Airship AI Holdings, Inc. (AISP) stock intrinsic value fair value estimate & owner earnings analysis

What the business is. Airship AI sells AI-driven video, sensor and data-management software: Acropolis (the enterprise platform, in Commercial and Law variants), Outpost AI (edge devices that turn raw video into structured data such as plates and vehicle types) and Airship Command (the viewing applications), plus support contracts and professional services. Most revenue comes from U.S. federal agencies in Homeland Security and the Justice Department. It is small and highly concentrated: in 2025 four customers produced 87% of revenue, and in 2024 one order produced 57%. It became public through a SPAC merger in December 2023. What it has actually earned. We measure owner earnings as operating cash flow, stripped of working-capital timing swings, less stock-based compensation; the company has essentially no capital spending. On that basis it has lost money in every year on record: about -$1.6M in 2022, -$6.6M in 2023, -$3.4M in 2024 and -$6.9M in 2025 (about -$0.12, -$0.33, -$0.14 and -$0.18 per share). Revenue has been lumpy rather than rising: $14.5M, $12.3M, $23.1M, $15.3M. Even the best revenue year, 2024, lost about $3.4M. Reported net income is not a guide: it swung from a $57.5M loss in 2024 to a $29.3M profit in 2025 almost entirely from non-cash revaluations of warrants and earnout shares, which we ignored. We use the median of the four years, a loss of about $5.0M a year, as the starting level. The company has no borrowings and paid no cash interest in 2025, so the business and the shares carry the same cash flow. Near-term path. In 2025 the loss roughly doubled from 2024. For the first year of the base case we assume the median loss of $5.0M, which is better than 2025. The reason is in the June 2026 quarterly filing: first-half 2026 revenue was $10.5M against $7.6M a year earlier, gross margin near 60%, and the operating loss narrowed to $3.1M from $3.7M, over two consecutive quarters. That filing also shows positive operating cash flow of $0.6M for the half, but this came from collecting $2.7M of year-end receivables, not from the business turning profitable; after removing that and stock pay, the half still lost roughly $3M. What we credited and what we left out. Secured work is small: backlog was $3.3M and remaining contracted support obligations $8.6M at December 31, 2025, including a $1.9M Homeland Security award for 2026 security events and a $2.8M commercial hardware refresh. We gave no credit to the $173.4M sales pipeline, to border-technology funding in the July 2025 federal budget law, or to management's aim of reaching positive cash flow during 2026: none of these is a contract, and the company has never produced a year of positive owner earnings to show the aim is achievable. The 10-K also says it plans significant additional hiring in 2026, which cuts against a near-term profit. Bear case ($0.00). Losses run at the 2025 rate of about $6.9M a year, driven by the planned hiring and the concentration of revenue in four customers, with no large order replacing 2024's. The $12.4M of cash held at June 30, 2026 is used up by about mid-2028. The operating business is worth about -$12.4M in present-value terms, the company owes no debt, and it holds $12.4M of cash, leaving nothing for today's shareholders, who would then face heavy dilution or a wind-down in which roughly $8.9M of prepaid customer support still has to be delivered. Base case ($0.03). Losses continue at the median $5.0M a year. Cash covers about two and a half years. With no year of profit on record, we credit no value beyond that point. The operating business is worth about -$11.2M, it owes nothing, and cash is $12.4M, leaving about $1.2M, or $0.03 per share. In short, today's cash roughly pays for the losses we can foresee. Bull case ($0.44). The company's own levers work. Revenue holds near the first-half 2026 run rate of about $21M a year, supported by the new awards and support renewals, at the 60% gross margin it is now earning. Management brings operating costs back to about the $11M a year it ran in 2022. Losses shrink to about $3M in the first year and $1M in the second, then the business earns about $1M a year of owner earnings through year 10 on its installed Acropolis base and support contracts, fading to zero over the following four years because nothing we can evidence sustains it beyond that. The business is worth about $2.8M, owes nothing, and holds $12.4M of cash, for about $15.2M or $0.44 per share. Dilution. There are about 34.4 million shares, plus about 16.1 million public warrants expiring December 21, 2028, roughly 6.3 million other warrants and remaining earnout shares. The $13.3M warrant liability and $2.6M earnout liability on the balance sheet are settled in shares, not cash, so we did not treat them as debt; they would reduce each share's slice if exercised. The company has repeatedly funded losses by selling stock and through warrant exercises ($7.3M offering and $7.7M of exercises in 2024, $9.5M of exercises in 2025), and further losses would likely be funded the same way. What would make this wrong. If Airship converts part of its federal pipeline into multi-year contracts and shows several consecutive quarters of positive owner earnings, the bull case becomes the realistic path and the value moves up sharply. If the four main customers slow their spending, the value is simply the cash less the losses, and the cash lasts less than two years. Margin-of-safety price. Because the business has never earned money for its owners, price alone does not provide a margin of safety. The margin-of-safety price applies only after the business shows at least four consecutive quarters of positive owner earnings without shrinking revenue, and even then it is about $0.15, a steep discount to the $0.44 bull value, given the tiny size, customer concentration and dependence on federal budgets.