Graham Corporation is a Batavia, New York maker of mission essential fluid, power, heat transfer, vacuum, and mixing equipment for nuclear submarines, torpedoes, refineries, and rocket engines, and its investment case now rests on whether a record defense backlog can keep converting into earnings while the stock has already run up sharply from a year ago. The case is not a growth story in the traditional sense, it is a multiple story, and the multiple is the number that has moved the most.
The defining recent event is the acquisition of FlackTek, a bladeless mixing business closed in January for $37,022,000 in cash and stock. The mechanism matters because the deal imports an installed base of consumables and service revenue that behaves more like a recurring annuity than an engineered order. A performance earnout capped near $25,000,000 ties future payments to adjusted EBITDA targets, which gives the transaction a built in margin test that the market can watch each quarter.
The central tension is valuation. The shares stood near $97 in mid August, after a late June close of $121.50, and a price like that underwrites a string of flawless quarters. It also sits against a base year of adjusted net income near $15,600,000, and defense now accounts for 60% of revenue. A budget or shipbuilding stumble therefore lands on the largest segment with the least offset from Energy and Process.
The timing trigger is the conversion of the record order book, which reached $557,217,000 at the end of June. Management expects more than a third of it to reach revenue within one year. Each report card tests whether the book converts at gross margins near the high 20s or slides back toward the fiscal 2026 mix.
Graham designs and manufactures custom engineered equipment where failure is not an option: ejectors and surface condensers for aircraft carrier power plants, turbines and pumps for Virginia and Columbia class submarines, MK48 torpedo hardware, cryogenic turbopumps for rocket propulsion, and vacuum and heat transfer systems for refineries, small modular nuclear plants, and geothermal projects. The fiscal 2026 revenue mix tells the story of where the company has landed. Defense produced $147,445,000 of net sales, or 60% of the total. Energy and Process contributed $83,343,000. Space added $14,505,000, about 6% of net sales. The company describes itself as having transitioned from a highly cyclical energy business into a diversified supplier to markets with enduring tailwinds, and the numbers show that transition is substantially complete. What that shift means in practice is that the revenue the company books now is tied to defense procurement cycles and long lead time engineering rather than to the capex decisions of a handful of large industrial customers, and the difference between those two drivers is the entire point of the investment case. The defense programs run on multi year schedules that do not reset with a budget cycle, and the engineering work is done close enough to the customer that a substitution decision is not a commercial conversation but a technical risk conversation.
The strategic logic of the past three years is a deliberate reweight toward the United States Navy. Graham serves the CVN aircraft carrier program, the Virginia fast attack submarine program, and the Columbia and Ohio ballistic submarine programs, plus all size classes of Navy torpedoes. It has reinforced the position with infrastructure that the customer helped fund. A new defense production building in Batavia, roughly 30,000 square feet, finished in mid 2025. Its funding came primarily from a $13,500,000 strategic grant from a defense customer. A cryogenic propellant testing facility near Jupiter, Florida came online in early 2026. Customer funded expansion is the signature of a supplier relationship. It means the customer deems Graham's capacity an essential path item for its own programs rather than a discretionary purchase, and it is one of the strongest signals a small industrial supplier can show that it has moved from being a vendor to being a program partner. The Navy does not hand out grant funded construction to companies it is willing to replace, and the fact that the grant was tied to a defense program rather than a commercial one tells you which side of the table Graham is now sitting on.
Two smaller acquisitions extend the technology base. Xdot Bearing Technologies was bought in late 2025 for $900,000 in cash. It carries up to $600,000 in milestone earnouts, and it folds foil bearing expertise into the Barber-Nichols operation while opening a path into high speed rotating machines. FlackTek, described below, adds the third core technology platform, advanced mixing, and the company now frames itself around three platforms rather than one. Each deal has been priced with an earnout, which keeps the purchase cost aligned with whether the acquired business actually performs. The earnout structure is a form of price protection, and it means the company is not paying full price for a business until the business has demonstrated that it is worth that price.
