What is claimed, what is proven

High-altitude flight is a genuine test for small UAVs: thin air cuts rotor and wing lift, batteries lose capacity in the cold, and strong valley winds complicate station-keeping, so a trial at 16,500 feet is a reasonable way to show a system works outside benign conditions. Swarm coordination and vision-based geolocation are both useful building blocks for reconnaissance and targeting, and a swarm that coordinates its aircraft is a step beyond individually piloted drones if it holds up under independent testing.

The geolocation method itself deserves a careful reading. Converting a camera pixel to a ground coordinate from telemetry, as described, depends on the aircraft's own GPS position and attitude estimate, so its accuracy is bound to the quality of that telemetry; a sub-2 metre figure would need to say against what reference it was checked, and whether it held at the ranges and angles a reconnaissance or strike mission would require.

The company has not published how many aircraft flew together, what tasks they performed, what sensor and communications fit each carried, or whether any armed-forces personnel observed the trial, so this remains a company-reported demonstration. A clear account of swarm size, the datalink between aircraft, and an independent check of the geolocation accuracy would let readers judge how close this is to a fieldable capability.

A hybrid fixed-wing and multirotor layout is a sensible choice for a tactical ISR UAV. The wing gives better endurance and cruise efficiency than a pure multirotor, while vertical take-off and landing removes the need for a runway or catapult, which matters for forward units without prepared strips. Two to three hours of endurance fits a close-in reconnaissance role rather than a persistent one.

The certification needs reading carefully. DGCA type certification is a civil airworthiness process, and the company describes this one as certification for training purposes. That is a useful mark of design and production discipline, and it opens civil and training markets, but it is not a military qualification, and the Army's acceptance of the aircraft it has bought rests on its own trials. The reported ₹72 crore contract and first delivery are the more significant defence facts here, though the number of aircraft delivered has not been published.

A small turbojet airframe in this weight class is a plausible way to get a loitering munition to cruise speeds well above propeller-driven designs, trading some endurance and cost for much shorter time to target and a harder target for air defences to engage. The specific figures reported this week do not yet agree with each other: a 290 km range at Mach 0.8 (roughly 980 km/h) would take under 20 minutes of flight time, well inside the stated 40 minute endurance, so there would be meaningful loiter time left over, but the same coverage headlines a 1,000 km range that the detailed specification does not support. Readers should treat the larger figure as unconfirmed until the company publishes a consistent specification.

The more basic open question is what kind of weapon this is. "Loitering munition" implies the aircraft can circle and search for a target before committing to an attack, which is a different engineering and doctrinal problem from a jet-powered glide bomb flying a pre-planned route to a fixed target, even though both can share an airframe and a turbojet. JSR Dynamics has a real track record in precision-guided munitions through its Gaurav family, which is a reasonable basis for a new product, but the company has not yet said whether the JPLM has flown in any configuration, or clarified its relationship to its other jet-powered products.

This is a company announcement and a static display, not yet an aircraft that has flown, and its figures are design objectives. Taken together they are broadly consistent for a subsonic strike drone. A combined thrust of 5 kN for a 1,000 to 1,500 kg aircraft is a thrust-to-weight ratio of about 0.35 to 0.5, higher than an airliner that cruises at Mach 0.85, and 1,500 km at the stated 500 to 600 km/h cruise needs a little under three hours aloft, which fits "more than two hours" of endurance if the fuel fraction allows.

The gap is in the role rather than the range. The same thrust class that suits efficient cruise gives little margin for climb, acceleration and sustained turns, which is what escorting fighters demands. The company is candid about where the programme stands: Supreme Heavy's detailed design is about half finished, and the smaller Supreme Lite has completed taxi trials but not flown, with first flight now estimated for late November or December 2026 after an earlier target was missed.

An AI-piloted combat aircraft that makes mission decisions with minimal human intervention also raises a question the published material does not answer: what human authorisation remains over weapon release. Funding is the other open point, since about $3.44 million has been recorded against the company's own estimate of $100 million for development.

The significance of this order lies in where the drones were tested. The Army ran the trials at up to 17,500 feet in Ladakh and Spiti, not on a plains test range, before choosing a supplier for a small FPV buy. FPV drones have become a staple of the front line in recent conflicts, and the Army is now buying them in quantity from Indian firms.

High altitude is unforgiving for small rotorcraft. Thin air cuts the lift a given rotor and motor combination can generate, and cold reduces battery capacity and flight time, so a drone that performs well at sea level cannot be assumed to perform the same way in the Himalayas. Public figures on the drone's endurance, payload and control link, and the contract terms, would let readers judge how far these drones extend the Army's reach from forward posts.

