$13.93
$97.32
bear $13.93
base $59.82
bull $97.32
margin of safety $36.00
market price* $299.59
Business condition attached to the margin of safety
Air Products reports a full fiscal year (fiscal 2027 or later) with capital spending, on the company's own measure that excludes partner-funded NEOM spending, at or below $2.5 billion and with no new project-exit or write-down charge during that year.
Air Products is one of the world's three big industrial gas companies, alongside Linde and Air Liquide. It makes oxygen, nitrogen, argon, hydrogen and helium and delivers them by pipeline, from on-site plants built beside customers, and by tanker and cylinder. Many large customers sign take-or-pay contracts of 15 years or more with cost pass-through and price escalators, which is why the core business is steady. Sales are reported in four regions: Americas, Asia, Europe, and Middle East and India. What has changed is the last four years. Air Products spent far more than it earned on clean-hydrogen megaprojects built ahead of signed customers: the NEOM green ammonia plant in Saudi Arabia, a Louisiana clean energy complex, an Arizona liquid hydrogen plant and an Alberta hydrogen complex. Capital spending was $4.6 billion in fiscal 2023, $6.8 billion in 2024 and $7.0 billion in 2025, against operating cash flow of $3.2 to $3.6 billion. Exits cost about $3.0 billion of charges in fiscal 2025, and on 30 June 2026 the company walked away from the Louisiana complex and the Arizona plant, taking a further $2.9 billion pre-tax charge ($2.2 billion after tax) and a $1.44 billion net loss for the June quarter. The good news underneath is that the base business is still producing cash: operating cash flow for the nine months to June 2026 was $3.31 billion, revenue was up about 6%, and capital spending is being cut to about $3.5 billion this year. How we measure earning power. We start with operating cash flow, strip out swings in working capital (including about $560 million of tax paid in fiscal 2025 on the 2024 sale of the LNG equipment business), subtract stock compensation as a real cost, and subtract maintenance capital spending. Only spending that visibly supports revenue growth is treated as growth spending, at 29% of each year's revenue increase (the ten-year average ratio of capital spending to sales); in years when revenue was flat, all spending counts as maintenance. That is a conservative test for this company: in fiscal 2022 it counts $2.2 billion as maintenance against about $1.3 billion of depreciation, because Air Products builds plants a year or more before they earn revenue. We then add back cash interest so we value the business before its lenders, and subtract the debt separately. Which years count. We use fiscal 2016 to 2022, the seven years before the megaproject build-out, which include the 2020 shutdown. Fiscal 2023 to 2025 are left out because the spending in those years went into unfinished plants that earn nothing yet and the method cannot separate it from upkeep; the cash still needed to finish what remains is charged through the debt instead. Over the window, cash flow before lenders was $6.63, $7.85, $7.00, $5.06, $3.64, $6.23 and $5.09 per share. The middle value, $6.23 per share (about $1.39 billion a year), is our starting level. The nine-month 2026 results support that level still holding but are not enough on their own to raise it. Growth. In fiscal 2025 the measured stream improved by about 1% (it was negative in both 2024 and 2025 because of project spending), and the last clean year, 2022, fell. We therefore start the base case at 1% a year, not higher. Credited growth: 1% a year for ten years in the base case, backed by contractual price escalators and energy cost pass-through, the 15-year fixed-fee Uzbekistan gas processing contract, and the company's long record of adding contract-backed plants from its own cash. The bull case gives 3%, still below the roughly 5% a year revenue growth of the past decade. After year ten, cash flow is held flat, not grown: long contracts, pipeline networks and on-site plants have kept this business's cash flow steady for decades, but nothing we can evidence lifts it further. Left out: hydrogen demand growth, any profit from NEOM, the Alberta complex, margin improvement and management's earnings targets. The filings say the large clean-hydrogen plants were built before most of their output was sold, so we give them no value. Debt. Reported borrowings were about $17.7 billion at 30 June 2026, of which $5.5 billion belongs to the NEOM joint venture and its lenders have no claim on Air Products. We set the NEOM plant, its debt and the partners' share aside. That leaves about $12.2 billion of Air Products' own debt; less $1.0 billion of cash and plus about $0.6 billion for outside owners of other subsidiaries, net debt is about $11.8 billion, or $52.90 per share. Air Products' own borrowing costs about 3.9% overall, below our 6% yardstick, so we reduce the debt figure by $1.6 billion ($1.2 billion in the bear case). We then add amounts the company still has to pay: an estimated $0.5 billion of contract-cancellation cash for the Louisiana and Arizona exits (the company has not split the $2.9 billion charge between write-downs and cash; payables rose about $0.3 billion in the quarter), and about $1.0 billion of spending to finish committed projects beyond normal upkeep, based on the $3.5 billion 2026 budget. Maturities are spread out at roughly $0.8 to $1.4 billion a year over the next five years, and nothing in the filings points to trouble refinancing them. The three cases.