The counterweight to the defense concentration is the Energy and Process franchise, which still produces 34% of revenue and serves refineries, fertilizers, ethylene, and alternative energy applications including small modular nuclear, lithium extraction, and geothermal. Management notes that large capital project orders in that market have pushed out on geopolitical uncertainty, while New Energy demand offsets the slowdown. The split keeps the segment relevant but no longer positions it as the growth engine, which is precisely the point of the strategic shift. The Energy and Process business is now a margin and diversification cushion, and it also supplies the aftermarket revenue that keeps cash flowing in quarters when defense milestone timing is lumpy. The segment is smaller than it was, but it is the part of the business that behaves most like a traditional industrial company, and that is what gives the overall mix some balance.
The moat is not any single product but the combination of qualification history, low volume high mix manufacturing discipline, and early engineering engagement on programs that span a decade. Graham holds a position on every class of United States Navy nuclear submarine and on all size classes of Navy torpedoes, which means its equipment is qualified inside the most safety regulated procurement process in the country. Requalification of a substitute supplier is measured in years and in program risk that the Navy has little appetite to take, so the switching cost is structural rather than contractual. That durability is why a backlog of the size the company now carries can persist through a budget pause without vanishing. It is also why the revenue the company books is less correlated with the macro cycle than the industrial peer group it used to trade with, and that decoupling is the reason the multiple has expanded even as the absolute growth rate is modest. The qualification moat is not a moat that can be priced in a single quarter, because it is built from years of demonstrated reliability under conditions where a failure has a national security consequence, and it is not something a lower cost competitor can shortcut even in principle.
Barber-Nichols, the Arvada, Colorado turbomachinery subsidiary acquired in 2022, extends that moat into space and cryogenics. It supplies rocket turbopumps, cryogenic propellant pumps, and nuclear propellant pumps, and the Xdot foil bearing acquisition adds an oil free bearing technology that removes lubrication constraints from high speed rotating machines. The mechanism is that bearing technology is a component level advantage that multiplies across every rotating machine design, rather than a single product win, and the combination positions Graham to bid into markets it has not historically served. The new Florida cryogenic facility deepens the same advantage by letting the company test against real propellant rather than simulant, which is a differentiator in propulsion work. In rocket and submarine propellant systems, the customer cares more about demonstrated performance at the limit of the envelope than about price, and a facility that can demonstrate it is a gate that keeps competitors out of the bid. The cryogenic capability also gives the company a seat at the table for next generation propulsion work that does not yet have a design frozen, which is where the long term value of the space franchise lives.
FlackTek contributes a different species of moat: an installed base. Its bladeless centrifugal mixers are process essential in adhesives, sealants, composites, electronics, energetics, and battery production, and once a customer qualifies a mixer into a production line the replacement decision is rare. The installed base generates recurring demand for consumables, accessories, and services, which is why the purchase agreement prices in a performance earnout capped near $25,000,000 against four years of progressively higher EBITDA marks. That earnout is the cleanest visible test of the thesis, because if FlackTek cannot hit those marks the sellers do not get paid and Graham has paid a premium for a business that did not perform. The installed base also gives the company a recurring revenue line that is not tied to any single capital project, which is the structural change in revenue quality the acquisition is actually buying. The consumables and service revenue is lower margin than the defense work, but it is more predictable and it does not require a procurement review, which makes it a meaningful stabilizer for the overall revenue mix.
The production environment itself is a capability. It is low volume, high mix, customer driven engineering changes during production, rigorous quality documentation, and the ability to manage outsourced fabrication in China and India to meet content requirements. These are unglamorous advantages, but in engineered to order markets they determine who gets shortlisted in the first place, and they are not something a new entrant can buy on a timeline short enough to matter before the current backlog converts. The combination of qualification, manufacturing discipline, and early engineering engagement is what lets the company win work that a lower cost competitor cannot touch, and it is the reason the defense mix shift is happening at all. The moat is not visible in any single product spec sheet, and it is not something that shows up in a competitor announcement, which is exactly why it is durable. It shows up in the fact that the company is still on the program after a decade of engineering changes and content revisions, and that record is the asset.
