Air Products and Chemicals, Inc.（APD） · Diversified Industrials

Air Products: A Long-Term Value Study of Industrial Gases

May 23, 2026·~23 min read

Quick ReadPlain-language overview · read this first

Air Products is one of the global big three in industrial gases, alongside Linde, Air Liquide, and Messer; more than 90% of revenue comes from supplying oxygen, nitrogen, argon, and hydrogen to refining/chemical, electronics, metals, and healthcare customers, of which about half is 15–20-year on-site long-term contracts carrying fixed monthly fees, minimum-take volumes, and indexed price escalation, with remaining performance obligations of about $26 billion; the underlying business is close to a utility-like cash machine. Rating: Watch.

The core tension is capital allocation, not the business. Over the past few years the company turned part of its steady gas platform into a big-project, long-cycle, capital-heavy energy-transition execution vehicle; in fiscal 2025 it took a one-time recognition of about $3.6 billion of project-exit costs, activist shareholder Mantle Ridge forced a leadership change, and new CEO Menezes took over to cut losses. Underlying segment earnings are actually smooth and operating cash flow is stable in the $3-billion-plus range, but 2025 capex was still as high as $5 billion and 2026 falls only to $4 billion. At the current $289, the trailing P/E is 30.6x and conservative owner earnings about 24x — an inadequate discount to Linde.

The three intrinsic-value tiers are set at conservative $190–220, fair $240–290, and optimistic $320–380; the current price sits right at the upper edge of fair value, and combined with the 10-year Treasury at 4.57% versus a dividend yield of only 2.5%, the margin of safety is inadequate. The real downside risk is renewed impairments on large projects such as NEOM, capex that won't come down, and the market compressing valuation back to the ordinary industrial-stock range — not vanishing demand — at which point a fall to $150–180 would mean roughly a 40% permanent loss of capital. The ideal buy point is $190–220; today's price does not warrant rushing into a heavy position.

Lead
Air Products is one of the global big three in industrial gases, with large FCF swings through its LNG/hydrogen project transition; at $289.47 the stock sits inside the fair-value band, and its 30.6x P/E pays at once for both a good business and a successful turnaround, leaving too little discount. Rating: Watch.

Full report

Prices in the article are as of publication; see the valuation band above for the live price.

Conclusion First

Investment rating: Watch

Core judgment: Air Products and Chemicals is, at its core, still a highly understandable, demand-stable, structurally excellent industrial-gas business. Over 90% of the company's revenue comes from industrial gases, and about half comes from long-term on-site contracts that typically run 15–20 years and carry fixed monthly fees, minimum-take volumes, and indexed price-escalation clauses. As of the end of fiscal 2025, the company's remaining performance obligations were about $26 billion, roughly half of which is expected to be recognized over the next five years — and this excludes new on-site projects not yet in service. In other words, the underlying business has a "utility-like" resilience.

But from an "owner of the enterprise" perspective, today's APD is not a simple case of a good company on sale cheap. The problem lies not in the core industrial-gas business but in the fact that capital allocation over the past few years has turned part of what was a steady gas platform into a big-project, long-cycle, capital-heavy energy-transition execution vehicle. In fiscal 2025, because of project exits and asset actions, the company booked about $3.6 billion of project-exit costs and $3.75 billion of business and asset action charges, and in the same year experienced a proxy contest, a CEO change, and a board reshuffle. These actions show that the problem has been formally acknowledged, and that the repair is still in progress.

As of the U.S. market close on May 22, 2026, APD traded at about $289.47, with a market cap of roughly $64.5 billion and a trailing P/E of about 30.6x. Judged by the standard of a balanced-to-conservative value investor holding for 10 years or more, this price looks more like paying for "a good underlying business AND a successful turnaround" at the same time, rather than exploiting a market error. The margin of safety is not obvious.

Whether there is a margin of safety at the current price: not obvious

Suitable investor type: Better suited to long-term value investors willing to track capital allocation over time and to accept project-execution uncertainty; less suitable for conservative ordinary investors who simply want "a stable dividend utility substitute."

Biggest uncertainties: First, whether the eventual returns on large clean-energy projects such as NEOM, Louisiana, and Alberta will be enough to compensate for the enormous upfront capital outlays. Second, whether new management can re-anchor the company on "high-return incremental capital in industrial gases" rather than "high-narrative megaprojects." Third, whether the current valuation has already over-reflected the turnaround expectation.

In this report, I try to distinguish four categories of information: Fact (from company disclosures and the latest market data), Inference (a logical extension of facts), Assumption (valuation-model inputs), and Opinion (the investment conclusion).

Business, Industry, and Moat

How the business actually makes money

Fact. APD is a leading global industrial-gas company serving refining, chemicals, metals, electronics, manufacturing, healthcare, food, and other industries, primarily selling oxygen, nitrogen, argon, hydrogen, helium, carbon dioxide, carbon monoxide, syngas, and specialty gases; it also sells related equipment and develops, builds, owns, and operates large clean-hydrogen projects. In fiscal 2025, the regional industrial-gas businesses accounted for more than 90% of consolidated sales.

Fact. APD's two main gas-supply models are on-site and merchant. On-site serves large customers, typically by building a plant near the customer or connecting to a pipeline network, with contracts generally running 15–20 years, carrying fixed monthly fees and/or minimum-take requirements, and often with price-escalation clauses tied to external indices; this business is about half of total sales. Merchant is mainly bulk liquid and packaged (cylinder) gas, with contract terms usually within 5 years, slightly less recurring but highly diversified. The fiscal 2025 revenue mix was: on-site $6.18 billion, 52%; merchant $5.34 billion, 44%; equipment sales $520 million, 4%.

Fact. The company explicitly discloses that no single customer accounts for more than 10% of consolidated sales; its customer base is diversified, and the business spans about 50 countries and territories. As a complement to predictability, at the end of fiscal 2025 APD's remaining performance obligations were about $26 billion, of which the company estimates roughly half will be recognized over the next five years and the rest over a longer horizon.

Inference. The "base layer" of this business is very easy to understand: continuously supplying irreplaceable molecules to industrial settings, monetized through long-term contracts, adjacent facilities, pipeline networks, and reliability charges. What really raises the bar for understanding is not oxygen, nitrogen, or argon themselves, but the portfolio of large clean-energy projects layered on in recent years. In other words, the base industrial-gas business is highly understandable, but complexity in capital allocation at the corporate level has risen markedly.

Opinion. If the stock market closed for 5 years, I would be happy to hold APD's core industrial-gas business; but I would not be willing to hold, at any price, today's APD — a blend of a high-quality gas platform and capital-heavy energy-transition projects. The deciding factor is not the quality of the business, but the purchase price and project discipline.

Business understandability score: 4/5.

Industry structure and competitive landscape

Fact. The industrial-gas industry is not a "high-tech, explosive-growth arena"; it is closer to "mature oligopoly + stable demand + cyclical disturbance." In its 10-K, APD directly lists its main global competitors as Linde, Air Liquide, and Messer, and explicitly states that competition is mainly based on price, supply reliability, and application-development capability; the company especially emphasizes that regions with a pipeline network enjoy an economical and reliable supply advantage.

Fact. The industry's underlying demand comes from basic industrial activity such as manufacturing, refining/chemicals, electronics, metals, food, and healthcare, and will not suddenly disappear because of shifting consumer preferences — but it is affected by the macro-industrial cycle, energy prices, regional manufacturing cycles, and regulatory policy. APD's fiscal 2026 Q1 and Q2 disclosures also show that some European end-markets are still under high-cost and import pressure, and helium prices were relatively weak year-over-year in fiscal Q2 2026, but growth in on-site volumes and new electronics and aerospace projects provided an offset.

Fact. The operating quality of the industry leaders is generally very strong. Linde's 2025 sales were about $33.99 billion, operating cash flow about $10.4 billion, adjusted operating margin 29.8%, and return on capital 24.2%. Air Liquide disclosed for 2025 that its operating margin was already above 20%, recurring net profit exceeded €3.5 billion, and operating cash flow (before working-capital changes) exceeded €6.8 billion. These comparisons show that industrial gases is an industry that can generate high-quality cash flow and high returns on capital over the long term.

Inference. This is a good industry. The question is whether APD is still the most capital-disciplined player in this good industry. Compared with Linde, APD's current strengths remain long-term contracts, customer stickiness, and a global network; its weakness is a project portfolio that has been noticeably more aggressive, more capital-heavy, and noisier in execution over the past few years.

Opinion. Framed as "a good company in a good industry, or an excellent company in a bad industry," APD clearly belongs to the former; but if we press further and ask "is it the company in this good industry most worthy of paying a high valuation premium for," I lean toward giving that title to Linde rather than today's APD.

Industry attractiveness score: 4/5.

Whether the moat is durable

Breaking down the moat, APD's advantages come mainly from the following.

Scale advantage and network density. The industrial-gas business is inherently asset-heavy, especially on-site plants, pipeline networks, on-site customer facilities, and distribution systems. The larger the scale, the stronger the equipment procurement, engineering capability, logistics scheduling, backup capacity, and delivery reliability; and once a pipeline network is formed, the cost and time to replicate it are both very high. The company explicitly writes "the competitive advantage of pipeline-network regions" into its annual report.

