Dupont De Nemours Inc DD

Thesis

DuPont de Nemours is the rump of the 2017 Dow-DuPont mega-merger that has now spent eight years dismantling itself: Dow (2019), Corteva (2019), and now a tri-spin into Qnity Electronics (completed November 2025), a future Water/Healthcare entity, and an Industrials RemainCo. The composite scorecard score is 62/100, which sounds respectable until you read the components: profitability 11, capital allocation 16, valuation 17. The reason is simple — the 10-year ROIC average is 0.0% and FCF conversion is 0.0%, both crushed by the 2018-2020 goodwill impairments, the agriculture/materials carve-outs, and a TTM owner-earnings figure of negative $432 million. The stock at $46.24 trades at 27.5x trailing earnings versus a 10-year average of 12.9x, while the scorer's reverse-DCF spits out an intrinsic-value range of negative $14.85 to negative $10.30 per share — i.e. the model cannot find a positive IV from the consolidated cash flows because the base CAGR was clamped from -12.9% to -5.0%.

The compounder case requires you to ignore the consolidated numbers and underwrite three new pure-plays separately, plus a stub. That is not a Buffett-Munger setup [4]; that is event-driven arbitrage. The honest verdict is that you cannot buy DD as 'a wonderful business at a fair price' until the spins clear, the PFAS liability tail is bounded, and the RemainCo Industrials entity reports two clean quarters of standalone cash flow. Net debt/EBITDA reads -1.13x (technical net cash from spin proceeds), which is comforting but transient. Price/IV math: with IV negative on every scenario, there is no margin of safety at any price the scorer can compute. Wait.

Moat

DuPont (post-spin RemainCo) will be a portfolio of specialty chemicals businesses — Kevlar/Nomex aramids, Tyvek protective materials, Spectra/Sorona fibers, and water-purification membranes (FilmTec, IntegraTec) until that segment is also separated. We score moats by the five Buffett-Damodaran categories.

Pricing power. Mixed. Aramids (Kevlar, Nomex) genuinely have brand equity inside specifying engineers — a body-armor or aerospace customer cannot trivially swap to Teijin's Twaron or Hyosung's Alkex without re-qualification. That is real, but limited: there are only two or three end-uses (ballistic, fire-resistant apparel, optical-fiber reinforcement) and each has a credible second source. Tyvek is the closest thing to a brand moat — Damodaran's Coca-Cola test [1] requires that taking the brand away would destroy value, and in housewrap and sterile medical packaging, 'Tyvek' is genuinely the generic term. That is narrow but real.

Switching costs. Stronger than pricing power, weakest where it matters most for valuation. Specifying a Kevlar-grade aramid into a Mil-spec or NIJ-certified armor plate creates re-qualification costs that are 18-36 months and seven figures — Microsoft-style switching cost lock-in [2] applied to industrial chemistry. But the installed base is finite and the qualifications are eventually granted to competitors. Water-membrane elements have similar dynamics: a desalination plant that designs around FilmTec spirals will not swap mid-cycle, but new plants are bid competitively. Verdict: NARROW where it exists.

Network effects. None. Specialty chemicals do not get better with more users.

Intangibles (patents, brands, regulatory). This is where DuPont historically lived — Nylon, Teflon, Lycra, Kevlar were all genuine 20-year-patent monopolies that compounded for decades. The problem in 2026 is that the most valuable intangibles have already been spun out (Corteva took the seed and crop-protection patents; Qnity is taking the semiconductor materials IP including photoresists and CMP slurries). The RemainCo will hold mature, off-patent chemistries. Damodaran's warning [2] is directly on point: 'the companies that will see the greatest increases in value are not necessarily the companies that spend the most on R&D, but those who have the most productive R&D departments.' Post-spin RemainCo R&D productivity is a question mark, not a fact.

Cost advantages. DuPont has scale-based feedstock advantages in some integrated sites (Spruance for Kevlar, Richmond for Nomex) but feedstock for these products is naphtha- and benzene-derived — i.e. the company is a price-taker on its largest input cost. There is no Saudi-Aramco-style structural cost moat. Manufacturing-process know-how on Tyvek's flash-spinning and on aramid spinning is genuine but defensible only as trade secret, not as cost advantage versus a determined competitor with $10B over five years.

Competitor stress test. Give Teijin, Toray, 3M, or Honeywell $10B and five years and they could meaningfully erode the aramids and protective-materials franchises. Tyvek would be hardest to dislodge because of building-code specification inertia, but a determined Chinese entrant (Sinopec/CNPC) could fund a cheaper substitute housewrap.

Erosion risk. Real and ongoing. Carbon-fiber composites are quietly displacing aramids in some ballistic applications. PFAS regulation is forcing reformulation of multiple Tyvek and protective-fabric grades. Water-membrane technology is commoditizing as Chinese producers (Hydranautics, Vontron) close the technology gap.

