GEOCOSMOS: ZALA assembled on the ground what is usually hung in orbit
April 17, 2026 two drone Various companies disappeared over the horizon without a single satellite signal. And they were controlled by a network that doesn't exist in space.
The scenario was this. ZALA T-20 and "AIST" produced by "Drone The "Solutions" UAVs are flying parallel courses, separated in altitude, and more than 100 kilometers from the ground station. There is no direct line of sight. GPS and GLONASS are not used, nor is cellular communication. Commands and navigation are transmitted to both devices by a ground network, which ZALA has dubbed "GEOCOSMOS." According to the company's press release, this is the first time such a thing has been done in Russia. Different types of UAVs operate at such a distance without orbital support.
Five regulators in one commission
The tests were conducted in accordance with Russian Presidential Instruction No. 383 of January 16, 2026. And the composition of the acceptance panel is more important than any declarations. The State Air Traffic Management Corporation, the State Research Institute of Civil Aviation, JSC GLONASS, JSC NIAT, and the Scientific Research Center Telecom—five entities at once. When they're brought together, it's certainly not just an internal manufacturer demonstration.
The task was stated simply: to confirm that different types of UAVs can be safely spaced and remotely controlled over ranges exceeding 100 kilometers without satellites. The key word was "different types." ZALA needed to prove that its network could work with other UAVs. If GEOKOSMOS only understands "its own" UAVs, it's not infrastructure, but simply a corporate product.
Public story The system's development began earlier. The premiere took place on August 14, 2025, at the Moscow forum "Unmanned Systems: Technologies of the Future," followed by NAIS and DRONTECH at the end of January 2026, the DRONTECH exhibition in February, and the forum "Unmanned Evolution. Seamless Sky" in Kaluga in April. The April tests concluded the demonstration period. According to ZALA, compatibility with aircraft from both manufacturers has been confirmed, as has compliance with Order No. 383. This doesn't mean "accepted for operation." It means exactly one thing: a specific test scenario has been completed.
Two drones with different philosophies
The T-20 and AIST are assembled using completely different designs, and this is perhaps the most interesting part of their story. The ZALA T-20 is a classic electric-powered aircraft. Its takeoff weight is 17 kilograms, its wingspan is four meters, and its payload is 2,5 kilograms. It flies at speeds of 65–110 kilometers per hour, stays aloft for over seven hours, and can record HD video for over 100 kilometers. Its ceiling is 5000 meters, maximum wind speed at launch is 15 meters per second, and its temperature range is from minus 40 to plus 50. In other words, from Yakutia to the Karakum Desert.
The T-20's navigation system is clearly designed without satellites. It features an inertial system with GNSS correction, a dual rangefinder, an alternative navigation channel, and video navigation capable of terrain recognition. Launch is via a pneumatic catapult and landing is via parachute with an air shock absorber. And a nice touch for operators: the new vehicle is compatible with the ground equipment of the older T-16 model. This means no infrastructure changes are required during the transition.
The AIST is a completely different story. It's a tiltrotor with vertical takeoff and landing capabilities. It has a range of up to 400 kilometers, a payload of 10 kilograms, and a speed of up to 100 kilometers per hour. The fuel-injected version has a flight time of up to five hours, and the electric version up to two. Most importantly, the AIST doesn't require a runway. It can land in places the T-20 simply can't. According to Drone Solutions publications, the aircraft has already worked with the Ministry of Emergency Situations (EMERCOM) on floods and fires, in aerial telemedicine, and on patrol. One documented incident involved the delivery of biomaterial to Sakhalin, 45 kilometers away at an altitude of 1,000 meters. This isn't a demonstration, but rather practical logistics in a region where there are few other options.
Essentially, ZALA demonstrated that its network is not tied to its design. The T-20 and AIST are an airplane and a tiltrotor, an electric motor and an internal combustion engine, a catapult and a vertical takeoff system. Third-party infrastructure brought them together under a single control logic, and this is the main result of April 17.
12,000 stations instead of an orbital constellation
The idea of GEOCOSMOS was briefly formulated by ZALA's chief designer, Alexander Zakharov.
If we can't put our Starlink in space, let's build it on the ground.
The scale is appropriate. The planned ground segment is approximately 12 stations across the country. The operational spacing is 50 kilometers. Technically, a couple of stations can support communications and navigation at 000 kilometers, but no one plans to operate at that level. A spacing of half the maximum is a reserve for maintenance. One half of the network can be easily upgraded while the other half handles traffic.