Switching costs. For large refining/chemical, chemical, electronics, and metals customers, gas supply is tied to continuous production, safety responsibility, process stability, and shutdown losses. In theory a supplier can be switched at contract expiry, but in reality customers do not take the risk lightly, especially when the on-site model is deeply coupled with the customer's production. APD's long-term contracts, minimum-take volumes, and price-adjustment mechanisms are themselves reflections of this structural stickiness.

Cost and pricing power. This is not the simple "whoever builds it cheaper wins" story of low-cost manufacturing. The real cost advantage comes from networked supply, pipeline density, plant reliability, and energy-cost pass-through mechanisms. The company discloses that electricity is the largest cost item in air-separation gas production, and that natural gas and hydrocarbons are important feedstocks for hydrogen and syngas production, while also disclosing that it mitigates energy and feedstock price swings through pricing formulas, surcharges, cost-pass-through clauses, and tolling arrangements. In 2022 the company also explicitly stated that pass-through clauses in on-site contracts helped it cope with the European energy-price shock.

Brand and operating capability. In B2B industrial gases, "brand" is not consumer impression but safety record, supply stability, engineering delivery, and application-development capability. The company positions itself around "innovation, operational excellence, and safety"; at the audit level, revenue recognition for complex on-site contracts is listed as a key audit matter, which indirectly shows that this kind of business requires long experience and process control rather than easily replicable commodity trading.

Weaker moat elements. Network effects are not strong, data advantages are not strong, consumer brand is not strong, and channel advantages show up mainly as industrial distribution and on-site footprint rather than traditional channel monopoly. Patents and licenses help, but are not the primary factor determining the industry landscape. What really matters is "the physical embedding of projects and molecules in geography."

Inference. APD's moat is broadly stable in the base industrial-gas business, and may even widen in certain high-end applications, electronic specialty gases, and regional networks; but at the corporate level, the "visibility" of the moat has been diluted by the execution risk of large clean-energy projects. That is, the core-business moat remains strong, while the certainty of the corporate-level moat is somewhat weaker than in the past.

Opinion. The moat is not absent; rather, "a good moat has been temporarily obscured by a poor phase of capital allocation." This is a completely different thing from "a bad business," but it raises the bar for valuation.

Moat strength score: 4/5.

Management and Capital Allocation

Fact. In January 2025, Air Products went through a proxy fight launched by activist investor Mantle Ridge; Reuters reported that Mantle Ridge won multiple board seats, and one of the core disputes was CEO succession and capital-allocation risk. On February 4, 2025, the company formally appointed Eduardo Menezes as CEO, effective February 7, succeeding Seifi Ghasemi.

Fact. After new management took over, the company in fiscal 2025 recognized about $3.6 billion of project-exit costs and a total of $3.747 billion of business and asset action charges; in the first half of fiscal 2026 it took an additional roughly $22 million of charges related to previously exited projects, and stated explicitly that the related exit activities are expected to be substantially complete within fiscal 2026, though they could still be extended or generate further costs due to future events. This approach at least shows that the company is not continuing to "drag it out and deny," but has chosen concentrated exposure and correction.

Fact. Unlike many U.S. companies, APD has for more than a decade barely used large-scale buybacks to "manage" earnings per share. The 10-K shows that although a $1 billion buyback program was authorized in 2011, the company has not repurchased shares since fiscal 2013; as of September 30, 2025, $485.3 million of authorization remained. The main uses of cash have always been dividends, project investment, and debt arrangements.

Fact. On shareholder returns, the company had by fiscal 2025 achieved 43 consecutive years of dividend increases; cash dividends in fiscal 2023, 2024, and 2025 were about $1.497 billion, $1.565 billion, and $1.584 billion respectively, and in the first six months of fiscal 2026 it paid another $797 million.

Fact. The corporate-governance framework is relatively sound. The 2026 proxy statement discloses: the CEO's stock-ownership requirement is 6x base salary, and other executives 3x base salary, to be met within 5 years of taking the position; all executive officers are currently disclosed as compliant. The company also prohibits executives and directors from hedging, pledging, or holding shares in margin accounts, and has a clawback policy.

Fact. But looking at actual direct ownership, as of October 31, 2025, CEO Eduardo Menezes's direct beneficial ownership in the proxy statement was only 751 shares; directors and executive officers as a group were deemed to hold about 1.9% of the company, the vast majority of which comes from the roughly 4.108 million shares held by Paul Hilal. This shows that management's "economic alignment" relies more on equity incentives and ownership policy than on naturally high holdings.

Inference. My judgment on current management is "two-part." For the old management era's capital allocation, I give a low score, because the huge project-exit costs of 2025 are themselves the counter-evidence; for new management's corrective direction, I lean toward a neutral-to-positive assessment, because they are at least doing three reasonable things: acknowledging the problem, cutting losses and clearing out, and lowering capex. Fiscal 2025 capex was about $5.06 billion, and the company expects it to fall to about $4 billion in fiscal 2026.

Opinion. This is not a management team I would describe as an "outstanding capital allocator," but neither is it one I would outright dismiss. More precisely, they are correcting back from a stretch of irrational capital-allocation history. This deserves continued observation, not an immediate valuation premium.

Management and capital-allocation score: 3/5.

Financial Quality and Owner Earnings

Key financial metrics

The table below presents key data on the company's disclosed basis as far as possible; the 2026E column is a range-based judgment based on data disclosed through March 31, 2026 and the company's full-year guidance, and is an Assumption, not an established fact.

Metric	2020	2021	2022	2023	2024	2025	TTM to 2026-03
Revenue ($ billion)	8.86	10.32	12.70	12.60	12.10	12.04	12.46
Operating income ($ billion, GAAP)	2.24	2.28	2.34	2.49	4.47	-0.88	2.29
Net income attributable to APD ($ billion)	1.89	2.10	2.26	2.30	3.83	-0.39	2.11
Operating cash flow ($ billion)	3.26	3.34	Data to be supplemented	3.21	3.65	3.26	4.12
Capex ($ billion, company-adjusted basis)	Data to be supplemented	Data to be supplemented	Data to be supplemented	Data to be supplemented	5.15	5.06	3.91
Dividends paid ($ billion)	Data to be supplemented	Data to be supplemented	Data to be supplemented	1.50	1.56	1.58	1.60 annualized approx.
Total debt, period-end ($ billion)	Data to be supplemented	Data to be supplemented	Data to be supplemented	Data to be supplemented	14.23	17.70	17.76
Cash, period-end ($ billion)	Data to be supplemented	Data to be supplemented	Data to be supplemented	1.62	2.98	1.86	0.95

Note. Revenue, operating income, and net income for 2020–2022 come from the 2022 10-K; 2023–2025 come from the 2025 10-K; TTM is derived by taking fiscal 2025, subtracting the first half of fiscal 2025, and adding the first half of fiscal 2026. On operating cash flow, 2020 and 2021 can be confirmed from 2021 10-K search results as about $3.26 billion and $3.34 billion respectively; the 2022 figure for this line was not directly found in a single complete official statement in this research, and is therefore marked "Data to be supplemented."

The most important thing about this table is not the "2025 loss" itself, but seeing through the accounting noise of 2024–2025. Profit in 2024 was substantially lifted by the gain on the LNG business sale, and profit in 2025 was hugely depressed by project-exit charges; focusing only on those two years' P/E, ROE, and net margin would badly distort the conclusion. More informative: the company's segment operating income for the three years 2023–2025 was $2.74 billion, $2.95 billion, and $2.86 billion respectively, so underlying earnings power is actually far smoother than GAAP net income.

How to read the financial quality

Revenue and margins. From 2020 to 2025, APD's revenue rose from about $8.86 billion to $12.04 billion, but 2023–2025 was basically flat in the $12.0–12.6 billion range, showing this is a "low-to-moderate-growth but highly resilient" business. Gross margin was compressed to 26.5% in 2022 by energy inflation, was 30.2% in 2021, and 33.9% in 2020; by 2024–2025 it returned to around 31% on the consolidated income-statement basis. Operating margin and net margin, however, were significantly distorted by the 2024 LNG sale gain and the 2025 project-exit losses.

Cash-flow quality. Although the income statement is very noisy, cash-flow resilience is clearly stronger. Operating cash flow in fiscal 2023, 2024, and 2025 was about $3.21 billion, $3.65 billion, and $3.26 billion respectively; operating cash flow for the trailing 12 months to March 31, 2026 is estimated in this report at about $4.12 billion. This shows the core business is still continuously "spitting out cash," and did not suffer an operating-level cash collapse alongside the large 2025 write-downs.

Free cash flow. Here we must strictly distinguish Fact from basis. On the strictest GAAP cash view — operating cash flow minus increases in long-term assets and related investments — APD's free cash flow over the past two years is not pretty, because capex on large projects is extremely high. The company itself discloses that fiscal 2025 capex was about $5.06 billion, and fiscal 2026 is still expected to be about $4 billion. So APD is currently not an "asset-light cash cow," but rather in a state of "very strong core-business cash that is partly consumed by heavy project investment."