Armoured vehicles are among the targets that cheap loitering munitions and FPV drones have proved most effective against in recent conflicts. Giving tanks their own counter-drone sensors and effectors, rather than relying only on area air defence, is therefore a sound direction, and the two main in-service types between them make up most of the armoured fleet.

What has been reported so far is a direction, not yet a system. No vendor, specification or budget has been disclosed. Fitting a radar, an electro-optical sensor and a jammer onto a turret already crowded with sights, antennas and the main gun is a real integration problem, on top of the extra load on the tank's electrical system. A tank-mounted system will also need to share its picture with the Army's layered air defence networks rather than work as an isolated bolt-on. An official announcement with the supplier, cost and schedule would turn this from a reported approval into a capability we can assess.

The class itself is credible. Propeller-driven delta-wing drones with ranges near 1,000 km and warheads of tens of kilograms have been mass-produced and used in combat. The figures quoted for KAL do not yet fit together, however: endurance has been given as both three to five hours and seven to eight hours, which implies quite different cruise speeds for the same range.

The reported Pokhran trial moves this entry from a company claim to a company-reported flight test, which is a real step, but the public account of it is thin: no range, speed or altitude achieved has been published, and no independent witness, such as an armed-forces officer, is named, unlike the comparable Bhargavastra trials elsewhere on this page. Readers should treat "validated the integrated architecture" as a company characterisation of its own test rather than an external assessment.

The export pitch raises a point beyond engineering. India is a member of the Missile Technology Control Regime, and an unmanned aircraft able to deliver a payload to 1,000 km is a controlled item under India's SCOMET list. Any Gulf sale would need government authorisation, which usually follows user evaluation at home. A clear, consistent specification, a fuller account of the Pokhran trial, and a word on the trials behind "fully tested" would help observers judge the claim.

A fast, small anti-radiation drone fills a real gap: suppressing enemy air defences is costly with missiles, and a cheap jet-powered emitter hunter is what several air forces are pursuing. The company has a vehicle-launch test and a speed figure behind it, which is more than many announcements in this edition.

The step from the demonstrated 350 km/h to the claimed 700 km/h is large in physics terms. Throttle percentage on a small turbojet is not linear with thrust, and doubling the speed roughly quadruples the drag the engine must overcome. The anti-radiation seeker is the other demanding element. Accurate direction finding against low-frequency surveillance radars needs antenna aperture that is hard to fit into a small airframe, and modern air defence units shut down and move. Positioning the first deliveries as target drones is a pragmatic route to revenue and flight hours while the strike variant matures.

Speed is the attribute that matters most for an interceptor against one-way attack drones, which typically cruise at about 150 to 200 km/h. A speed margin of more than twice the target's lets an interceptor catch up from behind and recover from a missed first pass. A record-class multirotor is therefore relevant engineering, not just a stunt.

The trade-off is endurance. The power a multirotor needs rises steeply with speed, so at full speed endurance is usually measured in a few minutes, which limits the defended radius. The guidance chain is the other half of the problem: a cue from ground radar, then an onboard seeker that can hold a small target at very high closing speed. Interception trials against representative targets would turn a speed record into a counter-UAS capability.

Swarm control software is where much of the value in mass drone warfare now lies, and it is a field in which Indian start-ups can compete without heavy hardware investment. Running at the edge, with no cloud dependence, is the right architectural choice for a battlefield system.

The key test of any swarm system is graceful degradation. A swarm that relies on a central controller or a ground link is brittle under jamming. The Army's own RFP explicitly seeks decentralised autonomy, in which drones keep coordinating among themselves when cut off. Public evidence of the number of drones flown together, and of behaviour under jamming, would show how far ARMANET has progressed along that path.

The features described, decentralised control, frequency-hopping encrypted mesh links and dynamic role allocation, are the right ones for a modern swarm. They match what the Army has asked for in its own swarm RFP. The report gives no figures, location or independent observers, however, so the claims cannot yet be weighed. "Simulated jamming" in particular can mean anything from a laboratory attenuator to a field jammer, and the difference matters. We will upgrade this entry as soon as the company publishes trial details.

A vehicle-mounted counter-drone system for convoys and posts answers a real and growing need, and the reported order shows user interest regardless of which buyer is correct. The 4 km engagement envelope is realistic for RF and cyber effectors against small commercial-class drones.

