GRAHAM CORP (GHM) stock intrinsic value fair value estimate & owner earnings analysis

Graham Corporation makes vacuum, heat-transfer and fluid-handling equipment. Over the last five years it has turned from a small supplier to refineries and chemical plants (sales of about $90 million in fiscal 2020) into a U.S. Navy and space supplier ($245 million in fiscal 2026; Defense was 60% of sales, 85% of sales were domestic). It holds what it describes as a leading, and in most instances sole-source, position on certain Navy systems. Fiscal 2026 orders included $101.5 million for the Virginia-class submarine program, $75 million of Columbia-class follow-on orders and $25.5 million for the MK48 torpedo, plus $35.3 million from commercial space launch customers. Backlog was $533 million at March 31, 2026 (85% Defense), roughly three years of fiscal 2026 Defense sales, and remaining performance obligations were $557 million at June 30, 2026. Two customers took 19% and 16% of fiscal 2026 sales. How we measured earnings. Owner earnings are operating cash flow less swings in working capital, less stock compensation, less maintenance capital spending. Working-capital swings are large here because customers pay progress deposits (+$25.6 million in fiscal 2024, +$12.1 million in fiscal 2025, +$16.4 million in fiscal 2026), so we strip them out. Capital spending of $16 to $19 million a year in fiscal 2025 and 2026 is mostly expansion (new Batavia, NY manufacturing building, a Florida propellant test facility, Arvada, CO capacity) that the filing calls primarily discretionary; management puts true maintenance at about $2.5 million. We did not use that figure. We set maintenance spending at the 2.07% of sales Graham spent in the seven years before the build-out (fiscal 2017 to 2023), which is $3.8 to $5.1 million in fiscal 2024 to 2026 and about $6 million on guided fiscal 2027 sales, in line with depreciation. Interest is under $1.1 million in any year and is added back when valuing the business. On this basis owner earnings per share were $0.56 (fiscal 2023), $0.67 (2024), $1.40 (2025) and $1.53 (2026); including interest, the business cash flow per share was $0.66, $0.75, $1.42 and $1.57. Earlier years ranged from a loss of $0.92 (fiscal 2022, the year of the Barber-Nichols acquisition) to $0.70. Starting level. The business changed with the Barber-Nichols acquisition (fiscal 2022) and the Defense and Space ramp. We exclude fiscal 2022 and use the two years since profitability stepped up, fiscal 2025 and 2026: median business cash flow of about $16.6 million, or $1.50 per share. Using fiscal 2023 to 2026 would give a lower $1.09, so this window is a judgement we want readers to see. The June 2026 quarter does not break it: sales were $71.3 million (up 29%) and net income was 5.5% of sales, in line with fiscal 2026's 5.1%. Operating cash flow was negative $12.7 million in the quarter, but that is receivables up $16.0 million against customer deposits up $10.8 million, which we treat as working-capital timing. Growth reconciliation. Business cash flow per share grew 10.3% in fiscal 2026. Our base case assumes 9.1% per share in fiscal 2027. In total dollars that is a 16% rise, to about $20.3 million, because the company guided fiscal 2027 sales to $285 to $295 million in its June 2026 annual report, which the $533 million backlog covers, and we hold the margin at 7.0% (7.1% in fiscal 2026, 7.5% in fiscal 2025). The roughly 6.5% increase in diluted shares from the equity offering absorbs part of that. Base case, $19.24 per share. Sales stay at about $290 million and business cash flow at $20.3 million a year for ten years, with no further growth credited. The Defense and Space part of that stream, about $18 million, then fades evenly to zero over the following 15 years, while the older refining and petrochemical equipment business keeps contributing $2.2 million a year (the median of fiscal 2017 to 2021, when sales ran $78 to $98 million). The business is worth $222.7 million. Net of debt it owes nothing: the $13 million credit line was repaid in April 2026 and a $50 million equity offering was completed in the June 2026 quarter, leaving no borrowings (finance leases are negligible) and $26.95 million of cash at June 30, 2026. We count $12.0 million of that as spare (the rest runs the business and holds customer money) and deduct $6.5 million for the FlackTek acquisition earn-out (up to $25 million over four years from fiscal 2027), so net cash is $5.5 million. That gives $228.2 million over 11.86 million diluted shares (11.74 million outstanding in August 2026 plus dilutive awards). Credited because it is secured: the guided fiscal 2027 sales and Defense persistence supported by backlog and the named Navy programs. Left out because not yet evidenced: Space growth, the Arvada and Batavia expansion paying off, the guided gross-margin rise to 24.5 to 25.5% from 23.5%, small modular reactor demand, and any margin expansion. Bear case, $9.09 per share. Fiscal 2027 sales miss guidance at $255 million, and the margin falls to 4.5%, the level Graham earned in fiscal 2023 and 2024 while the Defense ramp was absorbed, with tariffs (about $1 million extra in fiscal 2026) and overruns on large fixed-price contracts, which the filing names as risks. Business cash flow is $11.5 million for two years. Then comes a full energy downturn like fiscal 2015 to 2016, when sales fell 33% (from $135.2 million to $90.0 million). Applied to the roughly $85 million energy and process business, that is about $28 million of lost sales at about a 20% drop-through, so cash flow falls to $6.0 million for three years, then returns to $11.5 million. Defense fades to zero over years 11 to 20, leaving $1.0 million a year. The business is worth $102.3 million; with $5.5 million of net cash, that is $107.8 million. Debt is not a problem in any case. Graham's goodwill and other write-offs in fiscal 2018 and 2019 ($14.8 million and $6.4 million) are non-cash but are a reminder that its newest acquisitions, including FlackTek, could disappoint. Bull case, $28.84 per share. Backlog converts and the gross-margin guidance is met: business cash flow of $20.3 million, $24.0 million, then $28.0 million a year (about $330 million of sales at 8.5%) from the third year, held for 15 years and then fading over 15 more, with $3.0 million a year from the legacy business. The business is worth $336.5 million; with $5.5 million of net cash, $342.0 million. What could make this wrong. If Graham's maintenance spending really is $2.5 million, the base case is about $22.28. If margins fall back to the 4.5% of fiscal 2023 and 2024, it is about $12.96. If Defense earnings end after year ten instead of fading, it is about $14.79. We could not fully reconcile management's $2.5 million maintenance figure to Graham's history, and the quarterly cash outflow is explained only by the receivable and deposit movements above. Margin-of-safety price, $11.00. Earnings are moderately certain: backlog covers fiscal 2027 and part of 2028, but the profitable level rests on two years, and the business cash flow margin was only 4.4% in fiscal 2024. The worst plausible outcome combines a Navy contract overrun, an energy downturn and a delay or loss at one of the two customers that supply 35% of sales. The business could survive it because it has no debt and a hard-asset base, but business cash flow would fall by close to half, as in the bear case. If we are wrong on the margin or on how long Defense lasts, the shares are worth $13 to $15. The price assumes a three-to-five-year holding period for the Defense stream to be visible in results, and the company paid no dividends in the last three fiscal years, so any return has to come from retained cash flow. From $11.00 the bear case implies roughly breaking even over three years, and the base case implies above 20% a year. Previous figures (September 2026) were built on margins rising toward 15% and 6 to 8% annual sales growth, neither yet evidenced, and on a different basis, so they are not comparable. The business facts behind them (record order book, debt repaid after the equity offering) are unchanged; this analysis simply credits only the sales covered by backlog and today's margin. The June 2026 quarter brought higher sales but a lower net margin than a year earlier (5.5% against 8.3%), which is why we do not assume any margin improvement.