For investors · Caligo Dynamics Inc. · Delaware C-Corp · Pre-seed
Self-funded to a flying prototype, patent pending, AWS 10,000 AIdeas Innovation Award 2026. Raising to build the aircraft that can carry a real payload and to fly the rotation with more than one aircraft.
01 · Why the gap is widening
Nothing on this page is a criticism of the fire service. Crews hit their marks. What changed is what is burning, and no amount of professionalism moves a fire station closer to a house.
UL's Fire Safety Research Institute burned four pairs of identical rooms: one furnished with natural materials, one with modern synthetics. The legacy rooms took 29 to 34 minutes to flash over. One never did. The synthetic rooms took 3:20, 3:40, 4:45 and 4:50.
NFPA 1710's four-minute standard is a travel-time target, and it says nothing about the call handling and turnout that precede it. Measured call-to-arrival in England for the year ending March 2026 was 9 minutes 25 seconds. Stations sit where councils put them decades ago and roads are what they are. The crews are not the variable here.
A 2025 study of 15,576 Dutch building fires found the probability of large damage rises roughly one percentage point for every additional minute of response, near-linearly3. Each minute past the ten-minute mark is worth more than the one before it.
02 · Why now
Autonomous first response has been technically plausible for years and administratively impossible. That changed recently enough that most people have not noticed.
In April 2025 the FAA replaced the drone-as-first-responder approval path with an expedited Part 91 waiver reviewed by one office rather than two. Between 15 April 2025 and 23 February 2026 it approved 1,004 of them for US public-safety agencies, including 873 to law enforcement and 113 to fire and emergency management. Each runs four years and permits flight beyond visual line of sight8.
7,783 fires mapped in 25 of the 27 member states burned 1,079,538 hectares, the highest EU-27 total in the EFFIS record, which begins in 2006, and close to double the 2006-2024 average. Germany, Spain, Cyprus and Slovakia each set national records in the same year. Our launch market had its second-worst season since satellite monitoring began.
The January 2025 Los Angeles fires produced roughly $40 billion in insured losses on both Munich Re's and Swiss Re's independent counts, around 37% of the entire world's insured natural-catastrophe losses for the year, from one event, over a few days9.
03 · The wedge
In our launch market, the national forestry directorate went into the 2025 fire season owning nine helicopters and fourteen aircraft. To cover the season it leased eighty-one more (68 helicopters and 13 fixed-wing) at a reported cost of about ₺10.2 billion for the year12.
That is a budget line that already exists, that recurs annually, that grows with every bad season, and that is spent on assets which cannot fly at night, cannot fly in smoke, and cannot be first on scene.
Two figures, honestly. The national forestry service counted 81,473 hectares burned on forest land in 2025; the EU's satellite service mapped 162,188 hectares across all land cover. Both are official; they measure different things. We quote both rather than the more flattering one.
Every number in this box comes from a commercial research vendor rather than a primary source. They are not independently auditable and they disagree with each other. We present them as estimates, with the vendor named, because pretending otherwise wastes everybody's diligence time.
The four questions you are about to ask
Isn't this just detection? The literature says faster detection doesn't pay for itself.
It does say that, and the paper is worth reading. Huang & Wichmann (PLOS ONE, November 2024) analysed 4,934 Western Canadian fires and found each additional hour of reporting delay raises suppression cost by roughly a quarter of one percent. Their own conclusion is blunt: investment in detection is not justified on suppression-cost savings alone10.
We agree with them, and it is why we are not a detection company. Detection is a prerequisite, not the product. The value is in avoided loss, which is a different and much larger number: the Australian National University modelled responding to all fires within thirty minutes and put the saving at AU$8.2 billion over thirty years11. A firm that sells you a faster alarm is selling into the weak case. We intend to sell the intervention.
Why hasn't DJI or Skydio already done this?
Because they build superb sensors that fly, and this is not a sensing problem. It is a work problem: lift a mass, hold a geometry in thermal turbulence, discharge it accurately, and then do it again without pausing. That demands a payload architecture, an adaptive controller, a rotation logistics model and a command-authority story that satisfies a fire chief, none of which improve a camera platform's core business.
It is also a channel they do not want. Selling into fire services means safety cases, procurement cycles and shared liability for an operational outcome. That is a moat if you are willing to walk into it and a reason to stay away if you are not.
This is one founder with an off-the-shelf drone.
Largely, yes. The flight-control integration, the ground station, the thermal and detection lane, the automated flight-data pipeline and the beyond-line-of-sight link were built and flown by the founder, on hardware bought with his own money, and they work, including the unglamorous parts, like an aircraft that logs every flight without being asked and picks its own video bitrate from the radio it woke up on.
That is the signal. A team that has already put an autonomous aircraft in the air and knows precisely which of its own claims are simulation is a considerably better bet than a larger team that has not yet found out.
Where is the company, actually?
Caligo Dynamics Inc. is a Delaware C-Corporation. You can invest in it on standard instruments today, with no cross-border structuring exercise.
Engineering and flight testing run out of our launch market, which burned 162,188 hectares in 2025, has extensive terrain to test over, and where the cost of iterating on real hardware is a fraction of the Bay Area equivalent. The problem is outside the window and the runway goes further.
04 · Business model
No large capital purchase for the customer. Impact is measured in the field, and the price is reviewed as the measured value grows.
Covers the aircraft, nest, software and updates. This is a proposed starting price, not current revenue, an offer or a contract.
≈ ₺241 million a year in recurring revenue potential from one city. Simple arithmetic to show scale, not a forecast or a commitment. The launch market has 1,401 municipalities, 30 of them metropolitan; Istanbul's fire service answered 27,820 fires in 2025.
Sources: İstanbul Fire Department 2025 activity data (ANKA, 7 January 2026); TÜİK municipal data (2024); IMARC Group, Firefighting Drone Market 2025-2034 (~8.2% annual growth). Vendor market estimates are not primary sources.
05 · What the money buys
We have written this as proof points rather than deliverables, because a deliverable can be redefined and a fire that goes out on camera cannot.
Reliable autonomous takeoff, mapped navigation at held altitude, arrival scan, live thermal and visible perception with hotspot alerting, and automatic flight-data capture on every armed flight.
Validate the beyond-line-of-sight link at 500 m and 1 km with recorded range-versus-link-quality data. Move to a purpose-built airframe with genuine payload capacity. First sustained single-aircraft suppression on a controlled test fire.
Three aircraft flying a real orbital rotation with live hand-over, so suppression continues while individual aircraft cycle. Detection moved onto the aircraft so it survives link loss. Begin seeking pilot partners among fire services.
A long-endurance vertical-take-off platform for wildland distances, where a multirotor's physics stop being generous. Formal certification path and operational safety case developed alongside the airframe, not after it.
Dates are intentions, not commitments, and depend on the size and timing of the round. We will tell you when one moves and why.
The brief
The investor brief carries the raise size and use of funds, the full capability status with flight logs behind it, the unit-economics model and the risk register, including the risks we have not solved. We will send the raw flight data if you want it.
Sources