This week's check of sources found a basic disagreement in public reporting that is worth flagging rather than resolving by picking one account: The Week's original report names the Army as the buyer of the ₹155 crore order, while other coverage from the same month describes it as a Ministry of Home Affairs order for border and urban security, split into two contracts that add up to about the same figure. Both are plausible buyers for a mobile counter-drone vehicle, which is exactly why an official source matters here; until one appears, readers should treat the buyer's identity as unconfirmed even though the order itself has been widely reported.

Two parts of the wider pitch would still benefit from more detail. Cyber takeover depends on recognising and exploiting a drone's control protocol, which works well against commercial drones with known protocols, much less against encrypted military links, and not at all against drones that fly autonomously or by fibre optic. The 4,000 sq km coverage claim should be read as a networked command-and-control footprint rather than one system's sensor reach: low-flying drones are masked by terrain and the earth's curvature long before 36 km, so the figure implies many fused nodes. A separate set of Ministry of Defence contracts worth more than ₹100 crore for Army and Navy sites, mentioned in this month's company profile coverage, is a reminder that this company's order book may be larger and less publicly itemised than the headline ₹155 crore figure suggests.

This is a firm, officially confirmed procurement that went through a competitive technical evaluation, and it places loitering munitions inside artillery regiments as an organic deep-fire weapon. The headline figures give an all-in average of a little under ₹1.9 crore per munition. That ratio includes launchers and accessories, so it is not a unit price.

The main questions are about integration and indigenous content, not the aircraft. A loitering munition at 100 km is only as useful as the kill chain behind it: a sensor that finds the target, a fire direction system that clears the engagement, and a datalink that keeps the operator in the loop well beyond radio line of sight. One supplier's product is reported to derive from a proven Israeli design. That reduces technical risk, but it puts weight on who controls the software and supply chain in a crisis.

The most significant part of this announcement is the engine. A small Indian turbojet that has flown in an operational-type airframe is a genuine step, because small gas turbines are a known gap in the Indian ecosystem and are often imported.

What the system can do is still open, because no performance figures have been published. The physics sets the frame: small turbojets burn fuel at a much higher rate per unit of thrust than propeller or turbofan propulsion. A jet loitering munition therefore trades endurance for speed, and the jet earns its place only if the design flies fast enough to compress the defender's reaction time. The secondary-source figures of 250 to 300 km/h sit in a speed band where a propeller is usually more efficient. Clarity from the company on the design speed and fuel fraction would show whether this is a fast strike weapon or a loiterer.

Interceptor drones have become one of the most cost-effective answers to one-way attack and reconnaissance drones, so a hand-launched, vision-guided interceptor is a relevant idea. At this stage the system is described rather than demonstrated, and coverage has drawn comparisons with fielded foreign interceptors that the company has not itself claimed. The engineering challenge in this class lies less in the airframe than in the end game. The interceptor must detect a small target, track it against clutter, and steer into it at high closing speed, with very little time to correct. Flight test results would allow a proper assessment.

Bhargavastra addresses the hardest counter-drone problem, a dense swarm, with an area-effect weapon. In principle that is the right choice: guns and single interceptors run out of time and ammunition against many simultaneous targets, while a fragmenting rocket can damage several small drones at once. The Make-II route and repeated Army attendance show genuine user engagement.

What has been shown so far is launch performance and sensor-to-shooter cueing, not interception. The 2025 firings measured launch parameters, and in the 2026 review no rockets were fired. The decisive data will come from live engagements against drone targets: fuze reliability, miss distance at range, kill probability per salvo, and the debris footprint. High-altitude operation to 5,000 m is a welcome design goal, since thin air changes rocket and radar behaviour.

A layered architecture that fuses radar, RF and EO/IR into one picture is the accepted approach to counter-UAS, and Adani has the scale to field it. Without published ranges, though, the sensors cannot be assessed on performance.

The headline feature, taking control of hostile drones, is attractive because it can bring a drone down intact and without collateral damage. Its reach is bounded by the target. It works best against drones using known commercial protocols. Against military drones with encrypted, frequency-hopping links, or autonomous and fibre-optic drones, there may be nothing to take over. Presenting it as decisive against swarms would be stronger with evidence of simultaneous takeovers, the protocols covered, and how the suite falls back to hard-kill effectors.

There is not yet enough public information to assess this platform on performance. We list it because it was shown at a symposium hosted by the Army, and because media coverage has already attached range and guidance figures that the company itself has not given. The pusher-propeller fixed-wing layout is a conventional, sensible choice for an endurance-oriented one-way attack drone. A partnership with an established space and avionics firm could address the guidance and electronics that usually decide whether such designs work. We will update this entry when the company publishes specifications or a flight test.