Balance sheet. As of March 31, 2026, the company had cash of about $951 million, total debt of about $17.76 billion, and net debt of about $16.81 billion; on APD's shareholders'-equity basis, book net assets were about $15.65 billion. On the surface leverage is not low, but part of it is related to the NEOM green-hydrogen project, and the company explicitly discloses that NEOM VIE liabilities are non-recourse to Air Products' general credit. On the same date, VIE assets available to service debt were about $7.896 billion, and non-recourse liabilities about $5.296 billion. So from a shareholder-analysis perspective, one should look at both "consolidated leverage" and "shareholder risk after excluding non-recourse project financing."

Interest coverage and debt-service capacity. Using GAAP operating income for the trailing 12 months, interest coverage is about 10x; on an operating-income basis excluding abnormal items, it is roughly 13x. At the end of March 2026 the company disclosed that all of its debt covenants were compliant, and that it maintains a $3 billion five-year revolving credit facility and a 364-day facility, neither of which was drawn during the period. For a large industrial-gas leader, this means short-term liquidity risk is manageable — but it does not mean capital-allocation risk has disappeared.

Working capital and accounting risk. As of the end of March 2026, accounts receivable were about $1.938 billion, slightly up from $1.901 billion at end-September 2025; inventory was about $768 million, slightly down from $777 million at end-September 2025; and payables and accrued liabilities fell from $3.238 billion to $2.857 billion, which weighed on first-half operating cash flow. I see no abnormally inflated receivables or inventory, and no control deficiencies disclosed at the audit level. The accounting-sensitivity points to watch are not "signs of fraud," but revenue/cost estimates on long-term on-site contracts and equipment projects. In the 2025 10-K, Deloitte listed revenue recognition for complex on-site customer contracts as a key audit matter; in addition, the company disclosed in the first half of fiscal 2026 that estimate changes on certain equipment projects had an unfavorable impact on operating income of about $58 million. This looks more like the estimation noise inherent to project-type businesses than clear evidence of manipulation.

Financial-quality conclusion. Looking only at GAAP profit, APD appears to "swing wildly"; looking at the underlying industrial-gas operating cash flow and segment operating income, it is still a very profitable company. What really drags on "distributable cash" is not core-business deterioration, but the capital consumption of large recent projects and the correction costs of the wrong projects. This distinction is crucial.

Owner Earnings estimate

Fact. Buffett-style owner earnings do not simply equal free cash flow; they are closer to "net income attributable to shareholders + non-cash charges − maintenance capex ± working-capital normalization." For APD, the biggest difficulty is that the company does not separately disclose maintenance capex, and total capex is mixed with a large amount of growth-project spending, especially on NEOM and other clean-energy projects. The company only disclosed that fiscal 2026 capex is expected to be about $4 billion, of which roughly $1 billion will go to traditional industrial-gas projects, while still including core-business maintenance and replacement spending.

Conservative estimation assumptions. I use the following conservative assumptions: For net income attributable to shareholders, I use a level close to a normalized full-year 2026 rather than the 2025 anomaly; referencing the company's raising of full-year adjusted EPS guidance to $13.00–13.25 after fiscal Q2 2026, on about 223 million diluted shares, this corresponds to normalized net income of about $2.9–2.95 billion. Adding annualized depreciation and amortization of about $1.55–1.6 billion; estimating maintenance capex at $1.4–1.6 billion; and estimating working capital at a neutral-to-slightly-outflowing $100–200 million. This yields conservative owner earnings of roughly $2.7–3.0 billion, with a midpoint of about $2.7 billion, equivalent to about $12.1 per share. The biggest Assumption here is maintenance capex, not the profit figure itself.

Inference. If the capital invested in NEOM and other large growth projects is treated as "reinvestment" rather than spending required to keep the existing gas network running normally, then APD's true owner earnings are clearly higher than current strict-GAAP free cash flow; conversely, if all that project spending is treated as maintenance, then owner earnings are markedly depressed. To be conservative, I lean toward not adding back all the large-project spending, but I also do not accept the surface conclusion that "APD's true free cash flow is negative over the long run."

How many times owner earnings does the current valuation represent. Using the current market cap of about $64.5 billion and the conservative owner earnings of $2.7 billion above, APD currently trades at roughly 24x owner earnings. For an excellent but capital-heavy industrial company whose capital-allocation repair is still in progress, this multiple is not cheap.

Valuation and Margin of Safety

Current price and relative valuation

Fact. As of the U.S. market close on May 22, 2026, APD traded at about $289.47, with a market cap of roughly $64.5 billion and a trailing P/E of about 30.6x. Based on shares, cash, and total debt as of March 31, 2026, enterprise value is about $81.3 billion; on a rough estimate using fiscal 2025 adjusted EBITDA, current EV/adjusted EBITDA is around 16x, and would be somewhat lower after excluding the incremental net debt driven mainly by NEOM non-recourse project financing. On APD's shareholders'-equity basis, price-to-book is currently about 4.1x. These are all "not cheap" quality-stock valuations, not distressed-stock valuations.

Fact. Compared with its strongest competitor Linde, the latter currently trades at about $517.58, with a trailing P/E of about 34.3x; but Linde's 2025 adjusted operating margin of 29.8%, operating cash flow of $10.4 billion, and return on capital of 24.2% make its operating quality clearly superior to APD, which is still in a repair phase. In other words, although APD's valuation is slightly below Linde's, it is not low enough to compensate for "a weaker execution history + greater project uncertainty." Air Liquide's 2025 operating margin also exceeded 20%, again showing that APD has not been handed a meaningful "discounted turnaround price."

Inference. On a relative-valuation view, today's APD looks more like "a bit cheaper than the best peer" than "clearly undervalued." For the second-best player in a high-quality industry, this is usually not a good value-investing entry point. The truly ideal entry point is when the company's quality has not deteriorated but the market has knocked it down to a clear discount because of short-term noise. Right now looks more like the market has already priced in more than half of the turnaround expectation in advance.

Intrinsic-value estimate

Below are three methods. Note: valuation is driven by assumptions, not by wishes.

Method 1: owner-earnings discount. Using conservative owner earnings of $2.7 billion as the base, I assume three scenarios:

Conservative scenario: owner-earnings annual growth of 4% over the next 10 years, a discount rate of 9%, and a terminal growth rate of 2%. This corresponds to an intrinsic value of about $190–220 per share.
Neutral scenario: growth of 6% over the next 10 years, a discount rate of 8.5%, and a terminal growth rate of 2.5%. This corresponds to an intrinsic value of about $240–290 per share.
Optimistic scenario: growth of 8% over the next 10 years, a discount rate of 8%, and a terminal growth rate of 3%. This corresponds to an intrinsic value of about $320–380 per share.

The core logic of these ranges is not "can APD keep growing," but "can it truly convert large-project spending into high-quality owner earnings rather than remaining stuck long-term in a capital-heavy story." The basis for the input assumptions includes: the underlying business supported by long-term contracts, the current expectation of falling capex, the 2026 normalized EPS guidance, and the long-term high-margin characteristics of the industry leaders.

Method 2: relative valuation. APD currently trades at about 30.6x trailing P/E, Linde at about 34.3x. On the surface APD is slightly cheaper, but Linde's margins, cash flow, and return on capital are significantly stronger, so this "slightly below the leader" valuation does not automatically constitute cheapness. On owner earnings, APD currently trades at roughly 24x; on price-to-book, about 4.1x, with no asset discount. The relative-valuation conclusion is: reasonable, but not cheap; more precisely, "a fair-to-expensive price for a quality-repair stock."

Method 3: asset or liquidation value. This is not the most useful method for APD. On the books, as of the end of March 2026, APD's shareholders' equity was about $15.65 billion, and after deducting goodwill and intangibles, rough tangible net worth was about $14.4 billion, equivalent to about $65 per share; but industrial-gas assets are highly specialized, and on-site plants, pipeline networks, and on-site facilities are usually worth far less in liquidation than in continued operation, so book value does not represent a "floor price." The only thing worth viewing positively is that the large NEOM-related liabilities are largely non-recourse, reducing part of the direct harm of surface leverage to the parent's shareholders. What the asset method tells me is not "cheap," but "the floor is not as high as one might imagine."

Valuation conclusion and price ranges

Combining the three methods, I make the following judgment:

Valuation range	Price per share
Conservative intrinsic-value range	$190–220
Fair intrinsic-value range	$240–290
Optimistic intrinsic-value range	$320–380

Conclusion. At the current level of about $289.47, APD sits roughly at "the upper edge of the fair-value range" as I define it. If you take a more optimistic view of new management and large-project returns, today's price can barely be justified; but if what you require is a Buffett-style margin of safety, this price is hard to call comfortable.

I would divide the price bands as follows:

Ideal buy range: $190–220. In this range, even if the repair is slower than expected and growth is only low-to-mid single digits, the long-term return is more acceptable.
Acceptable holding range: $220–290. This is more like "holding excellent underlying assets" than "building a new high-odds position."
Clearly overvalued range: above $330. This is already paying for a more optimistic realization of projects and repair of returns on capital.

Whether the margin of safety is adequate

My conclusion: the current margin of safety is inadequate.

The most fragile valuation assumptions are not "whether industrial-gas demand will disappear," but the following three: first, that the large projects will ultimately convert into high-quality cash flow; second, that capex will fall from about $5.06 billion in 2025 to about $4 billion in 2026 as management says, and continue to decline; third, that the market will keep granting APD a quality valuation close to that of a top-tier leader. If any one of these fails, the current pricing will look elevated.

