Why This Matters Now

Drones costing 30,000 rubles (~$320) destroy assets worth billions

A single FPV drone costing 30,000 rubles, hitting a power substation, leaves an entire district without electricity for months. Hitting an oil storage facility causes a fire with billions in damages. This happens every day.

30K ₽

cost of
one hostile drone

150 km/h

attack
speed

5 seconds

from detection
to impact

billions ₽

damage from
a single hit

Existing defenses are failing

Anti-aircraft missiles cost tens of millions of rubles — a thousand times more than the target. It's like using a cannon to shoot a sparrow.
Jammers (EW) are useless when a drone flies on a pre-loaded route with no radio signal.
Small arms — hitting a plate-sized object flying at 150 km/h is nearly impossible.

A fundamentally new solution is needed — one as cheap and scalable as the threat itself.

Our Solution — the ECHELON-Z System

Sphera Platform + Sokol Interceptor

The ECHELON-Z system is two products in one: the Sphera detection platform locates hostile UAVs, and the Sokol interceptor drone catches and disables them. No human involvement. In seconds. Cost per interception — 500 rubles (~$5).

3D MODEL — SOKOL

Specifications

Size	Fits in a palm (25x28 cm)
Weight	Under one kilogram (870 g)
Speed	250 km/h — faster than any hostile drone
Launch readiness	3 seconds — launches automatically
Range	3 km around the protected site
Neutralization method	Deploys a cloud of polyurethane foam — foam envelops propellers, drone falls
After interception	Returns, reloads in 10 seconds, ready again
Human involvement	Fully autonomous — from detection to neutralization

500 ₽ (~$5)

Consumables per interception

Foam canister — 300 rubles, component wear — 200 rubles. The interceptor itself is undamaged, returns and is ready for the next sortie in 10 seconds. Lifespan — hundreds of flights.

350K ₽ (~$3,700)

Cost of one interceptor

At serial production of 1,000+ units. Reusable — pays for itself in a few interceptions.

∞

Number of uses

The interceptor is reusable. Only the foam canister is replaced. Airframe lifespan — hundreds of flights.

Why Foam

The simplest solution — the most effective

Ordinary expanding foam, the kind used to insulate windows. Except instead of a window — a hostile drone.

✓

Not a munition

Foam is a civilian material. No weapons certification required. No explosives licenses needed. Can be manufactured at any facility.

✓

Safe for people

If foam lands on the ground — no harm done. Non-toxic, non-flammable, biodegradable. Unlike buckshot or ram debris — safe for the civilian population.

✓

Works on any drone

Foam covers everything: propellers, cameras, sensors. Works on small FPVs and large reconnaissance drones alike — size doesn't matter.

How It Works

Fully Autonomous

No human involvement. The system detects the threat, launches the interceptor, guides it, and neutralizes the target on its own. The operator only observes on screen.

STEP 1

Detection

Radar picks up an object. Or detects the sound of propellers. Or intercepts the control radio signal.

0 sec

STEP 2

Identification

The computer determines: this is a hostile drone, not a bird and not a friendly aircraft.

1 sec

STEP 3

Launch

The Sokol interceptor automatically launches from the launch platform. Accelerates to 250 km/h.

3 sec

STEP 4

Approach

Onboard camera and radar track the target. The computer calculates the intercept point.

10-20 sec

STEP 5

Foam!

At 3-5 meters range — a cloud of expanding foam is released. Propellers are coated — drone falls.

0.5 sec

STEP 6

Ready Again

The interceptor returns. A crew member replaces the foam canister in 10 seconds. Back in action.

10 sec

What We Protect

Any site within a 3 km radius

⚡

Power plants and substations

One drone hitting a transformer — a district without power for months. Damage — billions of rubles.

🛢

Oil storage and refineries

A hit on a storage tank — fire, environmental disaster, production shutdown.

