UAV Initiative · Strategy & Execution Plan

Where We Are, and What to Do Next

Everything from four conversations and seven research threads, synthesized into one plan. Built on the discipline of clear phases, decision gates, and kill criteria, not hope.
Contacts  Dimas, Serhii, Illia, Sayed Research  7 threads, 50+ sources Stage  Discovery to validation Updated  2026-05-26

What this is

Your decision document. It states honestly what you actually have, what the Illia call changed, the one strategy that fits the market, and a staged plan with gates so you spend only as fast as the evidence justifies.

Picture this: you fly to Dnipro in two weeks to sit with Dimas. You can arrive as an enthusiastic foreigner with a vague offer, or as a man with a named manufacturing partner, a diagnosed spec, a quality guarantee no one else is offering, and a clear ask. This plan is how you arrive as the second man.

Executive summary

Stop thinking of this as "build a better drone." The market is drowning in drones and short on trust. The winning play is to be the quality, traceability and intelligence layer on top of a proven Ukrainian manufacturer (Illia), sold first to a warm, funded customer (Dimas's brigade), with your AI/IoT as the irreplaceable piece, not the hardware.
~30 / 100
odds of this becoming a real money-maker, up from ~20 a week ago. Illia (a credible local manufacturer), commander approval, and a clearer software wedge raised it. Still capped by thin margins, the cash-flow trap, and the risk you end up a middleman. The number moves to ~40 if you lock a defensible software/IP layer and a real partnership; back to ~10 if all you can do is resell commodity drones.

01 What we actually have

Four assets, and most ventures would kill for the first one.

Warm, funded relationships

  • Dimas (UAV commander, 155th Brigade; WhatsApp contact "Dimas Boxing Trainer"): his commander approved working with you; he can buy with brigade money ("I can buy these drones from you by money of my brigade"); offered an in-person Dnipro meeting; sending specs.
  • Serhii (separate unmanned battalion, via Sayed): a second unit and channel.
  • Sayed: your 10+ year friend and the connector who opened both Serhii and Illia.
  • You are past the barrier that kills most ventures: getting the funded customer to engage.

Illia / the manufacturer

  • Illia Cheherst, founder, defense-only since 2022; team 20+; makes large drones (13in hexa, 18in, 22in) and deliberately avoids commodity 10in.
  • Makes his own electronics, flight controllers and ESCs, brand SkyPulse (spec sheet confirms FC F405 8S, GORO 12S 120A ESC for 15-22in builds). A genuine, strategically valued niche. (skypulse.ua / H743 variant are research-added leads, not in the spec sheet, verify.)
  • The 3-channel anti-jam controller: his flight controller works on 3 channels simultaneously (auto-reconnect if one is blocked) plus rare/uncommon channels. Note the commander is skeptical channel-hopping beats current jamming, get the spec sheet so he can react.
  • Network: a friend producing ~4,000 fiber FPV/month with government orders, plus another for +4,000; can cover 10,000/month.
  • ~2,000 kg of parts on hand now (~2,000 13in drones); can build a batch in ~2 weeks; can build prototypes (~90% chance he or a friend already has any requested prototype).
  • Quality-first ethos; refuses to float unpaid production ("Do you have money to pay for it?").
Due diligence (project-research, not in the call): his personal link to SkyPulse is not publicly verified; the hypothesis that his documented company is SageSeller (an Amazon tool) traces only to research, the transcript merely says he had "an Amazon project", treat as an unverified DD lead. Verify SkyPulse ownership and the 3-channel controller spec before committing capital.

Your stack

  • AI/IoT software company (Poland) + a small Ukrainian entity + AJ (VP Eng) and a team.
  • Capital and the willingness to invest, rare in a market where everyone is cash-starved.
  • A reputation you guard, and a Ukrainian family, home and property, deep local commitment and trust.
  • Western footing for the export / co-production angle.

Deep market intelligence

  • You now understand the procurement system (DPA, Brave1, DOT-Chain, e-points), the pricing tiers, the jamming reality, and the foreigner barriers better than most insiders.
  • You know the real failure modes and who is at fault.
  • You know the cash-flow trap before it bites you.
  • This is on the hub: The Catch, the combined intel, the price decode.

02 What the Illia call changed (the reframe)

One hour of insider education flipped the problem from a hardware problem you cannot win into a trust problem you can.