From an opportunity-cost perspective, the current U.S. 10-year Treasury yield is about 4.57%, while at the current price APD's dividend yield is only about 2.5%, its trailing earnings yield only about 3.3%, and even on my conservative owner-earnings basis the yield is barely over 4%–5%. That is, you are not getting an equity risk premium that is very friendly to conservative capital today.

So this is very much a classic case of "a good company, but not necessarily a good price."

Risks, Comparison, and Final Recommendation

Key risks and the strongest bear case

The most important risk is not short-term price volatility, but permanent loss of capital.

The first type of risk is capital-allocation and project-execution risk. The roughly $3.6 billion of project-exit costs in fiscal 2025 already told investors: not every energy-transition narrative converts into shareholder returns. If follow-on projects such as NEOM, Louisiana, and Alberta again see budget inflation, downward return revisions, delayed start-ups, or customer/policy changes, shareholders may well keep paying for poor capital allocation.

The second type of risk is "the second-best executor in a good industry" risk. The industrial-gas industry is excellent, but compared with Linde and Air Liquide, APD has been less stable in capital discipline and project selection in recent years. The strongest bear case is actually very simple: APD is a company that could have compounded as steadily as Linde, but was dragged by the past few years' high-risk projects into a stock with a turnaround story; and the market's current pricing does not give enough discount for the turnaround failing.

The third type of risk is cash-flow and valuation mismatch risk. After book profit recovers, investors may mistakenly think "the problem is over"; but if growth continues to depend on heavy capital investment rather than on cash naturally released by the existing gas network, truly distributable cash flow will still lag. If the market then compresses valuation from 30x earnings to a more ordinary 18–22x, then even if the company is not a bad company, it will inflict long-term return damage on new buyers.

The fourth type of risk is accounting-estimate and contract-complexity risk. I see no obvious signs of financial fraud; the company disclosed effective controls at both the 2025 year-end and the end of fiscal Q1 2026. But revenue recognition on complex on-site contracts is a key audit matter, and equipment-project cost estimates also affect profit, which requires investors to keep watching project cost updates, collections, performance, and impairments.

Which facts would overturn the investment logic. If there is another large project exit or impairment in the next 12–24 months; if capex does not fall while owner earnings continue to be consumed; if adjusted operating profit growth stalls while net-debt pressure keeps rising; if the quality of on-site contract renewals deteriorates or cost pass-through can no longer be maintained; if dividend coverage starts to depend on adding leverage — I would lean toward admitting the judgment was wrong.

Comparison with other opportunities

Comparison with the strongest peer competitor. If I could only choose between APD and Linde today, I would admit APD is slightly cheaper on valuation, but not cheap enough to cover the quality gap. Linde currently has stronger operating quality, return on capital, and cash-collection ability, while APD is still repairing the project legacy left by its predecessor.

Comparison with a broad index. Over the five-year cumulative return through September 2025, APD significantly lagged the S&P 500: APD went from about 100 to 141, while the S&P 500 reached 247. The past does not determine the future, but it reminds investors: even in a good industry, if capital allocation goes astray, a stock can underperform the index for a long time. And if the current buy price is not cheap either, then "actively choosing APD" must rest on a sufficiently high conviction that the repair will succeed. For most investors, the index may be the more worry-free choice.

Comparison with the risk-free rate. The current 10-year Treasury yield is about 4.57%. If APD's conservative long-term annualized return is only 4%–6%, then the operating, execution, policy, and valuation risks you bear may not be adequately compensated. Only at a lower price would APD's expected return/risk ratio improve significantly.

If I could only hold 5 assets, does it qualify for the portfolio. At the right price, I think it qualifies; at today's price, I lean toward reserving the slot for more certain or cheaper opportunities.

Investment checklist

Check item	Conclusion
Can I understand this business?	Pass
Does it have long-term stable demand?	Pass
Does it have a durable moat?	Pass
Does it have pricing power?	Pass
Can it generate stable free cash flow?	Uncertain
Is its return on capital excellent?	Uncertain
Is management trustworthy?	Uncertain
Is capital allocation rational?	Fail
Is the balance sheet sound?	Pass
Is valuation below intrinsic value?	Fail
Is the margin of safety sufficient?	Fail
Does long-term holding put me at ease?	Uncertain
Which key facts would make me sell?	Large re-impairments, capex that won't come down, owner earnings that don't improve, deteriorating dividend coverage
Am I buying only because of a rising price or emotion?	Should be actively wary

The core tension of this checklist is clear: the business and industry pass most tests, while price and capital allocation do not.

Final investment conclusion

【Final Rating】 Watch

【One-Sentence Investment Thesis】 APD is an excellent industrial-gas business, but at the current price, what you are buying is not just a stable molecule business — you are also buying a large-project turnaround story that has neither been fully disproven nor fully delivered.

【Core Bull Case】

The industrial-gas core business is understandable, demand is stable over the long term, customers are diversified, and no single customer exceeds 10% of sales.
More than 90% of revenue comes from industrial gases, about half from 15–20-year on-site long-term contracts, with strong revenue recurrence and predictability.
In regions with pipeline networks it has hard-to-replicate scale and network advantages, and the industry is a global oligopoly.
The operating-cash-flow base is still steady, at basically over $3 billion in fiscal 2023–2025, with TTM further recovering to over $4 billion.
New management has at least begun to acknowledge the problem, cut project losses, and lower capex.

【Core Bear Case】

The huge project-exit losses in fiscal 2025 show that past capital allocation had major missteps.
Large clean-energy projects still tie up substantial capital, and the recovery of truly distributable cash flow remains to be verified.
The current roughly 30.6x trailing P/E and roughly 24x conservative owner earnings are not cheap.
Compared with Linde, APD is inferior in quality but has a limited valuation discount.
For conservative investors, the current equity risk premium relative to 10-year Treasuries is not compelling.

【Key Assumptions】

Capex declines gradually in 2026–2028 as management expects.
NEOM and other large projects do not see large exits and impairments like those in fiscal 2025 again.
The core industrial-gas business maintains its current pricing power, contract-renewal rates, and cost pass-through ability.
New management sustains the capital discipline of "cut losses + focus on high-return projects."

【Fair Buy Price】 $190–220. The basis is a conservative owner-earnings discount and the margin of safety demanded for project-execution risk, not a simple application of peer multiples.

【Target Holding Period】 At least 5–10 years; but on the precondition of buying at a better price and subsequently tracking confirmation that capital allocation is truly repaired.

【Expected Annualized Return】

Conservative scenario: 4%–6%
Neutral scenario: 7%–9%
Optimistic scenario: 10%–12%

【Maximum Loss Risk】 If there are multiple more rounds of project impairment, capex is hard to bring down, and the market pulls the valuation back to a more ordinary industrial-stock range, the price could return to around $150–180, implying roughly 38%–48% downside for current buyers; in a worse scenario, add years of opportunity cost on top. This is the real permanent-loss-of-capital risk.

【Tracking Metrics】 What to watch most in the future is not quarterly EPS, but the following: the actual path of capex decline; whether owner earnings/operating cash flow keep improving; adjusted operating margin; changes in net debt, especially after excluding non-recourse project financing; the quality of newly signed on-site projects; renewals and price adjustments on existing on-site contracts; milestone and return updates on projects such as NEOM/Louisiana/Alberta; whether large business and asset action charges recur; dividend coverage; and whether the board's and management's statements on capital discipline are consistent.

【Signals That Trigger Re-evaluation】 Another $1-billion-scale project impairment or exit; capex above guidance without corresponding high-return evidence; a stall in TTM operating-cash-flow improvement; continuously rising net debt; a decline in core industrial-gas segment operating income; management shifting back toward a "big-and-comprehensive" project narrative; dividend growth needing more debt to support it.

【Final Recommendation】 Put plainly, APD deserves a place on a high-quality watchlist, but does not deserve rushing into a heavy position at today's price. If your goal really is to own it long-term the way you would acquire a company, then you should care more about "at what price you own a good business" than about "whether the company's story has started to improve." APD's business foundation is still excellent; unfortunately, the current price does not leave a conservative long-term owner enough room to be wrong.

Open questions and limitations

This research still has three points that need ongoing supplementation. First, the company does not separately disclose maintenance capex, so owner earnings can only be estimated conservatively. Second, Air Liquide and APD use different accounting bases, so directly comparing cross-company EV/EBITDA and ROIC requires more granular, unified adjustments. Third, all the cash-flow line items for 2020–2022 were not fully retrieved in this round of official-source searching, so parts of the historical tables use a "Data to be supplemented" note rather than conjecture.

This report is based on public information and does not constitute investment advice. Markets carry risk; invest with caution.

Baillie Framework · Ten Questions for Growth Investing

Hunting ten-year five-baggers among great growth stocks — pressing the upside question: "Can it get much bigger?"