🔥

Gas pipelines

Compressor stations, pumping facilities. Transit disruption — economic losses.

🎖

Military installations

Airfields, depots, barracks, command posts. FPV attacks are an everyday reality.

🏘

Border cities

Residential areas, hospitals, schools, government buildings. Civilians under attack.

🌉

Bridges and transport

Railway junctions, bridges, airports. One drone — traffic disruption for days.

The Key Factor — Price

Interception for 500 rubles (~$5)

For the first time, the cost of defense is comparable to the cost of the threat. You can afford a miss and a retry.

Anti-aircraft missile

30,000,000 ₽ (~$320K)

Reliable, but 1,000x more expensive than the target

AA gun (100 rounds)

500,000 ₽ (~$5.3K)

Hard to hit — 50% miss rate

Jammer (EW)

0 ₽

Doesn't work✗ if drone flies without radio signal

SOKOL

500 ₽ *

90-95%, reusable✓

* 500 ₽ — cost of consumables (foam canister + wear). The interceptor is reusable, lifespan — hundreds of flights. Interceptor cost: 350,000 ₽, amortization per interception over 500 flights — 700 ₽. Total cost per interception: ~1,200 ₽ (~$13).

Cost to protect one site

One Sokol interceptor	350,000 ₽ (~$3,700)
Launch platform (8 units)	800,000 ₽ (~$8,500)
Site protection kit 32 interceptors + 4 launchers + radar + command system	8M ₽ (~$85K)
Annual consumables foam canisters, batteries, spare parts	500,000 ₽/yr (~$5.3K)

8M ₽ (~$85K)

Full site protection

32 interceptors, 4 launch platforms, detection radar, Sphera command system. Protection radius — 3 km. Operates without an operator, 24/7.

For comparison

A single surface-to-air missile system costs 1 to 10 billion rubles ($10M–$106M). For that money, you can protect 125 to 1,250 sites with Sokol systems.

Swarm Attack

20 drones at once?

Sokol operates as a swarm. 32 interceptors launch simultaneously, automatically distributing targets among themselves — without operator involvement.

Scenario

Attack of 20 FPV drones on a power plant. Simultaneously, from different directions.

Second 0: Radar detects 20 targets at 3 km range.
Second 3: 20 Sokol interceptors launch. Each assigned to its own target.
Second 20: 18 of 20 drones neutralized with foam.
Second 23: Reserve interceptors launch against the 2 that broke through.
Second 35: All 20 drones neutralized.

Result: power plant intact. Expenditure: 22 foam canisters (11,000 rubles / ~$117).

Numbers

35 sec

from attack to
full neutralization

11,000 ₽

cost of repelling
a 20-drone attack

humans involved
in the interception process

The Brain of the System

The Sphera Platform

Sokol is the fist. Sphera is the eyes and the brain. The core of the ECHELON-Z system — it merges all sensors into a single picture, decides what to attack, and commands the interceptors. Built on the AeroTab mapping engine and the Russian AI platform NeuralGate.

👁

Sees everything

Radars, cameras, microphones, radio signal sensors — all on one screen. No drone approaches undetected.

🧠

Thinks for itself

Artificial intelligence powered by the Russian NeuralGate platform distinguishes hostile drones from birds and friendly aircraft. Makes decisions instantly.

🎯

Commands the interceptors

Automatically launches the required number of Sokols, distributes targets, coordinates the swarm.

📊

Full reporting

Every interception is recorded: video, coordinates, result. For command — a complete real-time picture.

🔒

Works offline

Fully autonomous. No dependency on central connectivity. Fits in a single hardened case.

📈

Scales with the mission

From protecting a single site to an entire region. Multiple kits unite into a single network.

US Analog

The Pentagon uses a similar system — Palantir Maven. Contract worth $12 billion. 20,000 users in the US Army. NATO adopted it within 6 months. It is not available to us — but we know how it works, and we are building our own based on Russian technologies: the NeuralGate AI platform + the AeroTab mapping engine.