Zen is an established supplier to the services, so this is a credible company making a broad claim. The detection figures need context. Radar range scales with the fourth root of radar cross-section, so a radar that sees a 1 m² target at 40 km sees a 0.01 m² small quadcopter at only about 13 km, all else equal. Detecting the smallest drones at 20 km therefore implies either a large power-aperture product or a larger target class. In practice, terrain masking, clutter and bird discrimination limit low-altitude detection well before the radar horizon does.

The other point concerns the defeat mechanism. RF jamming has been the backbone of counter-drone systems, but the most effective drones in Ukraine now use fibre-optic links or onboard autonomy and give a jammer nothing to act on. Effectors are what count against them: guns, interceptors and directed energy. The share of the claimed 100 tracks that can actually be engaged, and how fast, is the number that would tell users most.

This is one of the more substantive Indian long-range strike claims, because the company has shown a flight that ended on the target. A flight of about 720 km in nearly five hours works out to roughly 145 km/h on average. That is consistent with a propeller-driven one-way attack drone of the Shahed-136 class, which carries a warhead of similar size over a similar distance. The 1,000 km figure is therefore plausible for the class, but it remains a design figure: the reported trial covered about 70 per cent of it, with an inert warhead.

The rocket-assisted take-off validation reported this week addresses a practical constraint: where the drone can be launched from. Earlier launches used vehicles and highways; a booster removes the need for even a road strip, which matters for dispersal and survivability, since a fixed launch point is an easy target. No booster specifications or launch data have been published, so it is not yet possible to say how small or rough a site the system tolerates.

The harder questions remain about the operational environment rather than the launch method. Ukraine has shown that slow, low one-way attack drones are intercepted at high rates by a dense, layered defence, so their value depends on numbers, routing and cost. Navigating precisely over 700 km or more against a GNSS-denied adversary needs high-grade inertial sensors or terrain and scene matching. Swarm coordination at these ranges needs a datalink that works well beyond line of sight. The company's answers on navigation, datalink and unit cost will decide whether this is a niche capability or a mass one.

Acoustic sensing is a sound complement to radar and RF detection. It is passive, cheap, hard to detect and indifferent to whether the drone transmits, which matters against autonomous and fibre-optic drones. A localisation error of under three degrees on a 15 kg class target is a credible result.

Physics sets firm limits. Sound weakens with distance, and wind, terrain and battlefield noise mask it, so acoustic ranges of 1 to 3 km are typical internationally. Doubling the range requires a substantial gain in signal-to-noise ratio. The roughly 6 second travel time of sound over 2 km means a drone at 150 km/h has moved about 250 m by the time it is heard. Acoustic arrays are therefore best used as a cueing layer that hands targets to radar or EO/IR, which is how the company describes its multi-sensor fusion.

Ghatak is the most ambitious Indian unmanned programme and now has formal approval, which is real user engagement. A stealthy flying wing is a demanding class. Its shaping, materials and inlets must hide the engine face and keep the signature low across the threat radar bands, and it needs an engine that is reliable without afterburner. The SWiFT demonstrator has already retired some flight-control risk for the tailless configuration.

The critical path runs through the engine and the industry partnership. The Kaveri derivative is central to the programme's indigenous character, and its certification timeline will set the pace. Choosing an industry partner able to build and later produce low-observable airframes at quality will be the next decision to watch.

Solar's loitering munition line has more operational engagement than almost any other Indian product in this category, so the Nagastra-2 figures start from a credible base. A 4 kg warhead is in the range that can defeat the thinner top armour of a tank if it is a shaped charge or explosively formed penetrator that arrives at a steep angle. The 25 km range and 90 minutes of endurance shown at Advantage Vidarbha were moderate and realistic for an electric tactical loiterer; the delivery reports this week give a 50 km fire-and-forget range and 60 minutes of endurance, which is not obviously the same specification, so the two sets of figures should be read as describing the same programme at different stages rather than necessarily the same airframe performance.

The reported delivery of 21 units is a smaller and more concrete step than the completed 450-unit Nagastra-1R consignment covered here last week: it is the first batch of the newer variant reaching the Army rather than a closeout of a large order, and no total order quantity, contract value or delivery schedule has been published for Nagastra-2. Taken together with the 1R delivery, it shows Solar moving a second variant through trials (Mahajan Field Firing Range, August 2026) into the Army's hands within the same year, which is a reasonable pace for an emergency-procurement-style programme.