Moat verdict: NARROW.

Management & Capital Allocation

CEO Ed Breen is the central capital-allocation fact. Breen ran Tyco through its 2007 tri-spin (Tyco Electronics → TE Connectivity, Covidien, Tyco International), built a track record of breakup value-creation, and was brought back to DuPont specifically to repeat the play. The 2017 Dow-DuPont merger and subsequent four-way split (Dow 2019, Corteva 2019, Qnity 2025, future Water/Healthcare) is the largest sequenced corporate breakup in U.S. industrial history. So the question is whether Breen has been a good capital allocator within that framework, judged across Buffett's five choices.

Reinvest. RemainCo capex has run at roughly 5-6% of sales — adequate maintenance plus modest growth on Kevlar and Tyvek capacity adds. ROIC on incremental dollars is unmeasurable here because the scorer notes 'Net capital return period; ROIIC not meaningful' and 'roic_10y_avg = 0.0' on consolidated numbers polluted by impairments. This is a genuine gap; we cannot grade reinvestment quality without the post-spin standalone numbers.

Acquire. The 2017 Dow merger destroyed enormous value at the SUM-OF-PARTS level (the merger cost $130B+ in combined market cap by the 2019 trough); however, it was always a means to an end — the breakup. The 2021 sale of Nutrition & Biosciences to IFF (~$26B Reverse Morris Trust) was generally well-priced. The 2024 Spectrum Plastics carve-out and the November 2025 Qnity tax-free spin used the most efficient corporate-tax structures available. Grade on M&A: above average, weighted by execution rather than thesis.

Debt. Conservative. Scorecard shows net debt/EBITDA of -1.13x (technical net cash, partly from Qnity dividend proceeds at spin). DuPont retired ~$8B of legacy Dow-Du-Corteva debt in 2020-2022 and has not re-levered. Interest coverage data is null in the scorecard but qualitatively the balance sheet is among the strongest in specialty chemicals. The PFAS settlement liability (Water District Settlement Fund — see filing tags) is the offsetting concern: the cost-share with Chemours and Corteva is a shared-cost arrangement that has so far been honored but is not capped against a worst-case multi-state attorney-general settlement.

Buybacks. Share count is down 11.59% over ten years (-0.1159 in scorecard). That is real shrinkage, but most of it occurred in 2022-2023 after the IFF cash hit at prices in the $60-80 range. With the stock now at $46.24, retiring that capital above current price is a mediocre trade in hindsight — Munger would call it a Type-1 capital-allocation mistake, buying the company's own stock above intrinsic value. The buybacks paused in 2024-2025 ahead of the spins, which is correct discipline.

Dividends. Steady ~$1.52 annual, ~3.3% yield. Will be re-set across the three SpinCos. Acceptable.

Communication quality. Breen's investor letters are clear, specific on segment cash generation, and honest about PFAS exposure. The 10-K discusses litigation reserves with specificity. No earnings-management red flags of the type Buffett warned about [3] in his 1984 insurance letter.

The central tension: Breen is genuinely good at the specific job he was hired for (orderly breakup), and that job is mostly done. The next CEO — running the Industrials RemainCo — is not yet identified at the time of this analysis. You are buying a transition.

Capital allocator: B.

Industry Structure

Specialty chemicals is a structurally average industry with pockets of excellence. Porter's Five Forces applied to the post-spin RemainCo Industrials business (aramids, protective materials, mobility, films):

Threat of new entrants — LOW. Capital intensity is high (a Kevlar plant is a $1B+ greenfield), permitting for any halogenated-chemistry facility takes 5-7 years in the U.S. and EU, and customer qualification cycles in aerospace and defense are 18-36 months. PFAS-related regulatory tightening actually increases the barrier for new entrants who would have to certify replacement chemistries from scratch.

Bargaining power of suppliers — MEDIUM-HIGH. Feedstocks (benzene, terephthalic acid, naphtha, fluorinated intermediates) are commodity petrochemicals with concentrated supply. DuPont takes price; it is not vertically integrated upstream the way Dow is. Energy costs (natural gas for the U.S. Gulf sites, electricity for Asian sites) are pass-through with a lag, which compresses margin in inflationary years.

Bargaining power of buyers — MEDIUM. Auto OEMs (mobility solutions), aerospace primes (Kevlar honeycomb), and water-utility EPC contractors are all sophisticated, multi-source-qualified buyers who run RFQs. Defense and law-enforcement (body armor) is more captive but smaller. Construction (Tyvek housewrap) sells through distribution to a fragmented installer base — best pricing power here.