Market

Who we sell to

DEFENSE

Ministry of Defense

Protection of military installations, positions, airfields. Massive drone attacks — every day. Demand: thousands of kits.

State defense order

INFRASTRUCTURE

Critical facilities

Power plants, refineries, gas pipelines. The law mandates protection. Demand: hundreds of sites across the country.

Corporate clients + government

EXPORT

Friendly nations

The drone threat is a global problem. No comparable price/performance ratio exists in the world. Potential: tens of billions.

BRICS, CIS, Middle East

Readiness

What we already have

✓ Complete technical documentation

32,000 lines of specifications. 34 sections. From electronics to combat tactics. Data model covering 91 object types. Ready for handoff to developers.

✓ World analog studied

Full analysis of the American Palantir Maven system: 218,000 lines of source code, 37 technical documents, 14 demo videos, 485 UI screenshots.

✓ Component base identified

All interceptor components are commercially available off-the-shelf. No dependency on sanctioned parts (alternatives exist). Prototype assembly can begin immediately.

          Timeline
          
            Interceptor prototype2 months
Flight testing4 months
Sphera platform prototype3 months
Integrated testing6 months
Serial production start12 months

        

Initial team

3-4 engineers (electronics, software, airframe design, testing). Scaling to 12-16 people by serial production.

Who Is Behind the Project

The Team — 7 Key Specialists

Each team member mitigates a specific investment risk: operational, engineering, regulatory, and financial. Combined experience — over 120 years in aviation, defense industry, investment, and public administration.

CEO

Chief Executive Officer

21+ years of operational management. Experience leading a trading and manufacturing group: coordinating production, logistics, procurement, and sales. Head of airline representative office in China. Fluent Chinese and English — critical for component sourcing and contract manufacturing.

STRATEGY

Business Development & Partnerships

13+ years in private equity and M&A. Built industrial cooperation networks of 12+ factories. Raised over 600M rubles (~$6.4M) in investment for manufacturing projects. International connections (Middle East, Africa) — export channels.

CTO

Chief Technology Officer / Digital Solutions

Creator of the AeroTab electronic pilot tablet (in operational use). Projects with GosNIIAS, FSB Administration, Almaz-Antey. Deep expertise: FPGA/SDR, GNSS, EW-resilience, phased arrays, multi-sensor navigation. Full cycle from prototype to production including China supply chain.

Chief Engineer

MIPT graduate. Experience as Chief Designer for aviation projects per GOST RV 15.203. Full R&D cycle: from concept to state testing and serial production handoff. Led a UAV program: prototypes, testing, small-batch production. Proven sanctions resilience — factory operating 3 shifts for 2 years without foreign support.

PRODUCTION

Head of Tooling & Mold Manufacturing

26 years in mold tooling and serial production. Full cycle: design, machining, assembly, launch. In 6 months — designed 10 and launched 30+ molds. Reduced project costs by 30% (~8M rubles) through design optimization. Experience with Chinese manufacturers.

INTEGRATION

Head of Systems Integration

25+ years in aviation instrumentation. MAI graduate, postgraduate at Gromov Flight Research Institute. Deputy Chief Designer at FGUP Piloting Research Center. Chief Designer at Aviaavtomatika and RPKB. Solutions used by Sukhoi, Irkut, GosNIIAS, TsAGI, Kronshtadt, RZD. VxWorks 653 replacement, SSJ-100 "electronic bird."

FINANCE

Investment Division

20+ years in investment finance. Structured deals exceeding 40 billion rubles (~$425M). Experience in PPP, M&A, financial modeling. PhD in Economics. Corporate track: AFK Sistema, participation in a ~$1.3B deal. Developed a PPP model for air ambulance services with Russian Government backing.

