earnings power.

Clearly damaged and not yet repaired as of 2026-09-13. RSI(14) at 32.75 is near but not below the classic 30 oversold line; price of $289.37 sits 7.87% below the 20-day SMA (~$314), 13.48% below the 50-day SMA (~$335) and 11.45% below the 200-day SMA (~$327), a fully bearish stack with the stock in the bottom third of its 52-week $262.75-$416.69 range and only 10.1% above the 52-week low. The one mildly constructive detail: the damage is aging — the 3-month return is -24.83% but the 1-week return is only -1.14% and the 6-month return is +0.82%, so selling pressure is decelerating rather than accelerating. Entry-timing implication: there is no momentum confirmation of a bottom, so a full position here is a bet against the trend; a first tranche on weakness and a second only on a close back above the 20-day SMA (~$314) is the disciplined read.

• Interest rates are the dominant swing factor for the multiple: the stock fell 24.83% over three months as of 2026-09-13 despite beating and raising, a textbook rate-driven de-rating of high-multiple software rather than an earnings problem.
• AI infrastructure capex — hyperscaler custom silicon, PCIe 6.0/CXL IP and advanced packaging are the fastest-growing source of incremental EDA seats and IP royalties, and Cadence is directly levered to that spend; it is also the most obvious place a digestion pause would first show up.
• U.S.-China export controls and trade tension — Cadence has posted double-digit growth in parts of Asia including China, making any tightening of EDA licensing rules a direct revenue and compliance risk rather than a headline risk.
• Slowing Moore's Law and rising design complexity (multiphysics, 3D-IC, advanced packaging) structurally raise EDA tool intensity per design — a multi-year tailwind that also underpins the Hexagon Design & Engineering integration.
• Sector rotation and high-multiple de-rating — capital has been moving out of expensive software into cheaper parts of the market, which argues the September 2026 decline was a multiple event compounded by forced momentum selling, not a fundamentals event.

• Q3 2026 earnings, expected late October/early November 2026 — the first real test of whether the raised FY2026 guidance ($6.26-6.34B revenue, $8.10 non-GAAP EPS) and the record $8.1B backlog convert to cash amid ongoing integration spend.
• Management's September 2026 appearances at the Goldman Sachs Communacopia + Technology Conference — commentary on long-term margin trajectory and AI/agentic monetization could reframe the multiple debate.
• Continued AI design-win momentum: first-pass PCI-SIG compliance for Cadence's PCIe 6.0 PHY and controller IP on TSMC N3 (announced September 2026) supports IP attach rates into high-end AI hardware.
• Any easing in rates or rotation back into quality software would mechanically re-rate a stock trading at ~35.6x forward EPS versus a 5-year average trailing P/E of 69.78x.
• Ongoing capital deployment — the Q1 2026 balance sheet step-up (total assets +$1.95B, equity +$1.09B, cash down ~$1.6B) is consistent with a stock-and-cash acquisition, and Hexagon D&E cross-sell plus further M&A remain live levers.

• Valuation still offers no margin of safety: ~35.6x forward EPS, 36.9x EV/EBITDA and roughly 48x TTM free cash flow (a ~2.1% FCF yield) mean a slide from ~19% growth to low-double-digits would compress the multiple hard — the September 2026 drawdown already demonstrated the speed.
• AI/hyperscaler customer concentration — a digestion pause in custom-silicon capex would hit the fastest-growing portion of the demand curve first and could show up in bookings before revenue.
• China export controls and geopolitical escalation — a direct licensing and regulatory risk to a region that has been a double-digit growth contributor.
• GAAP versus non-GAAP gap — TTM GAAP EPS of $5.05 against FY2026 non-GAAP guidance near $8.10 reflects heavy acquisition amortization and stock-based compensation, and the Hexagon D&E integration plus any further deal adds execution and margin risk.
• Competitive intensity — a combined Synopsys/Ansys in multiphysics and continued feature/price competition could cap Cadence's pricing power at the margin.
• Broken technical trend — price below the 20-, 50- and 200-day SMAs and only ~10% above the 52-week low means momentum-driven selling can continue even with the business executing, and the 52-week low of $262.75 is a realistic near-term downside reference.

This is a rare case where the business and the tape tell opposite stories, and as of 2026-09-13 I would side with the business. Q2 2026 revenue grew 24.2% YoY to $1.584B with a record $8.1B backlog and raised FY26 guidance; gross margin is ~88.5%, operating margin 30.8%, ROE 20.3%, debt/equity is only 0.36x and current ratio 1.74x, and the accruals ratio of -2.22% says cash generation is running ahead of reported profit — the healthy pattern. Meanwhile the stock has de-rated to ~35.6x forward EPS from a 5-year average trailing P/E of 69.78x, so the multiple, not the earnings, has taken the hit. Accelerating fundamentals plus a compressing multiple is usually a better setup than the reverse. The honest counterpoints: the trend is broken technically, the absolute multiple is still rich, and a China or AI-capex shock would hurt. Scale in rather than back up the truck.

I applied a 40x forward P/E — a ~2.2 PEG on the ~18-19% revenue and EPS growth the company is guiding to, and a discount to the 5-year average trailing P/E of 69.78x to reflect today's higher rate environment — to the $8.139 next-FY consensus EPS, giving $325.56, then cross-checked with a simplified DCF on TTM free cash flow of $1.679B (18% growth for eight years, 3.5% terminal growth, 8.5% discount rate, net debt of ~$1.04B), which produced ~$316; the two methods bracket $320-326 and I am using $325 as of 2026-09-13.