Threat of substitutes — MEDIUM-HIGH AND RISING. This is the single biggest industry-structure risk. Carbon fiber and UHMWPE (Honeywell Spectra, DSM Dyneema) are taking ballistic share from aramids on a 10-year arc. Mineral-based housewraps and rigid-insulation-with-WRB are taking some Tyvek share. In water, ceramic membranes are slowly entering high-fouling applications. PFAS regulation is forcing substitution within DuPont's own portfolio.

Rivalry among existing competitors — MEDIUM. Aramids: DuPont vs Teijin vs Hyosung — three rational players. Protective materials: DuPont, 3M, Honeywell, Ansell — four-way oligopoly with discipline. Water: DuPont, Toray, LG Chem, Hydranautics, Suez — fragmenting. Films and adhesives: highly fragmented.

Value pool location. Within the chemical value chain, the highest-return pools sit (a) immediately downstream of patent-protected molecules and (b) inside specifications and approvals. DuPont has historically owned the second pool well. But Damodaran's reversion-to-mean point [5] — 'there is a tendency, albeit slow, for the returns at companies to converge on industry averages' — is exactly the dynamic visible in DuPont's 10-year ROIC of 0.0%. The most patent-protected, highest-return pools have already been monetized via spin (Corteva seeds, Qnity semiconductor materials).

Trajectory. The end-markets DuPont serves are growing at GDP-or-slightly-better rates: water infrastructure (good demographic tailwind), defense (cyclical-up currently), construction (cyclical), automotive lightweighting (positive). But the share of value captured by specialty chemicals within those end-markets is flat-to-down as customers professionalize procurement.

Industry Verdict: Average.

Inversion (Bear Case)

I am short DuPont at $46.24 and here is why I expect to be paid.

The single event that kills this. A multi-state attorney-general PFAS settlement that breaks the existing cost-share arrangement with Chemours and Corteva. The 2023 public-water-systems settlement ($1.185B DuPont share) was always the floor, not the ceiling. There are at least three open litigation tracks — personal-injury MDL, state-AG natural-resource damages, and EU REACH-driven civil claims — any one of which can be a multi-billion dollar discrete event. The cost-share allocation MOU between DuPont, Chemours, and Corteva was a 2021 negotiated split; a Chemours bankruptcy (it carries $4B in debt against deteriorating margins) would re-impose joint-and-several liability on DuPont. That single event — Chemours Chapter 11 — could hit DuPont with $5-10B of incremental liability that is currently sitting off-balance-sheet as a contingent shared cost.

Why the moat is narrower than bulls think. Bulls point to Kevlar and Tyvek brand equity. The reality is that the highest-margin intangibles already left in the spins. Qnity took the semiconductor photoresists, CMP slurries, and advanced-packaging materials — those are the businesses that earn 25-30% EBITDA margins and grow at 8-12%. What is left in RemainCo Industrials is mature, GDP-growth chemistry with 18-22% EBITDA margins and 1-3% volume growth. Damodaran's reversion warning [5] applies in full. Tyvek is brand-strong but housewrap is a $2B end-market growing at construction-cycle rates, not a compounder. Aramids face a slow carbon-fiber substitution arc that bulls under-weight because it shows up over decades, not quarters.

Why management is worse than it appears. Ed Breen's tenure has been excellent at the breakup but mediocre at the operations within each piece. The 2018-2020 goodwill impairments — a cumulative $4-5B — were the unwinding of valuations Breen's predecessors paid in the Dow merger, but Breen approved retaining and writing down rather than selling early. Buybacks at $60-80 in 2022-2023 against a current $46 price are a clear capital-allocation error visible in hindsight. Most importantly: the next CEO of RemainCo Industrials is not yet named at the time of this analysis. Breen is the special-situations operator; he is not the long-term Industrials operator. The succession is the under-priced risk.

What bulls are extrapolating that won't hold. The bull case rests on three extrapolations. First, that each spin will trade at its peer-group multiple immediately and DuPont stub will hold at current EV. Reality: Qnity opened at a discount to MKS Instruments and Entegris peers, suggesting the market is haircutting governance and litigation tails. Second, that PFAS reserves are conservative. Reality: every legacy environmental liability in U.S. industrial history (asbestos, lead paint, MTBE, opioids) has cost 3-5x initial reserve estimates over a 20-year tail; PFAS will rhyme. Third, that defense and water capex super-cycles will lift Industrials revenue through 2030. Reality: defense spending is at a cyclical peak, and water infrastructure spending in the U.S. is dependent on IIJA appropriations that face Congressional re-authorization.