The last fiscal year was a scale year, and the first thing to understand about the result is the divergence between the top line and the bottom line. Net sales grew 17% to $245,293,000. Net income, by contrast, grew only 2% to $12,500,000. Defense revenue rose 21% to $147,445,000. Defense contracts carry lower margins, and the mix shift pulled gross margin down 170 basis points to 23.5%. A one time welder training grant that flattered the prior year did not repeat, which compounded the effect on the comparison. The divergence between revenue growth and net income growth is the single most important thing to understand about the company's economics, and it is the reason the adjusted figures matter more than the GAAP ones.
The adjusted figures better reflect run rate because they strip out acquisition charges, integration costs, and a bonus program that distorts the GAAP line. SG&A included roughly $4,258,000 of Barber-Nichols performance bonus expense, and that charge falls away in fiscal 2027. Other operating income benefited from $621,000 of earnout liability fair value changes, a credit rather than a cost. Adjusted EBITDA for the year was $26,000,000, up 16%. That is the figure management guides against, and the gap between EBITDA and GAAP net income is mostly integration and acquisition cost plus amortization.
Orders are the forward signal, and they accelerated sharply. Orders for the year jumped 56%. The book to bill ratio reached 1.5x. Defense now accounts for 85% of the backlog, which gives visibility but also concentration. The aftermarket base, by contrast, cooled. Aftermarket orders slipped to $36,572,000 from $48,462,000, a sign that the legacy industrial book is running cooler even as defense burns, and that asymmetry is the honest read of the current order profile.
The first quarter of fiscal 2027, filed in early August, showed the transition in motion, and it carries the clearest signal the company has given about the shape of the year ahead. Revenue rose to $71,342,000. FlackTek contributed $6,551,000 of that total, which is meaningful for a business that closed only a few months earlier. Defense revenue jumped sharply on milestone timing. Gross margin was 25.0%, down on the heavier defense mix. Net income of $3,912,000 fell against the prior year print. Adjusted earnings rose to $5,738,000, which is the more honest read of the underlying business. The backlog reached $557,217,000. The quarter is a preview of the mix problem and the backlog strength at once, and it is the quarter that sets the tone for the conversion test that follows.
The thesis turns on four variables: backlog conversion rate, gross margin trajectory, FlackTek integration and earnout progress, and the balance sheet repair around the T. Rowe Price transaction. Backlog conversion is the master variable because more than a third of the $557,217,000 order book is expected to become revenue within a year, which sets the fiscal 2027 revenue floor well above the prior year. The risk is that the conversion is not uniform. Defense programs can slip milestone timing, and the company itself flags that order lumpiness means quarters can vary significantly, so a slow quarter is not automatically a broken thesis. The backlog is a floor, not a schedule, and the gap between the floor and the actual print in any given quarter is where the multiple gets decided. The conversion rate is the number the market is watching, and it is the one that can surprise in either direction without any change in the underlying order book.
Gross margin trajectory is the second variable, and management has already set the frame. The guidance is for gross profit between 24.5% and 25.5% of sales. That is a modest step up from the 23.5% printed in fiscal 2026. First quarter fiscal 2027 already printed 25.0%, which is in range. If defense keeps rising toward 65% of revenue, the margin floor erodes even as revenue grows, and the offsetting mechanics are the drop in bonus expense, scale in SG&A, and the higher margin aftermarket and space content. The bear case is a business that grows at a double digit clip and earns a low 20s gross margin indefinitely. In that world the revenue growth is real but the quality of the earnings is lower than the multiple implies, and the stock has to earn its price on volume rather than on margin.
FlackTek is the third variable and the most binary. The company paid $37,022,000 for a business with roughly $30,000,000 of annualized revenue. That is a multiple above 1x, and the performance earnout means total potential outlay approaches $62,000,000 for a business the company itself pegs at a $30,000,000 run rate. That only makes sense if the installed base grows and the consumables attach rate rises. The first quarter already shows FlackTek generating orders at 2.0x its own sales, which is the encouraging signal, but four years of EBITDA targets are long odds, and any stumble resets the multiple debate. The earnout structure is a good price control for Graham, but it does not protect the multiple the market pays for the whole company.