How high is its market ceiling? Is it expanding an existing pie, or creating an entirely new market?5/10

Conclusion first: APD is mainly "expanding an existing, mature pie" rather than creating an entirely new market. Industrial gases is a century-old industry; demand comes from basic industrial activity such as refining/chemicals, chemicals, metals, electronics, food, and healthcare, and globally it is already a stock-market competitive landscape carved up among a few oligopolists — Linde, Air Liquide, APD, and Messer. Its ceiling is set by "total global industrial activity × gas-outsourcing penetration," a low-to-mid single-digit grower over the long run, and by no means the kind of "exponentially opening a new continent" story that Baillie Gifford favors.

The true scale of the existing pie can be seen from the leaders' size: industry number one Linde had full-year 2025 sales of about $34 billion, Air Liquide's 2025 recurring net profit was about €3.68 billion, while APD's own fiscal 2025 sales were about $12.04 billion. The three combined amount to only tens of billions of dollars, with relatively stable shares among them, which means APD's growth comes more from "building one more air-separation plant next to an existing customer, signing one more on-site long-term contract, grabbing a bit of share from a rival" than from opening up new zero-to-one demand. The report's judgment is consistent: the industrial-gas industry "is not a high-tech, explosive-growth arena; it is closer to mature oligopoly + stable demand + cyclical disturbance."

The only thing with a "new market" flavor is the clean-hydrogen/energy-transition line (large projects such as NEOM, Louisiana, and Alberta). In theory, green hydrogen could open an entirely new, potentially enormous decarbonized-fuel market — which is exactly the growth narrative the company told over the past few years. But to be honest: this "new market" has so far not converted into high-quality returns, and has instead become a drag. The company in fiscal 2025 booked about $3.7 billion of pre-tax business and asset action charges due to project exits (about $3 billion after tax, or $13.68 per share), and under the new CEO exited a batch of clean-energy projects. In other words, the company has itself admitted that this "new market's" ceiling cannot support its upfront investment at the current stage, and has voluntarily retreated to its core business of "expanding the existing industrial-gas pie."

The ceiling also has a structural floor supporting it, keeping it from shrinking: at the end of fiscal 2025, the company's remaining performance obligations were about $26 billion, roughly half of which is expected to be recognized over the next five years (per the report, citing the 2025 10-K), and about half of revenue comes from 15–20-year on-site long-term contracts, with no single customer exceeding 10% of consolidated sales. This makes APD's "pie" very stable and highly predictable — but that very "stability" is precisely why it is a utility-like stock business rather than a high-slope emerging market.

One-line characterization: the ceiling is high enough to let a good business exist steadily over the long term, but not enough to support a "5x-in-ten-years" market-size imagination — it is fighting for share within a mature pie, and the green-hydrogen slice that was meant to be a new pie has instead been self-disproven and voluntarily scaled back. On the Baillie Gifford LTGG framework, this dimension is not outstanding.

Jun 11, 2026
Can its revenue at least double over the next five years? Is growth driven mainly by volume, price, or new business?3/10

Conclusion first: doubling revenue over the next five years is almost impossible; the gap is enormous. APD is a low-to-mid-growth, mature industrial-gas business; revenue over the past three years has been basically flat in the $12.0–12.6 billion range, and growth comes mainly from "price (price increases + energy-cost pass-through) + a little volume (new on-site projects coming online)" rather than from a new-business explosion that could support a doubling. Baillie Gifford LTGG's "double revenue in five years" bar clearly does not hold here.

First, the actual slope of revenue. APD's fiscal 2025 sales were about $12.04 billion, down about 1% year-over-year — the official breakdown is "volume down 4%, partly offset by about 2% energy-cost pass-through and about 1% price increases." That tells us directly: the current growth engine is price, not volume, and volume is still contracting. The report's data point the same way: 2023–2025 revenue is stuck in the $12.0–12.6 billion range, a textbook case of "low-to-mid-growth but highly resilient." To double in five years to $24 billion would require about a 15% compound annual growth rate, which for a mature-oligopoly gas business is fanciful — even industry leader Linde's 2025 sales grew only about 3% year-over-year.

Now the driver structure (volume vs price vs new business):

Price: the current main engine, but with a limited ceiling. On-site long-term contracts carry indexed price-escalation clauses, the merchant side raises non-helium prices, and together with energy-cost pass-through this can contribute a steady single-digit price increment; but price increases are essentially tracking inflation and energy, not a sustainably scalable growth lever. In fiscal Q2 2026 helium prices were still about a 4% year-over-year headwind, so even the price side is not a one-way tailwind.

Volume: relying on slow ramp-ups as new on-site plants come online, a structurally slow variable. New signings require 15–20-year long-term contracts, build-then-ramp, so the incremental volume contributed each year is limited, and total fiscal 2025 volume was still negative.

New business: this was supposed to be where green hydrogen/clean energy provided the doubling imagination, but the company just voluntarily scaled it back. In fiscal 2025 it exited a large batch of clean-energy projects and booked about $3.7 billion of pre-tax charges, and cut fiscal 2026 capex guidance sharply from the prior year to about $4 billion — meaning the company is hitting the brakes, not the accelerator to grow new-business revenue.

Worth crediting: the repair slope on the profit side is better than on the revenue side. Fiscal Q2 2026 adjusted EPS was $3.20, up about 19% year-over-year, and the company raised full-year adjusted EPS guidance to $13.00–13.25 (corresponding to about 8%–10% year-over-year growth). But note this is "profit repair" (via price increases, productivity, portfolio optimization, and cost discipline), not "revenue expansion," and certainly not a doubling of revenue — the earnings improvement comes mainly from operating efficiency and portfolio actions, while the denominator (revenue) itself is still in the slow-growth zone.

Qualitative conclusion: on Baillie Gifford's "double revenue in five years" standard, APD does not qualify, and misses by a wide margin; its growth is a low-to-mid single-digit story of "price-led, volume-secondary, new business shrinking." What is truly improving is the normalization and repair of earnings per share, not high-speed revenue growth. This dimension should score markedly low.

Jun 11, 2026
Five years from now, what will take the baton as the next growth engine? Does this "second curve" exist today?4/10

Conclusion first: APD does not today have a clear, deliverable "second curve." The clean-hydrogen/energy-transition line that was supposed to take the baton is precisely the piece the company just voluntarily scaled back and cut losses on; over the visible next five years, the real growth engine is still the extension of the core industrial-gas business (new on-site projects + electronic specialty gases + regional networks), which looks more like "a natural extension of the same main curve" than an independent new growth pole. To Baillie Gifford's question of "does the second curve already exist today," the answer is somewhat negative.

First, the disproven curve. The growth story the company told over the past few years was exactly green hydrogen — large clean-energy projects such as NEOM, Louisiana, and Alberta, which in theory could open up the entirely new market of decarbonized fuel and could have become a second curve. But in fiscal 2025 the company booked about $3.7 billion of pre-tax business and asset action charges due to project exits (about $3 billion after tax), exited a batch of clean-energy projects under the new CEO, and cut fiscal 2026 capex guidance sharply to about $4 billion. The report's judgment is consistent: this business, meant to be a "new market," has so far not converted into high-quality returns and has instead become a drag. So strictly speaking, APD does not "own a validated second curve"; it has "just cut off a second curve that failed to work."

So what is actually taking the baton? A gradual extension of the core business, not a new species:

Electronics and specialty gases. Semiconductor/electronics end-markets are a higher-growth, higher-value-added niche within industrial gases, and the report mentions that new projects in electronics and aerospace provide an offset. This is quality incremental growth, but it still belongs to "selling industrial gases to higher-end customers," a high-end extension of the main curve, not an independent second curve.

New on-site project ramp-up. Remaining performance obligations of about $26 billion, roughly half to be recognized within five years (per the report, citing the 2025 10-K), plus new projects not yet in service, will continue to convert into revenue. But this is essentially the realization of the stock of the core on-site model — "doing more of the same business."

Surviving, disciplined low-carbon projects. The company has not entirely abandoned hydrogen; it has converged from a "big-and-comprehensive narrative" to "focusing on high-return projects." If some future green-hydrogen/low-carbon project truly delivers meaningful returns, it could grow again into a second curve — but this is a "possibility to be verified," not "a fact that exists today."

Placing it against a Baillie Gifford lens makes it clearer: the second curve Baillie Gifford wants is a new engine like "cloud is to Amazon, or data centers are to Nvidia" — one whose embryo is already visible today and can independently support market cap in the future. APD has nothing like that in hand today — what it can see is doing its mature core business more steadily, at a higher end, with more capital restraint. This is a healthy continuation of a good business, not high-slope regrowth.

Qualitative conclusion: the second curve "does not exist today (already cut) / must regrow in the future (to be verified)." Growth hand-off is more likely to come from core-business extension than from an independent new pole, which is a clear demerit on a growth-investing framework and also echoes the report's "Watch" rating — what you are buying is a good business in repair, not a growth stock holding a second growth engine.

Jun 11, 2026
What is its core competitive advantage? Will this moat widen or narrow over the next three to five years?6/10

Conclusion first: APD's core moat is real and durable — the high switching costs of long-term on-site contracts, regional pipeline density, scale, and operating reliability together form "the physical embedding of molecules and projects in geography." Over the next three to five years, this moat in the base industrial-gas business is broadly stable and may even widen in places (electronic specialty gases, regional networks); but the "visibility" of the corporate-level moat has in recent years been diluted by the execution risk of large clean-energy projects — the core is strong, while the certainty of the company as a whole is somewhat weaker than in the past.