Valuation trap. The stock at $46.24 trades at 27.5x trailing earnings. The 10-year average is 12.9x. Bulls argue 'the trailing number includes spin-related charges, the forward number is X.' That argument has been deployed continuously since 2018. The scorer's reverse-DCF cannot generate a positive intrinsic value at any scenario — IV low/base/high all sit at -$10 to -$15. Owner earnings TTM is negative $432 million. The only way the stock works is if you ascribe nearly all enterprise value to the spin proceeds plus a SOTP of forward-pro-forma RemainCo. That is a multiple-expansion bet, not an earnings bet, and multiple-expansion bets fail when (a) interest rates stay high, (b) cyclical earnings disappoint, or (c) litigation reserves prove inadequate.

If I am right, the stock could be worth $28-32 within 24-36 months — a 30-40% drawdown driven by a PFAS reserve-true-up, a Chemours liquidity event, or simply RemainCo Industrials trading at a 12x peer multiple on disappointing standalone EBITDA.

If I am right, the stock could be worth $30 within 3 years.

Lollapalooza Bias Check

Several biases are pulling on the analyst right now and the discipline is to name them rather than overrule them.

Authority bias. Ed Breen is rightly admired for the Tyco breakup and is treated by sell-side analysts as a name-brand operator who delivers. There is a temptation to give the breakup execution the benefit of the doubt because of the prior. Munger's antidote: judge the deal in front of you, not the deals in the rearview mirror.

Anchoring. The current $46.24 price is anchored against the post-Dow-DuPont 2019 lows in the high $20s and the 2021 highs in the $80s. That range makes $46 feel like the middle, hence 'reasonable.' But the relevant anchor is intrinsic value, which the scorer cannot calculate positively. Resist the chart.

Confirmation bias. I came into this analysis aware of the four-way breakup story, the Qnity spin completion, and Breen's track record. Most readily-available coverage is constructive. The harder work — finding the bear case — required looking at the consolidated cash flows, the goodwill impairment history, and the PFAS legal docket, all of which point negative.

Recency bias. The Qnity spin completed November 2025. The market's pricing of DuPont at $46 reflects very recent SOTP math from sell-side desks. Six months from now we will have two clean quarters of standalone Qnity data and one quarter of RemainCo data; that information will be priced in by then but is not priced now. Acting on recency means acting on incomplete data.

Commitment / consistency. Investors who own DuPont through prior spins are inclined to hold through this one because each prior spin paid off. This is exactly the bias the inversion section was designed to counter.

Deprival super-reaction. If I do not own DuPont and the post-spin SOTP works, I miss a 30-50% upside in 18 months. The deprival reaction creates pressure to take a position 'just in case.' The discipline is to recognize that 'Too Hard' is a defensible answer when the model gives a negative IV and the moat is narrow.

Incentive bias on management commentary. Breen is incentivized to keep narrating the breakup positively until each piece is separated. Management forecasts of standalone RemainCo margins and growth are not independent estimates and should be discounted accordingly.

Net: at least four biases are pulling toward 'own it' and the only bias pulling toward caution is the explicit Munger framework. The framework wins on this one.

10-Year Outlook

The ten-year-out test asks whether the same fundamental business will exist, with a wider moat, a larger customer base, and higher profit per customer. For DuPont in 2026 the honest answer is: I do not know which DuPont we are even talking about.

In ten years, the entity at this ticker symbol will be one of three things: (a) the Industrials RemainCo trading on its own, after the Water/Healthcare spin completes in 2026-2027; (b) something acquired by a larger industrial conglomerate (3M, Honeywell, Carlisle); or (c) further broken up into single-product specialty chemical pure-plays. The same-fundamental-business test fails immediately because the consolidated DD of 2026 is not what the holder of DD shares in 2036 will own.

If we narrow the question to the most likely RemainCo (aramids, Tyvek, mobility, films), the picture is still hard. Customer base in ten years: probably similar in count, larger in dollars, with continued GDP-plus growth in defense/aerospace and GDP-or-slightly-below growth in construction. Profit per customer: flat to slightly up if PFAS-reformulation costs are absorbed and pricing discipline holds. Moat: probably narrower than today on a 10-year arc as carbon-fiber substitutes mature, Chinese aramid producers improve, and PFAS regulation forces ongoing reformulation cycles.

The single biggest threat is litigation tail risk, not operational. PFAS, like asbestos and opioids before it, has shown the pattern of multi-decade reserve true-ups that exceed initial bookings by 3-5x. The cost-share with Chemours and Corteva is the second-derivative risk: if either of those entities runs into financial distress, joint-and-several liability mechanics could re-attach to DuPont in unbounded ways.

The step-4 circle-of-competence test does not fail outright — specialty chemicals can be analyzed with effort — but it sits on the boundary. The combination of mid-breakup ambiguity, environmental-litigation tail, and zero-base ROIC in the consolidated history puts honest confidence at LOW.

CONFIDENCE: low