The fourth variable is execution around the capital raise and balance sheet repair, which is the least glamorous but the most necessary of the four. The April 2026 Securities Purchase Agreement sold 600,000 shares. The price was $83.36 per share. Gross proceeds reached $50,000,000, of which a substantial portion repaid revolver borrowings. Cash stood at $26,953,000 in late June. It had recovered from a much lower trough at the end of March, and the recovery is the point. The raise was priced at a discount to the then current share level, which tells you the company prioritized balance sheet strength and acquisition dry powder over short term dilution optics. Capital expenditure guidance for the coming year sits between $18,000,000 and $22,000,000. That is well above the maintenance level, so discretionary investment continues, and the company is still choosing to grow the plant rather than merely maintain it.
The dominant risk is customer concentration at the macro level. The United States Navy is the effective end customer behind the majority of defense backlog, and shipbuilding programs run on budget cycles that can compress or re phase without touching Graham's fundamentals. A two year pause in Columbia class propellant pump orders, or a Virginia class rate cut, would hit the revenue floor built from the current backlog, and the 84% defense share of the June 2026 backlog means there is little cushion. Space and Energy and Process are too small to absorb a meaningful defense stumble, which is the structural trade of the reweight. The company bought a more durable revenue base by accepting that the bulk of it now lives in a single customer's budget, and there is no way to have one without the other at this scale. The concentration is not a flaw that management is working to fix, it is the feature that made the transition possible, and the risk is that the feature and the flaw are the same fact.
The second risk is margin dilution from mix. Every point of defense revenue share that replaces higher margin Energy and Process or aftermarket content pushes gross margin down, and fiscal 2026 already demonstrated the direction of travel with a 170 basis point decline. If the defense share of revenue keeps climbing toward 70%, the company could be growing at a double digit rate on a low 20s gross margin. That would quietly compress operating leverage and make the stock look expensive even in a growth quarter, because the multiple is a function of the quality of the earnings as much as the growth rate of the revenue, and a margin that settles at the low end of the range is a different company than the one the price implies.
The third risk is acquisition digestion. Graham has closed three acquisitions in roughly two years. Barber-Nichols landed in 2022. Xdot closed in late 2025, and FlackTek in early 2026. The SG&A line carried $1,827,000 of acquisition and integration expense in fiscal 2026 alone. Each deal adds amortization, earnout accounting, and management attention. The FlackTek deal also brought a contingent liability at acquisition and a set of transaction bonuses assumed from prior owners, both of which are one time hits that complicate quarter to quarter comparison. The pattern of serial acquisition also means management attention is split across three integration workstreams at once, which is a real execution tax that does not show up in the headline revenue line. The integration cost is a one time expense, but the management bandwidth it consumes is a recurring cost that shows up in the quality of the execution in other parts of the business.
The fourth risk is valuation itself, detailed in the next section. A stock that has run up this far prices in sustained execution, and the natural downside scenario is not fraud or collapse but a normalization. Orders hold, margins drift, and the multiple re rates from a growth multiple toward a cyclical industrial multiple. That re rating, more than any operational failure, is what creates the deep drawdown an owner should fear. A smaller fifth risk is the ERP implementation at Batavia, which has already generated a several million one time consulting cost and carries execution risk into the coming fiscal year. ERP projects at small manufacturers are notoriously prone to scope creep and delayed go live, and a slip in that project would add to the SG&A load at exactly the time the margin story needs to look clean. The ERP risk is small in absolute terms, but it lands in the same quarter as the margin recovery narrative, and that is when a one time cost is most visible to the multiple.
The framework is an earnings multiple applied to forward adjusted earnings, because GAAP results are distorted by acquisition amortization, earnout fair value changes, and bonus programs that management itself flags as non recurring. The choice of adjusted earnings as the base is not a preference, it is a necessity, because the GAAP line includes one time items that do not reflect the run rate of the business and that would make any multiple built on them meaningless. Shares outstanding in early August stood at 11,737,033, and the stock closed late June at $121.50. That gives a market value near $1,426,000,000. Against fiscal 2026 adjusted net income of $15,598,000, that is a multiple of roughly 92x, a level that only makes sense against a forward estimate.