Government Relations Advisor

Retired Major General, PhD in Military Science. Academy of the General Staff. Deputy Head of Roskomnadzor, Deputy Minister of Digital Development of the Russian Federation. Led R&D in the telecom sector. Navigating regulatory requirements and liaising with federal authorities.

For Technical Specialists

Technical Details

Sokol Interceptor — Specification

Flight Characteristics

Max speed	250 km/h (69 m/s) on 8S LiPo
Cruise speed	120 km/h (33 m/s)
Max acceleration	8g, thrust-to-weight ratio 10:1
Reaction time	<3 sec (catapult)
Radius	3 km from launch point
Flight time	6 min (loiter) / 2.5 min (max speed)
Ceiling	500 m AGL
Operating temperature	-30°C to +50°C
Max wind	15 m/s (54 km/h)

Mass & Dimensions

Takeoff weight	870 g (ram) / 960 g (foam)
Configuration	Quadcopter X-frame 5", foldable
Frame	3D-printed PA12-CF, snap-in modules
Span	280 mm (motors)

Electronics

Compute	NVIDIA Jetson Orin Nano 8GB (40 TOPS) Alternative: Rockchip RK3588s (6 TOPS, $40)
FC	STM32H7 + BetaFlight, IMU BMI270
ESC	4-in-1 BLHeli32 60A
Motors	2207 2750KV x4 (speed-optimized)
Battery	LiPo 8S 1300 mAh (29.6V, 38.5 Wh)
RGB camera	Sony IMX477, 12 MP, 120° FOV, 1080p@30fps
IR camera	FLIR Lepton 3.5, 160x120, LWIR 8-14 um
Radar	TI IWR6843AOP, 60-64 GHz mmWave, 30-100 m
RF scanner	Custom SDR 2.4/5.8 GHz, -90 dBm
UV detector	SiC photodiode, FFT propeller flicker
GPS	u-blox M10 + GLONASS

Sensor Fusion & Guidance

CV detection	YOLO v8n INT8 @ 30-45 fps (Jetson)
Tracking	ByteTrack + Kalman filter
Guidance	Augmented Proportional Navigation (APN), N=4-5
Terminal homing	CV-only, last 200 m, CEP <0.3 m
Sensor-to-actuator	38 ms (onboard) / 28 ms (via GCS)
Detection classes	multirotor (S/M/L), wing, FPV, loitering munition, bird

Multi-sensor fusion	RGB + IR + mmWave + RF + UV → unified track
Predictive RF	Protocol classification (DJI/ELRS/CrossFire) → type → speed
Target behavior ML	LSTM 2x64, 120 KB, prediction +0.5 sec
Paired interception	"Beater + hunter", hit rate 80%→95%
Decision checklist	8 mandatory checks (type, confidence, geofence, friendlies, civilians...)
Black box	All decisions recorded at 30 Hz, 64 GB eMMC, AES-256

Communications & Navigation

Priority 1	Tethered optical link (ultra-thin SMF 0.25 mm, up to 500 m, latency <1 ms, EW-immune)
Priority 2	LoRa FHSS on non-standard frequencies: 150-174 MHz (VHF), 400-470 MHz (UHF), 1900-2100 MHz, 2100-2500 MHz (S-band). 128 channels, 20 hops/sec. Sensitivity -137 dBm (20 dB below noise floor)
Priority 3	ELRS 900 MHz (standard)
Swarm mesh	ESP32-S3 2.4 GHz, deconfliction, swarm relay
GPS-denied navigation	VSLAM (Visual SLAM on preloaded map, ORB/SuperPoint + EKF fusion with IMU, accuracy ±2-5 m)

Neutralization (Foam)

Composition	Single-component polyurethane (MDI), CO₂ + HFO-1234ze (non-flammable), nylon microfiber 0.5 mm
Canister	150 ml, 8-10 atm, 60 g
Expansion	x30-40 → 4-6 liters of foam
Mechanism	Pyro charge 0.3 g, toroidal nozzle 60°x360°, range 3-5 m, <10 ms
Curing	3 sec (tack) → 30 sec (hard)
Safety	Non-toxic, non-flammable (HFO GWP<1), biodegradable 6-12 months, not a munition

