Safe Pro Group Inc. (SPAI) stock intrinsic value fair value estimate & owner earnings analysis

Safe Pro Group is a small Florida-based company with three units: Safe Pro AI, which sells software that scans drone photographs for landmines and unexploded ordnance (its SpotlightAI system has analyzed about 2.9 million images in Ukraine and flagged more than 54,000 explosive threats); Airborne Response, a drone-inspection service whose main customer is Florida Power & Light and whose work rises and falls with hurricane damage; and Safe-Pro USA, which sells body armor and security uniforms and was hit hard by tariffs on Chinese-made goods in 2025. The business has never made money. Measured as cash the owners actually keep after paying staff in stock and maintaining the equipment and software, it lost about $6.1 million in 2023, $7.1 million in 2024 and $13.4 million in 2025, roughly 66 to 79 cents a share every year. Stock-based pay is large here ($6.9 million in 2025) and we count it as a real cost because it is paid for by issuing new shares. Goodwill and intangible write-downs in 2025 ($0.8 million) were treated as non-cash and do not affect these figures. Interest is negligible (under $10,000 a year), so the business's cash flow before lenders is essentially the same as after. 2026 has brought real revenue growth: about $2.55 million in the first half versus $0.28 million a year earlier, driven by U.S. Army work on the NODE edge-detection system, a U.S. Air Force small-business research award and an integration award from a defense prime, with gross margins near two-thirds. But the cost base grew with it: the first-half operating loss was $6.2 million, slightly larger than a year earlier, and operating cash outflow was $4.1 million. No backlog is disclosed, and the awards are individual development and demonstration contracts rather than multi-year programs, so we cannot treat further growth as secured. The company had $10.2 million of cash at June 30, 2026 against only $146,000 of notes, about 50 cents a share of net cash, raised largely through two private placements in 2025 ($8 million and $14 million). Earlier filings carried doubt about the company's ability to continue; those raises resolved it for now, but at the current spending rate the cash lasts only into 2027. Management has also been buying back stock (about 475,000 shares in the first half of 2026), which uses cash the losses also need. Bear case: losses run at the 2025 level of about $13 million a year including stock pay. The cash is gone in early 2027 and further losses are funded by issuing shares. The present value of losses through 2027 (about $18 million) exceeds the $10 million of net cash by about $8 million, so today's shareholders are left with nothing: we value the shares at zero, not below, because shareholders are not liable for the shortfall; it falls on them as dilution. Base case: revenue holds near the 2026 run rate of about $5 million a year on the existing awards and Ukraine work, but overhead does not fall, because nothing in the filings commits to cuts. Losses of roughly $5.5 million in the second half of 2026 and $10 million in 2027 (about $14.5 million in present value) consume all of the cash with about $4 million still to be funded by new shares. Value to current holders: zero. Bull case: the company's identified levers work well. The Army, Air Force and defense-prime work continues and grows, Florida Power & Light storm work returns toward its 2024 level of $1.3 million in an active hurricane year, and costs are held so losses shrink to about $4 million in the second half of 2026, $4.5 million in 2027 and $0.5 million in 2028, reaching breakeven. Present value of those losses is about $8.4 million, leaving roughly $1.6 million of the cash, about 8 cents a share. We give no value to earnings after breakeven because no contract, backlog or recurring revenue in the filings yet supports them; management's expectation of subscription revenue is a goal, not evidence. In every case the business owes almost nothing to lenders; its problem is not debt but that it spends more than it earns. The shares appear to be worth between zero and about 8 cents. For this to be wrong, the company would need to win recurring, multi-year government or commercial programs large enough to cover an operating cost base of roughly $15 million a year, and do so before its cash runs out. That is possible, but it has not happened yet, and the estimate gives it no value until a full year of self-funded positive cash generation is reported.