Several pillars of the moat all have factual support:

Switching costs (the hardest pillar). About half of revenue comes from 15–20-year on-site long-term contracts carrying fixed monthly fees, minimum-take volumes, and indexed price-escalation clauses (per the report, citing the 2025 10-K). On-site plants are built at the customer's site or connect to a pipeline network, deeply coupled with the customer's continuous production, safety responsibility, and process stability, so switching suppliers means risking shutdowns and process disruption. The objective corroboration is the financial result of this stickiness: fiscal Q2 2026 adjusted operating income was $753 million, up about 19% year-over-year, and adjusted EPS grew about 19% year-over-year; holding earnings steady through price increases and renewals against a backdrop of unimpressive volume shows that contract stickiness and pricing power are still there.

Scale and network density. Industrial gases is inherently asset-heavy; once a pipeline network is formed, the cost and time to replicate it are both extremely high, and the company writes "the competitive advantage of pipeline-network regions" into its annual report (per the report). Customers are highly diversified, no single customer exceeds 10% of consolidated sales, and the business spans about 50 countries (per the report, citing the 2025 10-K), which reduces the impact of losing any single customer.

Cost pass-through and reliability. Electricity is the largest cost in air separation and natural gas is the main feedstock for hydrogen production; the company mitigates energy swings through pricing formulas, surcharges, cost-pass-through clauses, and tolling arrangements (per the report). This "energy-cost pass-through" mechanism lets it hold margins even under inflation and energy shocks, and is part of the moat.

But two weaknesses must be honestly flagged:

First, the moat type leans "stable" rather than "wide." Network effects, data advantages, and consumer brand are all weak — it is not the kind of winner-take-all moat that grows stronger as users grow, but rather "a steady-state oligopoly where each guards its region and charges on reliability." This can defend and compound, but does not easily open a fast gap over rivals.

Second, the strongest rivals "run the same moat better." Linde's 2025 adjusted operating margin was 29.8%, operating cash flow about $10.4 billion, with industry-leading return on capital (about 25.7% in Q1), and Air Liquide's 2025 operating margin was about 20.7%. With the same good moat, Linde converts it into a higher return on capital; APD in recent years has had the "certainty" of its moat discounted by aggressive projects. The report's judgment is accurate: the moat is not absent — rather, "a good moat has been temporarily obscured by poor capital allocation."

Will it widen or narrow over the next three to five years? Most likely "the core stays stable with a slight widening, and the corporate level recovers via repair": electronic specialty gases and regional network density have room to widen; as the company exits problem projects and pulls capex back from its high (fiscal 2026 guidance of about $4 billion), the "visibility" of the corporate-level moat should gradually recover. But as long as the returns on the legacy green-hydrogen projects remain unproven, the certainty of the overall moat is still discounted.

Qualitative conclusion: moat strength lands in the "real, sustainable, well-defended" range, one of the few dimensions in this report worth a higher score; but it is a defensive rather than an expansionary moat, and rivals run the same moat more excellently — enough to support the "good business" judgment, not enough to support a "great growth stock deserving a high premium" narrative.

Jun 11, 2026
If the core business were disrupted, does it have the genes to reinvent itself? How does it treat mistakes and bad news?4/10

Conclusion first: the probability of APD's core industrial-gas business being "disrupted" is already very low (it sells irreplaceable industrial molecules, embedded via long-term contracts and physical assets); and it recently happened to provide a ready-made stress-test answer sheet on "how to treat mistakes and bad news" — facing a capital-allocation disaster of its own making, the company chose to change leadership, admit fault, expose everything at once, and cut losses, rather than delaying and covering up. This "reinvention gene" is neutral-to-positive, but what it proves is "the ability to correct errors," not the disruptive-rebirth ability of "creating a new lifeline after the core is disrupted" in the Baillie Gifford sense.

First, address the implicit premise — the reality of "the core business being disrupted." Industrial gases sells basic molecules like oxygen, nitrogen, argon, hydrogen, and helium, with demand from evergreen industrial activity such as refining/chemicals, metals, electronics, food, and healthcare, which will not disappear because of a sudden shift in consumer preference (per the report); the on-site model locks in customers via 15–20-year long-term contracts and on-site physical assets. This kind of business almost has no "overnight disruption by new technology" script; the real risk is being dragged down by its own capital allocation rather than being disrupted — which is exactly what happened over the past few years.

Now look at how it treats mistakes and bad news, the most informative behavioral evidence from this company recently:

It did not cover up; it exposed everything at once. Facing the mess left by the previous stage's green-hydrogen bet, new management took a one-time charge of about $3.7 billion of pre-tax business and asset action charges in fiscal 2025 (about $3 billion after tax, or $13.68 per share), causing a full-year GAAP operating loss of $877 million, or a loss of $1.74 per share. Clearing bad assets in one shot, taking a huge book loss rather than hiding the problem, is a posture of "admitting fault" rather than "dragging it out and denying." The report's assessment is consistent: the company chose concentrated exposure and correction.

It changed people and reformed governance. The root of the worst news was the predecessor's capital allocation, and the company's response was a thorough reset at the governance level — in early 2025 activist investor Mantle Ridge launched a proxy fight and won multiple board seats, after which the company appointed Eduardo Menezes to replace Seifi Ghasemi as CEO (per the report). This shows the error-correction mechanism (board + external shareholder pressure) is working.

It adjusted behavior, not just words. The most convincing thing is the hard-dollar pivot, not slogans: fiscal 2026 capex guidance was cut sharply from the prior year's high to about $4 billion (about $1 billion less than the year before), converging from a "big-and-comprehensive project narrative" to "focusing on high-return projects." After cutting losses the profit side did begin to repair: fiscal Q2 2026 adjusted EPS grew about 19% year-over-year, and full-year guidance was raised to $13.00–13.25.

But the boundaries must be honestly flagged:

First, this mistake is itself evidence that the company's "capital-discipline gene" once failed — it took external activist pressure plus a huge write-down to complete the correction; this is not a company with "a strong built-in error-correction culture that self-corrects at the first sign of trouble." The report's low assessment of the old management era's capital allocation stems exactly from this.

Second, "error-correction" does not equal "reinvention." What Baillie Gifford asks is whether a new lifeline can grow after the core is disrupted (like Amazon from e-commerce to cloud). What APD demonstrates is contraction-style correction of "cutting a new business that failed and retreating to the core," not expansion-style rebirth of "creating a second growth pole." It proved downside protection (it won't infinitely amplify mistakes), but did not prove upside regeneration ability.

Qualitative conclusion: on the self-reinvention/error-correction dimension, APD's answer sheet is "adequate, candid, action-backed" — its way of handling bad news is commendable and clearly better than "delay-and-cover-up" management; but this is the maturity of "being able to stop the bleeding and correct course," not the disruptive gene of "rising from the ashes after the core is disrupted." For the Baillie Gifford framework it is neutral and not a bonus highlight, consistent with the report's judgment that "new management deserves continued observation, not an immediate premium."

Jun 11, 2026
Does management (especially the founder) have a long-term vision, with interests deeply aligned with the company? Are they willing to sacrifice current profit for five-to-ten years out?4/10

Conclusion first: APD is a professional-manager company with no founder; management's interest alignment relies mainly on ownership policy and equity incentives rather than naturally high holdings, so absolute economic alignment is on the weak side; on long-term vision, the company just went through a cautionary counter-example of "over-sacrificing current discipline for a long-term narrative" (the green-hydrogen gamble), and new management is now swinging the pendulum back toward "capital discipline." On balance, this dimension is neutral and does not constitute the "founder-style deep alignment + willingness to sacrifice the present for ten years out" strength that Baillie Gifford favors.

First, the hard facts on "alignment," the part that most warrants checking disclosures online rather than going by impression:

The new CEO's direct holding is minimal. As of October 31, 2025, CEO Eduardo Menezes's direct beneficial ownership in the proxy statement was only 751 shares (per the report, citing the 2026 proxy statement); directors and executive officers as a group were deemed to hold about 1.9% of the company, the vast majority coming from the roughly 4.108 million shares held by director Paul Hilal (Mantle Ridge). In other words, stripping out this activist-investor shareholder, management's own economic alignment is quite thin — far from the "founder/management with their net worth staked in the company" that Baillie Gifford favors. At APD's current market cap of about $63 billion (early-June 2026 price around $282), the absolute value of 751 shares is essentially negligible for a large-company CEO.

Alignment rests on rules rather than nature. The company has ownership requirements (6x base salary for the CEO, 3x for other executives, to be met within 5 years, all currently disclosed as compliant), and prohibits executives/directors from hedging, pledging, or margin-holding shares, with a clawback policy (per the report). This governance framework is sound and can force a degree of alignment, but it is essentially "using rules to tie interests on," not "management already treating themselves as owners." The report's judgment is consistent: executives' economic alignment relies more on equity incentives and ownership policy than on naturally high holdings.