Insight 1

It is a trust problem, not a parts problem

Drones fail not because good parts do not exist, but because producers cut corners: a 10in battery on a 13in drone, non-original Molicel cells (a pack is ~12-16 small cells; you must cut it open to see what is inside), weak ESCs that overheat under load, cheap frames, fiber spec'd 20 km that snaps at 16 km. QC is random sampling (a few units from a batch of ~2,000), not 100%. And "you never know who is at fault."

Illia's thesis: the problem is the producer, not the parts (Chinese parts are fine, Ukrainian parts are good). The ~30%-reliable-at-15-km vs Russian ~80%-at-20-km figure is a project-research estimate, not contact-reported.
Insight 2

Your wedge is verifiable quality, not better cable

You cannot out-supply the global fiber market. But component traceability, anti-counterfeit cell checks, assembly QC, and survivable black-box failure-attribution are genuinely open, software/IoT-shaped, and nobody serves them. DOT-Chain creates the score; nobody supplies the forensic evidence behind it.

Research verdict: this slice is open and defensible; the AI talent in Ukraine is all pointed at combat, not the factory floor.
Insight 3

Cash flow is the real killer

The state pays after delivery, the Agency of Defense Procurement (the largest gov buyer) pays net 30, sometimes 60 days; producers self-finance and have waited ~2 months, sometimes never paid. (The "up to 6 months historic", "no advance until 3 months clean DOT-Chain history", and "brigade channel ~10 days" figures are project-research estimates, not contact-reported.) Either way the state is the post-delivery payer.

So: small batches, your capital with eyes open, never float large unpaid inventory. This shapes the whole model.

03 The strategy

Positioning

Do not be a drone maker. Be the quality, trust and intelligence layer on top of a proven Ukrainian manufacturer, selling first to warm funded customers, where your AI/IoT and your capital are the parts no one else brings.

Pillar 1 · Product

Land with a verified drone, expand into a trust platform

Phase one is dead simple: deliver Dimas a correctly-spec'd, quality-verified fiber FPV (Illia diagnoses the spec, sources good cells/ESC/fiber, QC'd) that just works, plus a thin differentiator: a component traceability + QC report that proves it. The concrete target to brief AJ against: a fiber FPV with guaranteed X-km fiber, a 4-5 hour-class flight battery, ~10 kg payload headroom, better cable + better battery than what Dimas flies today. Phase two evolves that into the real product: anti-counterfeit cell verification, assembly-line computer-vision QC, and a survivable telemetry black box that turns "it fell out of the sky" into "undersized battery, fake cells, producer X, batch Y." That is the defensible, recurring-revenue layer, and it is yours, not Illia's.

Pillar 2 · Partnership

You + Illia, with clear lanes and protected IP

Illia brings manufacturing, SkyPulse electronics, the supply network, certification and local-entity market access. You bring capital, the software/QA/traceability IP, the funded customers, Western reach and reputation. Lock lanes early: Illia owns hardware and production; you own the software/IP and the customer relationships. Use your Ukrainian entity; decide co-produce vs license; get heads of terms in writing before any money moves. This is the structure the whole market demands of a foreigner, and you now have the local half.

Pillar 3 · Go to market

Land and expand from Dimas's brigade

Beachhead = Dimas: warm, funded with brigade money, fast field-testing, honest feedback. Win that one unit with a drone that measurably works, capture the DOT-Chain rating and the proof, then expand to Serhii's battalion and outward on reputation. Cash-flow-safe: small batches, brigade channel, your capital with the collection lag priced in, advances pursued once eligible. Never the one left holding unpaid inventory.

04 How money is actually made

Near term (months): margin on improved, correctly-spec'd drones sold via Illia's network to Dimas's brigade. Thin but real, and funded by brigade money through the fastest channel. This proves the relationship converts to revenue.

Mid term (the real business): the software / QA / traceability layer, sold to producers (to lift their DOT-Chain rating and pass certification), and potentially to the MoD itself, plus SkyPulse electronics bundled in. This is recurring, defensible, high-margin, and the part where your AI/IoT is essential rather than optional. This is what turns ~30 odds into ~40.