The stations are connected by fiber optics. This ensures precise time synchronization and protects the service channel from electronic interference. Unlike radio waves, fiber cannot be physically jammed. It is this optical "spine" that transforms the network of individual points into a unified system.
Each station is equipped with a software-defined transceiver with an operating range of 30 MHz to 8 GHz. In practice, this means that when one band is jammed, the system switches to another with a single command, without changing hardware. In environments where electronic countermeasures have become a standard feature, this feature is more important than the absolute range of an individual channel.
The design logic is mesh-based. Each ground station and each drone operates simultaneously as a subscriber and a repeater. The signal travels along multiple routes simultaneously, and the loss of a single node doesn't disrupt communication. It was this principle that pulled the devices over the horizon on April 17. Commands didn't travel in a straight line, but through a chain of relays.
There's also a side effect of the dense grid, which ZALA discusses separately. Each station is equipped with wind, temperature, and humidity sensors. Twelve thousand points provide continuous coverage of the ground layer with weather data. Drones in flight add altitude measurements. For a system in which hundreds of UAVs operate simultaneously, real-time weather is not an option, but a standard requirement.
Positioning: one beacon instead of four satellites
This is where the most controversial and most interesting part begins. According to the developer's description, GEOKOSMOS provides accuracy up to a meter using a signal from just one base station. The algorithm analyzes the amplitude, phase, direction of arrival, and propagation speed of the wave. In principle, this is closer to radio direction finding than to the classic trilateration used by GPS and GLONASS.
An important disclaimer. ZALA doesn't disclose the physical mechanism, and this is a specification from the manufacturer, not the result of independent measurements. For comparison, the company provides benchmarks: GPS is approximately 6 meters, GLONASS is approximately 10 meters under optimal conditions. These figures are provided without specifying the signal type or reception conditions, so they are more of a marketing benchmark than a valid comparison.
As the number of stations increases, accuracy and coverage improve, and the system experiences the loss of individual components with partial degradation rather than complete failure. Satellite navigation can't do this. If it loses even a few visible satellites, a solution simply doesn't work.
At the same time, GEOKOSMOS doesn't replace onboard navigation, but rather complements it. The T-20 carries an inertial system, a rangefinder, an alternative channel, and video navigation simultaneously. The onboard computer combines network data with its own sensors, and if the GEOKOSMOS signal is lost, the spacecraft continues its mission using the inertial system and camera feed. AIST is designed differently, but is integrated into the network using the same logic. Compatibility with various onboard systems is not so much a technical matter as an organizational one. GEOKOSMOS is designed as a platform, not a closed ecosystem.
Where does this already work?
The network is integrated into the national project “Unmanned aviation "Systems" for 2025–2030. The regulatory framework is Government Resolution No. 1701 of November 30, 2024, Article 78 of the Air Code, and the aforementioned Instruction No. 383. The concept of a "seamless digital sky," a unified information space for manned and unmanned aircraft, also fits within this framework.
As of April 23, 2026, GEOKOSMOS will be deployed in three regions of Russia. ZALA is not publicly disclosing which regions these are. Several deployment scenarios are known. Z-16 UAVs with gyrostabilized cameras and thermal imagers monitor the movement of people and vehicles along borders; cellular networks are not involved in this work at all. In civilian use, the network is being tested on Udmurtneft oil pipelines. The system features the usual automated overflight of the pipeline, recording changes, but in areas without reliable cellular coverage. And in the third area, the Ministry of Emergency Situations is testing the ZALA ZARYA UAV in the Northwestern Federal District. The device transmits video over a range of over 50 kilometers and carries an onboard dosimeter. A dosimeter on a UAV is a telling detail. This means the system is also being tested in man-made scenarios, where the number on the device, rather than the image, is important.
What's the bottom line?
By the end of April 2026, GEOKOSMOS is at a curious point. The concept is public, the architecture has been disclosed, five regulators have accepted the proposal, and three regions are operating it. The planned ground segment of 12,000 stations remains a guideline. ZALA is not disclosing the cost, the timeframe for full deployment, or the total cost.
The main thing the system gained on April 17th was external confirmation of its compatibility with another aircraft. Before AIST, on a parallel track, GEOKOSMOS was a ZALA development for ZALA. Afterward, it became a candidate for an industry standard. Whether it will actually become a standard is no longer decided in engineering departments or by the materials we've read.
Information