Now the tension of "long-term vision vs sacrificing current profit," where APD's history happens to offer a cautionary reverse sample:

The prior stage is a failed case of "over-sacrificing the present for a long-term narrative." For the long-term narrative of green hydrogen/energy transition, the former management invested enormous capital and took on long-cycle, capital-heavy projects, resulting in a one-time charge of about $3.7 billion pre-tax due to project exits in fiscal 2025 (about $3 billion after tax, $13.68 per share) and a full-year GAAP operating loss of $877 million. This shows that "willingness to sacrifice the present for five-to-ten years out," if it lacks return discipline, is value destruction, not a virtue. This is also the core reason the report scores the old management's capital allocation low and rates the whole only "Watch."

New management "recalibrated" its vision back to discipline rather than keep painting pies. After the new CEO took over, the actions were admitting fault, cutting losses, and lowering capex — fiscal 2026 capex guidance cut to about $4 billion, converging from a "megaproject narrative" to "focusing on high returns." The profit side began to repair (fiscal Q2 2026 adjusted EPS up about 19% year-over-year, full-year guidance raised to $13.00–13.25). This is a rational direction, but it reflects "capital discipline" rather than the growth aggressiveness of "daring to actively sacrifice current profit for the long term" in the Baillie Gifford context.

There is one long-term bright spot on shareholder-friendliness worth noting: the company is one of the few U.S. stocks that barely uses buybacks to dress up EPS — the 10-K shows it has not repurchased shares since fiscal 2013, with cash used mainly for dividends and projects (per the report); and it has raised its dividend for over forty consecutive years, raising the quarterly dividend to $1.81 in January 2026, its 44th consecutive year of increases. This reflects a long-term, steady capital-return culture that does not play financial games, positive evidence on the "trustworthy" side; but dividend growth has slowed sharply in recent years (increases of only about 1.1% and 2.3% in 2024 and 2025), which indirectly reflects cash being tied up in large projects.

Qualitative conclusion: management is "worthy of cautious trust" (sound governance, no buyback games, currently admitting fault and correcting course), but the Baillie Gifford core standard of "deep interest alignment + founder-style long-term vision" is not met — no founder, minimal direct CEO holding, alignment via rules, and the company just took a big hit for "sacrificing discipline to a long-term narrative." This dimension should get a neutral score, consistent with the report's neutral-leaning "management and capital allocation" and overall "Watch" tone.

Jun 11, 2026
If it disappeared tomorrow, how much would customers miss it? Is its growth model sustainable and free of dependence on harming society and regulation?6/10

Conclusion first (dual lens): on "indispensability," if APD disappeared tomorrow, its large on-site customers would miss it enormously — their continuous production depends heavily on stable on-site gas supply and can hardly be seamlessly replaced in the short term, so indispensability is high; on "social and regulatory sustainability," its growth model is clean, does not depend on harming society or regulatory arbitrage, and is instead tied over the long term to positive demand trends such as decarbonization, healthcare, and food safety. Both ends lean positive, one of the few dimensions in this report worth a higher score.

First, indispensability (how much customers would miss it):

For large customers it is "pull it and production stops"-level dependence. About half of revenue comes from 15–20-year on-site long-term contracts, with plants built at refining/chemical, chemical, electronics, and metals customers' sites or connected to pipeline networks, deeply coupled with customers' continuous production, safety, and process stability (per the report, citing the 2025 10-K). Once such customers lose gas supply, it means production stops, with losses far exceeding the cost of the gas itself, so switching costs are extremely high and renewal rates are high. The objective corroboration is that this indispensability converts into stable earnings: fiscal Q2 2026 adjusted operating income of $753 million, with adjusted EPS both up about 19% year-over-year, holding performance steady on contract stickiness even when volume was weak.

For social infrastructure it is an "invisible essential." Gases are widely used in healthcare (medical oxygen), food freezing, semiconductor manufacturing, refining, and more, spanning about 50 countries (per the report). It is the invisible base of industry and daily life, and its disappearance would cause cross-industry chain disruptions.

But two boundaries must be honestly flagged: first, the "missing" is industry-level rather than company-exclusive — if APD disappeared, Linde, Air Liquide, and Messer could take over most demand over the medium-to-long term, and Linde's 2025 sales of about $34 billion and larger scale show that industrial-gas supply does not have to be APD. So indispensability is more "short-term irreplaceable + long-term replaceable by peers." Second, the merchant side (about 44% of sales, liquid/cylinder bulk gas, per the report) is less sticky than on-site and more replaceable.

Now whether the growth model is sustainable and free of dependence on harming society and regulation (implicit premise: dual scrutiny):

Social dimension — clean and positive. It does not make money via addictiveness, information manipulation, regulatory arbitrage, or harming consumers; on the contrary, its incremental demand is tied to socially positive trends such as decarbonization (green/low-carbon hydrogen), electronics (semiconductor localization and capacity expansion), healthcare, and food safety. This is a "the more the better, won't be backlashed by society" business.

Regulatory dimension — a tailwind that does not rely on gray areas. The clean-energy/green-hydrogen business benefits over the long term from global decarbonization policy and subsidy frameworks, and the regulatory direction is a help, not a hindrance. The report also notes that some European end-markets face high-cost and import pressure, with weak helium prices, but these are cyclical/cost disturbances, not "regulation seeking to suppress its business model." It does not profit from regulatory arbitrage, and therefore has no structural risk of "collapsing when regulation tightens."

The one thing worth pointing out is the other side of policy dependence, not any harm to society: green-hydrogen project returns depend partly on subsidies and policy stability, and if policy is rolled back it will affect new-project economics — but this affects the return on growth projects (already partly reflected in the roughly $3.7 billion of project exits in fiscal 2025), and does not constitute an ethical or compliance problem of "harming society in exchange for growth."

Qualitative conclusion: high indispensability (especially for large on-site customers, nearly impossible to replace short-term), and a clean, sustainable growth model (tied to positive demand such as decarbonization/electronics/healthcare, not dependent on harming society or regulatory arbitrage) — on Baillie Gifford's dual question of "how much customers would miss it + whether growth can withstand social and regulatory scrutiny," APD answers quite solidly and should get a higher score. The only demerit is "the missing is industry-level rather than company-exclusive," consistent with the report's underlying judgment that "the moat is real and the business is excellent."

Jun 11, 2026
What are this business's unit economics (gross margin, incremental returns)? Do they get better or worse as scale grows? Where does the money it earns go?5/10

Conclusion first: APD's unit economics are themselves excellent — gross margin of about 31%, a high underlying segment operating margin, and steady operating cash flow, a good business that can spit out cash over the long term; but its "incremental returns" have clearly worsened in recent years. The problem is where the earned money goes, not business quality — over the past few years, large amounts of capital were poured into low-return, even to-be-exited, large projects, temporarily dragging down the good business's incremental capital returns. Growing scale should make unit economics better (rising network density), but wrong capital allocation once turned this curve in reverse.

First, the "quality" side of unit economics (gross margin, underlying earnings):

Gross margin is steady and resilient. On a consolidated basis, fiscal 2025 sales were about $12.04 billion, with a gross margin of about 31% (per the report; compressed to 26.5% during the 2022 energy inflation, then recovered). Being able to pull gross margin back through cost pass-through under an energy shock itself shows healthy pricing power and cost structure.

Underlying operating profit is solid and far smoother than GAAP net income. The report discloses segment operating income for the three years 2023–2025 of about $2.74 billion, $2.95 billion, and $2.86 billion respectively — this is the real earning power of the core business, not distorted by the 2024 LNG sale gain or the 2025 project-exit losses. The latest quarter corroborates the same way: fiscal Q2 2026 adjusted operating income of $753 million, up about 19% year-over-year, with a single-quarter adjusted operating margin above 23%.

Operating cash flow is real and did not collapse. Fiscal 2025 operating cash flow was about $3.26 billion, and first-half fiscal 2026 operating cash flow was about $2.0 billion, far above the $1.14 billion in the prior-year period. Even with the huge 2025 write-downs, cash kept flowing at the operating level — this is exactly the hard evidence of "good unit economics."

But "incremental returns" and "where the money goes" are the real questions on this dimension:

Incremental capital returns worsened in recent years, dragged down by capital allocation, not business deterioration. The report is explicit: what drags on "distributable cash" is the capital consumption and correction costs of large projects, not the core business. The most intuitive point is the loss of control and recovery in capex — fiscal 2025 actual capex was as high as about $7.02 billion (cash-flow-statement basis; the company also has an adjusted basis of about $5.06 billion), far exceeding operating cash flow in the same period, meaning the marginal return on incremental capital in these years was very low, with part even written off through the roughly $3.7 billion of project-exit charges. The report judging both "can it generate stable free cash flow / is its return on capital excellent" as "uncertain" stems exactly from this.

Growing scale should be a tailwind for unit economics but was offset. The larger the industrial-gas scale, the stronger the pipeline density, procurement, scheduling, and reliability, so unit economics should improve with scale; but over the past few years the scale dividend was invested into low-return large projects, breaking the "bigger scale → better returns" chain at the "capital allocation" link.

The comparison of "where the money goes" is most telling. As another scaled gas leader, Linde's 2025 return on capital was industry-leading (about 25.7% in Q1), operating cash flow about $10.4 billion, adjusted operating margin 29.8%, spending cash on high-return increments and buybacks/dividends; APD, by contrast, poured a lot of cash into projects that would be exited. The report's judgment is incisive: is APD still "the most capital-disciplined one in this good industry"? The answer is no.