For fiscal 2027, the construct starts from management's own guidance, which is the cleanest anchor the company offers and the one the market should actually use. Net sales are guided to between $285,000,000 and $295,000,000. That is driven by backlog conversion plus a full year of FlackTek. Gross profit is guided to a low mid 20s share of sales at the midpoint. Adjusted EBITDA is guided to between $35,000,000 and $40,000,000. That is a large step from the prior year print, and it is the single most important number in the outlook. After interest and a normalized tax rate, adjusted net income comes out around $17,000,000 to $18,000,000. That works out to roughly $1.50 per share. The guidance set is internally consistent, and every element of it sits inside a range the company has already committed to.
The base case values the company at 40x to 45x fiscal 2027 adjusted earnings. That reflects a defense supplier with a record backlog, a growing space business, and an acquired consumables franchise, but also a mix tilting toward lower margin defense work. At midpoint earnings, that range implies a share value of roughly $60 to $68. That sits well below the late June close. The bull case, where backlog conversion beats and FlackTek earnout targets are met, supports 50x to 55x. That is on a forward print near the next fiscal year level. It implies roughly $95 to $105. The bear case, where defense orders pause and the multiple re rates to 25x, implies $31 to $35 per share. That is a drawdown of nearly three quarters from the late June close. The distance between the bear and the base is the whole story, and it is wide enough that the price, not the business, is doing most of the work in the current share level.
The explicit counterargument to the bear case is the backlog, and it is worth stating clearly because it is the strongest piece of evidence the company has for the bull side of the argument. A $557,217,000 order book with more than a third converting within a year is not a hope. It is contracted work with milestones, and even a one year slip converts into the next fiscal year rather than vanishing, because defense procurement does not cancel qualified suppliers over a budget pause. The counterargument to the bull case is the math of the multiple. At 92x trailing adjusted earnings, even the bull case requires the market to keep paying growth prices for a company whose margin profile is still settling, and the valuation gap between the two views is wide enough that the share price is doing most of the work.
The judgment is that Graham is a fundamentally stronger business than it was two years ago and a more expensive claim on that business than the recent price history supports. The two statements are not in tension, and holding both of them at once is the honest read of where the company sits in the market right now. The company has successfully re weighted itself from a cyclical refinery pump maker into a defense and space supplier with a contracted order book it did not possess before fiscal 2026, and the FlackTek acquisition adds a consumables annuity that changes the revenue quality profile. That is the real achievement, and it is durable. Qualified positions on every United States Navy nuclear submarine program do not come off the table because of a budget cycle, and the consumables annuity from FlackTek means the company now has a line of revenue that does not depend on a single capital project clearing a procurement review. The two changes together mean the revenue quality profile has shifted from a cyclical industrial mix to a more durable defense and consumables mix, and that is the fundamental change the market is paying for.
But the stock has run from the mid $40s a year ago to $97 in mid August, and the multiple math no longer forgives ordinary execution. At roughly 92x trailing adjusted earnings, the market is underwriting the bull case. Flawless backlog conversion, margin recovery into the top of the guidance range, FlackTek earnout hits, and continued defense order flow. Any one of those variables softening is enough to justify a multiple re rate, and the bear case of $31 to $35 per share is not a catastrophe scenario. It is a normalization scenario, and it is the one the market has not priced in, because the recent run has done the work of the bull case on the assumption that all four variables hold at once.
The decision that matters for the shareholder is what the fiscal 2027 quarter reports do to the two numbers that drive the multiple, adjusted earnings per share and gross margin, and the reason those two numbers matter more than revenue is that the multiple is built on them, not on the top line. If the first half of fiscal 2027 delivers adjusted earnings at or above $0.85 per share, the backlog conversion thesis is proven and the re rating risk fades. Gross margin at or above 25% is the second number that has to hold. If adjusted earnings come in near $0.55 with margin at 24%, the multiple compresses before the stock does, and the drawdown is a valuation event, not an operational one. The backlog does the defensive work, the multiple does the offensive work, and right now the price has front run the evidence by a wide margin. The judgment for a new owner is that the business is worth owning and the price is the problem, and that the difference between those two facts is the entire question the stock is asking. The business has earned a higher valuation than it had two years ago, and the price has already moved most of the way to the level that valuation supports, which is why the remaining upside is smaller than the remaining downside.