Bear: $13.93 per share. Cash flow before lenders starts at the fiscal 2019 level of $5.06 per share, then a downturn like the 2008-2010 industrial recession (when Air Products' sales fell roughly a fifth) cuts it by a quarter to $3.80 for two years before recovering, with no growth. The business is worth about $81.65 per share ($18.2 billion). It owes $67.72 per share ($15.1 billion): core net debt, a smaller debt-cost benefit, $1.0 billion of exit cash, $1.5 billion of completion spending, and a $2.0 billion provision for NEOM obligations falling on Air Products as builder under the $6.7 billion construction contract and as the sole buyer of the ammonia under a long-term take-if-tendered agreement. Value left over: $13.93. Even at the trough this case covers its interest about 1.8 times.
Base: $59.82 per share. The business is worth about $112.27 per share ($25.0 billion), owes $52.45 per share ($11.7 billion) and leaves $59.82.
Bull: $97.32 per share. Cash flow starts at the fiscal 2018 level of $7.00 and grows 3% for ten years. The business is worth $147.53 per share ($32.9 billion), owes $50.21 per share ($11.2 billion) and leaves $97.32. Change since our last analysis (late September 2026). We previously estimated $2.89 / $65.78 / $100.87. The base and bull are lower mainly because we now charge the cash still needed to finish committed projects and the cash cost of the June 2026 exits, which we had not done before. The bear is higher because we now start it from the fiscal 2019 level with a modelled recession rather than holding the 2020 shutdown low forever, which better reflects how this business has actually behaved through downturns; we also now count the latest, higher NEOM borrowings, which reduces Air Products' own debt. Air Products has also finalized a marketing and distribution agreement with Yara for NEOM's ammonia, which gives the output a route to market, but its terms are not disclosed, so it does not change our numbers. Margin of safety price: $36, once the condition below is met. How sure are we of the earning level? Moderately. The core contracts are long and the 2016-2022 record is consistent, but our maintenance spending test probably understates earning power, and three years of project spending have hidden what the business earns today. The worst plausible outcome is that NEOM's ammonia cannot be resold near cost for years while Air Products remains its sole buyer, on top of further write-offs; the structure would survive this, since the debt is investment grade, spread out and serviced by hard, contracted assets, but the shares would be worth something like the bear case, in the low teens. The things we are most likely to be wrong about are the NEOM obligation (worth roughly $9 to $13 per share either way), the size of maintenance spending (if it is nearer depreciation, value rises by $30 per share or more) and whether management really has stopped building ahead of customers. The company pays about $7.18 per share a year in dividends, but that is more than our estimate of current owner earnings and has been partly funded with borrowing, so the estimate does not count on it continuing at that level. The price assumes a holding period of three to five years. From $36, with even a reduced dividend and the business valued near our base case within about three years, the base case implies well over 20% a year, while the bear case implies a loss of most of the price; only once spending is visibly back under control does the bear case become less likely than one in ten, which is why the price applies only after that has been shown.