Long-term optionality: Illia's larger-drone niche (18-22in, thinner competition, better margins); the export / Western-capital angle under the Drone Deals reciprocity framework (co-production funded from outside, built in Ukraine); and the fiber-import play Illia surfaced, good-quality fiber has no availability in Ukraine, he confirmed buyers exist for an imported container of 500 boxes of 20/30 km spools, and his last local quote was $34/km on 2026-05-06. Source from Brazil/China and test demand on the 19k-member Telegram market before committing a container. See the Brazil fiber-sourcing report.

05 The execution plan

Staged, with a decision gate at each step. Do not fund the next phase until the gate is met. Owners in the corner.

Rafael + AJ
Phase 0

Prep, before Dnipro

Now to the Dnipro trip (~2 weeks)
Gate   You have a diagnosed spec, an indicative price, and Illia checks out. If not, pause before committing travel and capital.
Rafael + Illia
Phase 1

Validate the wedge and the partnership

This week, including the Saturday (~2026-05-30/31) Krakow anti-drone meeting
Gate   A real partner + a real product path + a wedge that is defensibly yours. If your only role is passing orders to Illia's friend, restructure or walk.
Rafael + Illia + Dimas
Phase 2

MVP build and field test

After Dnipro (weeks 3-6)
Gate   Measurable reliability / hit-rate improvement, and the brigade wants to buy. If it does not beat what they already fly, iterate before spending more.
Rafael + Illia
Phase 3

First real order and the software product

Months 1-3
Gate   Repeatable funded orders at a workable margin + early software traction. Only then scale.
Rafael + Illia + team
Phase 4

Scale and expand

Months 3-12

06 Risks and how to handle them

RiskSevMitigation
You become an optional middlemanHighOwn the software/QA IP + the customer relationships + the capital. Get IP and lanes in writing. If you bring nothing Illia cannot replicate, you have no business.
Cash-flow trap (post-payment)HighSmall batches, brigade channel, your capital with the 30-60+ day collection lag priced in, pursue advances once eligible. Never float large unpaid inventory, the mistake Illia refuses to make.
Thin commodity marginsMedDo not compete on commodity FPV. Differentiate on verified quality + the trust layer; lean toward Illia's higher-margin big-drone niche.
Fiber demand is softeningMedSerhii's late-May steer: fiber is "really good but not really in demand right now because of its cost"; the momentum is retranslator drones and middle-strike (>30 km, up to 150-300 km) wing-type drones on Starlink/Mesh. A fiber-FPV beachhead (Pillar 1) is still valid for Dimas's brigade and KAMIK reports ~70% of its 2025 deliveries were fiber, but treat fiber as the wedge, not the destination, and keep the relay / recon-wing / long-loiter multi-camera concepts (which fit your AI/IoT/optics) on the roadmap.
Illia claims do not check outMedVerify SkyPulse founder/registration, spec sheets, the 4,000/month friend, QC. He offered other introductions if needed, so you are not locked in.
State swallows the QA/rating layerMedPosition as a producer-side tool that helps makers prove quality to the state standard, not a competitor to DOT-Chain.
War, corruption, certification delayMedBound your spend; treat as mission + business; heed Serhii's corruption warning; plan certification time into the roadmap.
Reputation (your stated fear)MedQuality-first, never ship bad kit to friends. Illia shares this ethos, a rare and valuable alignment. The QA layer is itself your reputation insurance.

Kill criteria, when to walk away

07 Immediate next actions (this week)

Do these now

  1. Chase Dimas for the written specs and problem descriptions. Everything downstream needs them. owner: Rafael
  2. Send Illia a focused ask: the specs to diagnose, plus SkyPulse founder/registration proof, FC/ESC spec sheets (incl. the triple-channel controller), the 4,000/month friend's profile, the QC process, and an indicative price for a Dimas-spec fiber FPV. owner: Rafael
  3. Saturday meeting (~2026-05-30/31, Krakow): with your anti-drone-company friend/professor (anti-drone company + Kyiv manufacturing plant); decide solo vs joint travel to Ukraine and scope the anti-drone adjacency. Separately, prep the Illia partnership agenda, lanes, IP, who finances, and the software wedge, not a hardware chat. owner: Rafael + AJ
  4. Write the one-page MVP + software-differentiator brief so Dnipro has something concrete behind it. owner: AJ
  5. Plan Dnipro: arrive with the diagnosed spec, an indicative quote, and the trust-layer pitch. Bring the maker into the room (or his proof) if you can. owner: Rafael

Open questions to close