The repair signal is worth a note: as fiscal 2026 capex guidance drops to about $4 billion and focus shifts to high-return projects, incremental capital returns are likely to recover and the visibility of free cash flow turning positive is improving — but this is a "repair to be verified," not "excellent incremental returns already delivered."

Qualitative conclusion: the business's "static unit economics" are excellent (high gross margin, strong cash flow, steady segment earnings), but its "dynamic incremental returns" worsened in recent years due to capital-allocation missteps, and "where the money goes" is a clear demerit. Scale should make unit economics better, but this was offset by low-return large projects. This dimension is a textbook split of "good business + bad capital allocation," fully consistent with the report judging "free cash flow / return on capital / capital allocation" as uncertain or failing.

Jun 11, 2026
To make it 5x in ten years, which conditions must hold simultaneously? Are these conditions realistic? What expectations does today's price imply?2/10

Conclusion first: to make APD 5x in ten years (about 17.5% annualized) requires a string of conditions to hold simultaneously, and these conditions are extremely unrealistic for a low-to-mid-growth, mature industrial-gas business; what today's price of about $282 (market cap about $63 billion, early-June 2026) implies is precisely that the market has already priced in more than half of the turnaround success in advance, not a "deep undervaluation waiting to be discovered" — so a 5x in ten years is a nearly untenable proposition here, and this core Baillie Gifford question should get a low score.

First, break "5x in ten years" into the conditions that must hold simultaneously (implicit premise: which elements must be true at once):

Condition one: earnings must grow about 5x in ten years, i.e., about 17% annualized. But APD is a mature business, with fiscal 2025 revenue down about 1% year-over-year, stuck at the $12 billion scale, full-year adjusted EPS of $12.03, down about 3% year-over-year, and fiscal 2026 guidance, even after being raised, of only $13.00–13.25, corresponding to about 8%–10% growth. Making EPS grow 5x at about 17% annualized over the long term far exceeds the core business's low-to-mid single-digit growth, and would require adding an entirely new, enormous, high-return growth pole — and that leg (green hydrogen) was just cut by the company itself.

Condition two: the valuation cannot compress, and preferably should expand. Today APD's P/E is already about 29.3x (on normalized earnings; on the FY2026 guidance midpoint of about $13.1, the forward P/E is about 21.5x), which is not cheap. For a 5x in ten years, if about 17% annualized earnings carries the whole thing, valuation must also hold at a high level without pulling back; but once the market re-rates it from the current high valuation back toward a more ordinary industrial-stock range (the report notes 18–22x), it becomes a strong headwind, not a tailwind.

Condition three: capital allocation must turn thoroughly from "value destruction" into a "high-return compounding machine." This requires that legacy projects such as NEOM, Louisiana, and Alberta no longer take impairments, and that incremental capital keeps going into high-return on-site with ROIC pulled to industry-leading levels. But the company just took roughly $3.7 billion of pre-tax exit charges in fiscal 2025 because of these projects, so turning from here into a 5x-in-ten-years engine has an extremely high bar.

Are these conditions realistic? Placing it in the industry coordinates makes it clear: even the clearly higher-quality leader Linde's 2025 sales grew only about 3% year-over-year, and Air Liquide's operating margin was about 20.7%, so the whole industry's compounding pace is "excellent but moderate." Expecting a turnaround-phase APD to run a 5x in ten years amounts to requiring it to both close the quality gap and surpass the industry's best in growth speed — a low probability of holding simultaneously. The report's three intrinsic-value tiers (conservative $190–220, fair $240–290, optimistic $320–380) also confirm: even in the optimistic scenario, the ten-year horizon falls far short of 5x the current price.

So what expectations does today's price imply (the reverse question Baillie Gifford values most)?

What it implies is "a successful turnaround + maintaining a top-tier leader valuation," not "distressed undervaluation." The report's judgment is precise: the current price looks more like paying at once for "a good underlying business + a successful turnaround" than exploiting a market error; the margin of safety is not obvious. About $282 sits right near the report's "upper edge of the fair-value range."

The equity risk premium is thinned. The current 10-year Treasury yield is about 4.55%, while at the current price APD's dividend yield is only about 2.6% (annualized dividend of about $7.24 ÷ about $282), and even on the conservative owner-earnings basis the yield is barely over 4%–5%. That is, the market has already priced in the "repair," leaving the expected return and safety cushion for new buyers both thin.

Qualitative conclusion: the three conditions needed for a 5x in ten years — "high-speed earnings growth + no valuation compression + capital allocation reversing into high-return compounding" — can almost never hold simultaneously; what today's price implies is "the turnaround is already priced in and the risk compensation is thin," not undervaluation. This core Baillie Gifford dimension should get a low score, fully consistent with the report's conclusion of "inadequate margin of safety, upper edge of fair value, rating Watch."

Jun 11, 2026
Why hasn't the market realized all this? Is it that it can't understand, looks down on it, or can't see far enough? What will become the "narrative inflection point"?3/10

Conclusion first: this Baillie Gifford question of "why hasn't the market realized it" must actually be answered in reverse for APD — the market has not mispriced it by "failing to understand, looking down on it, or failing to see far enough"; on the contrary, the market sees it quite clearly and has already priced "a good underlying business + a successful turnaround" into the price in advance. At today's roughly $282 (early-June 2026) there is no "cognitive gap waiting to be discovered"; instead it implies a not-low turnaround expectation. So this dimension is negative for a growth-investing framework: the market may be overestimating the certainty of its repair, rather than failing to notice its merits.

First, compare against Baillie Gifford's three types of "market blind spots" one by one, to see whether APD truly has an unpriced cognitive gap:

Not "can't understand." Industrial gases is a century-old, mature industry; APD is an S&P constituent, broadly covered, with ample sell-side research. Its merits (long-term contracts, moat, cash flow) and its wounds (the green-hydrogen gamble, about $3.7 billion of project exits, the CEO change) are all publicly disclosed and repeatedly discussed. This is not a hidden, thinly-covered cognitive lowland.

Not "looks down on it." Far from looking down on it, the market has actually given it a high valuation close to a top-tier leader's — a current P/E of about 29.3x, only slightly below the clearly higher-quality Linde (around 33–34x near June 2026). The report's judgment is incisive: APD's valuation is "slightly below the leader's, but not low enough to compensate for a weaker execution history and greater project uncertainty." This is precisely evidence that "the market does not think little of it."

Not "can't see far enough" (in the undervalued direction). If there is any "far-sighted" component, the market is instead looking ahead optimistically — paying in advance for a successful turnaround. Analysts' 12-month average target price is about $328, and target prices have been raised to roughly the $325–330 range, reflecting a consensus of "believing it will get fixed," rather than "failing to see its long-term value."

So where is the real cognitive divergence? It lies in whether the market's pricing of turnaround certainty is too optimistic, not in the market underestimating its merits:

The market may be overestimating the speed and certainty of the repair. The profit side is indeed improving — fiscal Q2 2026 adjusted EPS up about 19% year-over-year, full-year guidance raised to $13.00–13.25, and capex guidance cut to about $4 billion. The market linearly extrapolates this good news into "the problem is over." But the report cautions: after book profit recovers, investors easily assume the problem is over, while truly distributable cash flow may still lag capital investment.

What the market relatively overlooks is downside risk, not upside value. At the current price the dividend yield is only about 2.6% (annualized dividend of about $7.24 ÷ about $282), and against the roughly 4.55% 10-year Treasury, the equity risk premium is thin. In other words, the direction in which the market is more likely to "fail to see far enough" is — underestimating the tail risk of "turnaround failure / valuation falling back to the ordinary 18–22x industrial-stock range." The report estimates that in this scenario the price could return to $150–180, implying about 38%–48% downside for current buyers.

What will become the "narrative inflection point" (implicit premise)? Here we must look both ways, because the current narrative is "in repair, market leaning optimistic":

Upside inflection (reinforcing the optimistic narrative, but the room is already partly priced in): several consecutive quarters of adjusted-operating-margin expansion + continued improvement in owner earnings/operating cash flow + capex indeed falling year by year as guided + no more large project impairments. If all deliver at once, the market would fully re-rate it back to the "Linde-style steady compounding" tier — but because most is already priced in, the upside elasticity is limited.

Downside inflection (puncturing the optimistic narrative, more damaging): another $1-billion-scale project impairment or exit; capex that won't come down while owner earnings are still consumed; net debt that keeps rising while dividend growth relies on added leverage; management shifting back toward a "big-and-comprehensive" project narrative. Any one of these could compress valuation from the current high level toward the ordinary industrial-stock range — this is the more worrisome "inflection point" at the current price.

Qualitative conclusion: APD does not fit Baillie Gifford's "cognitive gap the market hasn't realized" paradigm — the market understands it, does not think little of it, and has even priced in the turnaround somewhat optimistically in advance. The real asymmetry lies in downside risk being relatively overlooked, not upside being underestimated. This dimension should get a low score, highly consistent with the report's judgment of "inadequate margin of safety, the market having priced in most of the turnaround expectation, rating Watch."

Jun 11, 2026

Ask about this report

Members can ask about this report; once answered it appears under "Reader Q&A" on this page. You can also highlight a passage in the text to ask about it directly.