CPI AEROSTRUCTURES INC (CVU) stock intrinsic value fair value estimate & owner earnings analysis

CPI Aerostructures (CPI Aero) builds structural aircraft assemblies, welded tubes and wiring, and integrated pod structures, almost entirely for military aircraft. Around 80% of revenue comes from subcontracts with defense primes, about 11% directly from the US government and about 9% from commercial work such as Embraer business-jet inlets. Concentration is high: four customers, led by Raytheon and Sikorsky, account for most of revenue. Total backlog at the end of 2025 was $504.5 million, but only $91.8 million of it was funded; the rest is the expected life of programs and can be cancelled or rescheduled without penalty. The defining feature of this business is that it takes on multi-year fixed-price contracts and periodically discovers that one of them costs far more than expected. That happened on the A-10 wing work in 2014, in the 2018-2019 accounting restatement that wiped out most of the reported equity, and again in 2025 when Boeing terminated the A-10 main landing gear pod contract in July after CPI asked for a price increase. The 2025 charge took about $10.2 million out of gross profit, revenue fell 14.6% to $69.3 million, the company missed a loan covenant in the second quarter and needed a waiver, and Boeing has asserted a claim for damages for alleged default that is not provided for in the accounts and remains unresolved. How we measured earnings. We measure the cash the business generates for its owners: operating cash flow, stripped of working-capital timing swings, less stock compensation and the capital spending needed to stay in business (which is tiny here, well under 1% of revenue). We then add back cash interest to see what the business earns before paying its lenders. On that basis the business produced about $4.3 million in 2021, $5.2 million in 2022, $6.7 million in 2023, $7.0 million in 2024 and $0.4 million in 2025. We use 2021-2025, the period since the restatement was cleaned up and the current contract base took shape, and take the median, about $5.2 million a year (roughly $0.40 a share). Two judgements matter here. First, the big net income figures of 2022 and 2023 were mostly the release of a tax valuation allowance and a forgiven PPP loan; neither is cash and both are excluded. Second, we remove working-capital movements, because contract assets (unbilled work, $34 million at mid-2026) swing heavily with milestone timing. Over the long run that build-up is real cash tied up, which is why the business has generated little cumulative free cash despite reported profits; we treat that as a reason for a wide range rather than as a lower normal level. The first half of 2026 is consistent with a return to normal: revenue rose 14% to $34.9 million and net income was $1.9 million, though operating cash flow was still slightly negative as receivables were collected and contract assets grew. Earnings in 2025 fell sharply versus 2024; the base case's first year assumes about $4.8 million of business cash flow, a return to the level already earned in 2021-2024 under the current programs rather than growth. No growth is credited in any case: the unfunded backlog depends on future appropriations and orders, and new programs (such as the Embraer inlet work starting in 2026) are not yet producing. Debt. At June 30, 2026 the company owed about $9.2 million on its Western Alliance revolving line and $9.9 million on its term loan, plus a small finance lease, against about $0.8 million of cash, for net debt of roughly $18.4 million. Both facilities mature December 12, 2030 and float at about 6.4%, close enough to our rate that no adjustment was made. The company also has about $68 million of federal tax loss carryforwards, which is why taxes paid are negligible and are expected to stay low for years; this is already reflected in the cash figures we use. Base case: business cash flow of about $5.0 million a year for the next decade, but with two years out of ten in which a fixed-price program loss or termination wipes out that year's earnings, as has happened roughly every five years in this company's record. After year ten the case assumes the core Black Hawk, F-16 and pod programs keep running but slowly roll off, fading from $4.5 million toward $3.6 million over ten more years and then ending, since no named replacement can be evidenced. The business is worth about $46 million; less $18.4 million of debt leaves about $28 million, or $2.14 a share. Bear case: Boeing's damages claim on the A-10 contract is paid at roughly $4 million in 2027, on top of another fixed-price loss of the 2025 kind, so the business consumes cash for three years and has to lean further on its revolver and covenant waivers; it then recovers to only about $3 million a year (the 2021 level) before fading out over the following five years as programs end. The business would be worth about $8 million against $18.4 million of debt, so the company would owe roughly $10.6 million more than it is worth, about negative $0.81 a share. In practice that means lenders would control the outcome and shareholders would likely be heavily diluted or wiped out. Bull case: the company sustains its best recent cash generation, about $6.5 million a year (just under the 2024 level), with only one moderate program setback in ten years, and the core programs keep running at about $6 million a year for a further decade. The business would be worth about $69 million, less $18.4 million of debt, or about $3.88 a share. Margin of safety price. We are not confident in the earnings level: the median of five years is about $5 million, but one of those five years was nearly zero, and over a longer span this business has lost money in as many years as it has earned it. The worst plausible outcome is a large adverse result on the Boeing claim combined with another program charge while the company carries $18 million of floating-rate debt and under $1 million of cash; with tangible equity of only a few million dollars beyond its tax asset, that structure would not absorb it without lender concessions or dilution, and shareholders could lose everything. The two things we are most likely to be wrong about are the size of the Boeing claim and how often fixed-price charges recur; if the claim is modest and charges recur only once a decade, the shares are worth around $3 to $4, and if both go badly, nothing. The company pays no dividend, so any return depends on the business rebuilding its earnings and paying down debt, which we expect to take three to four years to become evident. The margin-of-safety price is $0.85 a share, about $11 million for the whole company. At that price, roughly two years of normal business cash flow, applied to debt reduction, would shift about as much value from lenders to shareholders as the price itself, so a return to the 2021-2024 level of cash generation would more than double the share value within four years, while a single bad year does not eliminate the case. The margin-of-safety price applies only once the Boeing claim is resolved at $2 million or less, because until then the single largest swing factor is a number nobody outside the company and Boeing can see. This estimate is not comparable with our previous one, which was built on a different basis. What has changed in the business since then is mostly favourable: first-half 2026 results confirmed a return to profitability on the existing programs, while the Boeing claim remains unresolved.