Protection

Anti-spoofing	Multi-modal verification (min 2 of 4 sensors), ensemble CV (YOLO + RT-DETR)
Anti-laser	Notch filter 532 nm on lens + software glare detection + IR fallback
IFF	ELRS transponder + coordinate IFF via Sphera + mesh-IFF between interceptors
Zeroize	Firmware/key wipe <3 sec (triggers: command, 24h timeout, tamper switch, geofence)
EMC	Spatial separation + TDD (RF scanner during mesh pauses) + 10 g shielding

Sphera Platform — Architecture

Technology Stack

GIS engine	AeroTab — mature mapping backend and frontend for aviation and drone maps (proprietary, in production)
Frontend	React + TypeScript + AeroTab Map Engine (cartography) + Sigma.js (graph)
Backend	Python (FastAPI) + Go (high-load) + gRPC/REST/GraphQL
Data	Apache Kafka (bus) + Flink (streaming) + Spark (batch)
Ontology	PostgreSQL + Apache AGE (graph), up to 10B objects
Storage	MinIO (S3), TimescaleDB, Elasticsearch, Redis, Apache Iceberg
ML/AI	PyTorch, ONNX Runtime, TensorRT, MLflow, pgvector (RAG)
Infrastructure	Kubernetes (k8s/k3s), Cilium (eBPF), ArgoCD, Cosign
AI platform	NeuralGate — Russian AI platform incorporating best practices from industry leaders (Anthropic/Claude, OpenAI/ChatGPT, xAI/Grok, Alibaba/Qwen, DeepSeek). Pluggable LLM hub: local models (Llama, Mistral, GigaChat) or cloud

Key Modules (24 of 34)

Core	Ontology (91 types, 900+ fields), Entity Resolution, Provenance
Data	50+ connectors, CDC, Sensor Fusion (Kalman filter)
AI	CV Pipeline (47 classes), LLM Hub, Agent Studio (4 levels), Model Registry
C2	Battle management, OODA loop, operations documents, task Kanban
Comms	Auto channel switching, 5-level QoS, store-and-forward, MANET mesh
Security	ABAC, GOST crypto (modular), 5 classification levels, append-only audit
Deployment	Hub-Spoke, federation, air-gapped (USB), blue/green, Cosign

Configurations

Tactical edge	1 server: 14 CPU, 42 GB RAM, 1 GPU, 20 users
HQ	3-5 servers, 500 users, 50 video streams
Data center	10-50+ servers, 25,000 users, 10B objects

Palantir Maven Research

SDK analyzed	218,289 lines (TypeScript 89,870 + Python 128,419)
Packages	@osdk/gotham (Target Workbench, Gaia Maps), @osdk/foundry (25 subpackages), foundry-platform-sdk
Documentation	50 HTML pages, 37 PDF whitepapers (131 MB)
Video	14 files (579 MB): AIP Logic demo, Ontology Hydration, TITAN demo, Gotham hero
UI screenshots	485 images (260 MB): Gotham, Foundry, AIP, Apollo, TITAN
AIP Now catalog	154 ready-made AI workflows with descriptions, tags, implementation graphs
Repositories	16 cloned (blueprint, conjure, atlasdb, osdk-ts, defense-sdk-examples...)
Forensic	JS analysis (QA domain qa.fictivepalantir.net, 15 subdomains, 35 routes), Contentful Space ID, hidden pages (ShipOS, Chain Reaction, Sovereign AI OS)

Interceptor prototype	2 months
Flight testing	4 months
Sphera platform prototype	3 months
Integrated testing	6 months
Serial production start	12 months

ECHELON-Z:SPHERA + SOKOL