The open questions are about the battlefield and the specification, not just the airframe. The low-cost tactical loitering munitions used in recent conflicts have proved highly sensitive to jamming of their video links, so the reported ECCM capability is a claim worth testing rather than taking at face value. A man-in-the-loop weapon flying low at this range may also lose line of sight to its ground station behind ridgelines. A clean, official specification sheet that reconciles the February display figures with the figures reported around delivery, and an account of how the Army is evaluating the delivered batch, would help settle what has actually changed between claim and order.

This is a mature, combat-proven platform coming into India through a manufacturing partnership, so the technical risk is low by the standards of this page. Autonomy that keeps the mission going without GNSS or a datalink is exactly the capability that recent conflicts have shown to be essential.

The questions are about sovereignty and sustainment rather than physics. An SDK lets Indian firms build mission behaviours on top of Hivemind, but the core autonomy, its updates and its certification remain with the foreign owner. The services will want to know how much of the aircraft, engine and software they can maintain, modify and integrate with Indian sensors and networks without reference to the original manufacturer, particularly in a crisis.

This entry shows the procurement process working as it should. The Army bought after electronic warfare field trials and country-of-origin checks, the two filters that have tripped up several drone suppliers. ideaForge is a listed company, so its orders are disclosed through stock exchange filings rather than marketing material, which makes the engagement easy to verify.

The detailed performance of the baseline Zolt has not been widely published, so its range and endurance cannot be assessed here. The more interesting public data point would be how the GNSS-denied autonomy performed in the Army's EW trials, since that capability now separates useful tactical drones from vulnerable ones.

The new Hybrid and UCAV variants extend the family rather than replace it. A hybrid powerplant is a sensible way to buy endurance for persistent ISR, since batteries alone trade flight time for payload, though the trade is a larger infrared and acoustic signature than a pure electric airframe. A UCAV variant is a bigger step: ideaForge has not published a payload or weapons fit, and strike integration (target designation, release authority, weapon safety) is a different engineering and certification problem from ISR. Confirmation of which Zolt variant the Army's existing contracts actually cover would also clear up a point this entry has flagged before.

General Atomics' MQ-series is the benchmark MALE family, and India already flies leased MQ-9B aircraft and has ordered more, so the platform's performance is not in question. The open issue for this programme, and for every foreign-partnered bid in it, is how much capability is genuinely transferred. Flight control software, datalinks, satellite communications and weapons integration decide whether India can operate, modify and arm the aircraft independently. Competing bids reportedly pair Indian primes with Israeli partners, so the evaluation is likely to turn on indigenous content, sovereign control of software, and lifecycle support as much as on endurance and payload figures.

Archer-NG follows a difficult path. Its predecessor Tapas was cleared for limited series production in 2023 but did not meet its original endurance and altitude goals, and did not scale into operational fleets. A clean first flight is a positive milestone, and an indigenous MALE design offers the sovereign control of software and weapons integration that foreign-partnered offers have to negotiate for.

The usual MALE challenges lie ahead: endurance in high-altitude and hot conditions, a certifiable flight control and sense-and-avoid system, satellite communications for beyond-line-of-sight operations, and an engine that delivers its rated power at altitude. Clear published targets would help readers track progress against them.

The loyal wingman concept is now the centre of several major air forces' combat aviation plans, and CATS Warrior is India's entry in that class. The airframe is the more tractable part. The harder engineering lies in manned-unmanned teaming: a jam-resistant datalink between fighter and wingman, a cockpit interface that does not overload a single-seat pilot, and autonomy trusted enough to let the wingman act on high-level commands.

The programme's timeline has already moved once, which is common for first-of-type aircraft. A clear public milestone plan from HAL would help the industry partners and suppliers who are investing alongside it.

D4 and its Army variant are the most established Indian counter-drone system: developed by DRDO, built by BEL, in service with more than one service, and reported in operational use. Its stated detection range of 5 to 8 km is a realistic reference for small drones and a useful yardstick for newer claims of 15 to 20 km. The next challenge for this and every jamming-led system is the move by attackers to drones that do not radiate, which shifts the weight towards kinetic and directed energy effectors.

We list this as a benchmark. It is the reference point against which Indian directed energy claims should be read. A 30 kW class laser that kills small UAVs at a few kilometres is consistent with fielded systems elsewhere. The low cost per shot is real, although the system's own cost, size and power supply are not small.

Lasers have known limits. They must hold the beam on one spot for seconds, so they engage targets one at a time. Atmospheric absorption and turbulence reduce their range in dust, haze and rain. Heat build-up limits how many shots they can fire in a row. They work best as one layer in a mix with guns, missiles and electronic attack, not as a stand-alone answer to swarms.