Akashteer Is a Control System. Lattice Is an Operating System.
On 7 May 2025, when Pakistani drones and missiles came in waves after India's strikes on terror camps under Operation Sindoor, Akashteer was the system that held. The Bharat Electronics Limited (BEL) built air-defence command-and-reporting network, deployed at six nodes on the western front, tracked the inbound threats and assigned the closest weapon to engage each one. The Ministry of Defence later said it had a near 100 per cent success rate against Pakistani drones and missiles, and BEL said the system performed beyond users' expectations. Ten months later, in late February 2026, the United States and Israel opened Operation Epic Fury against Iran. Tehran fired more than 3,000 drones and over 500 ballistic missiles at coalition forces and Gulf states in the first weeks of the war, and the software running much of the American counter-drone response was a system called Lattice, built by the defence-tech company Anduril. Both systems did their jobs. But they are not the same kind of software, and treating them as comparable misses the point of what each was designed to do.
Akashteer is a control system. Lattice is an operating system. The distinction sounds semantic. It is not. It governs what each platform can absorb, how fast it can adapt, and what kind of fight it is built for. For an Indian Army that has just declared 2026 its Year of Networking and Data Centricity, that distinction is the most important software question of the year, and it is not yet being asked publicly.
What Akashteer was built to do, and did
Akashteer was contracted in March 2023 for ₹1,982 crore as the Automated Air Defence Control and Reporting System for the Indian Army. It is a joint product of DRDO, ISRO and BEL, with 455 systems planned and roughly 275 delivered by September 2025. Phased deployment is meant to complete by 2027.
The system's job is bounded and well defined. It pulls feeds from Indian Army and Indian Air Force radars, builds a single unified air picture, distinguishes friend from foe, and assigns the closest available ground-based air-defence weapon to engage each threat. It is integrated with the IAF's Integrated Air Command and Control System (IACCS) and the Navy's Trigun air-defence network, sitting inside a wider C4ISR framework. The picture reaches all the way down to the smallest Army Air Defence unit on the ground.
That is exactly what an automated air-defence C2 system is supposed to do. In Sindoor, against a real Pakistani drone and missile campaign, it worked. The architecture, however, is hierarchical and mission-specific by design. Radars in, weapons out, with a unified picture in the middle. The system is excellent at the problem it was scoped to solve. The architectural ceiling is the next problem.
What Lattice is doing differently
Anduril does not pitch Lattice as a control system. The contract documents themselves describe it as a proprietary, open-architecture, AI-enabled suite that integrates computer vision, machine learning and mesh networking to fuse real-time data from disparate sources into a single autonomous operating picture. Three architectural choices separate it from an Akashteer-class platform.
1. Open versus closed
Lattice is built on open APIs. New sensors and new effectors, including third-party hardware and legacy systems, can be plugged in without rewriting the platform. In Operation Epic Fury, Anduril's Fly-Away Kit bundled the Wisp 360-degree AI infrared sensor, the Heimdal mobile thermal-and-radar trailer, the Pulsar AI electronic-warfare platform and the Anvil ramming interceptor drone, all stitched together by Lattice as the integration layer. Akashteer integrates with a defined and BEL-anchored ecosystem. That works for the mission it was scoped against. It is harder to extend in a hurry when a new sensor type or a new interceptor needs to be brought in mid-conflict.
2. AI fusion versus operator-mediated picture
Akashteer's contribution to Sindoor was speed. It collapsed the time between radar detection and weapon engagement, automating what was previously a manually coordinated loop. That is real and valuable. Lattice operates a layer above that. Its command-and-control layer uses AI to fuse multi-spectral sensor data, classify targets and recommend or autonomously execute an effector pairing across cameras, drones, radars, RF sensors and acoustic detectors at once. Public descriptions of Akashteer reference real-time threat analysis and adaptive targeting, but the AI fusion layer is not characterised in the same terms.
3. Mesh versus hierarchy
Akashteer reports up and commands down, like a conventional military C2 hierarchy. Lattice runs on what Anduril calls its Lattice Mesh decentralised networking architecture, in which every node sees what every other node sees and the network holds together even when communications infrastructure is degraded. In a contested electromagnetic environment, where Iranian jamming reportedly forced coalition aircraft to adapt mid-mission, that resilience is not cosmetic. It is the difference between a network that bends and one that breaks.
The procurement architecture that matches the software architecture
The software is one half of the story. The procurement vehicle that carries it is the other. In March 2026, the US Army awarded Anduril an enterprise contract worth up to $20 billion over ten years, with completion estimated by March 2036. The deal consolidates more than 120 separate Anduril procurement actions into a single ordering vehicle, and the first task order, $87 million for Lattice as the C2 backbone for Joint Interagency Task Force 401's counter-UAS work, was placed within days. New sensors, new effectors and new software modules can now be added through task orders without re-tendering the underlying capability.
Indian defence procurement does not work this way. Each capability fights through its own AON, RFI, RFP, trials and L1 cycle, often running into multi-year timelines. By the time the next Akashteer upgrade clears the system, the threat envelope has usually shifted. An open, AI-fused, meshed software platform is only as fast as the procurement vehicle that buys it. The two architectures, software and contractual, have to move at the same speed, or the software's core advantage is neutralised before it reaches the field.
The Sindoor problem was not the next problem
None of this is an argument that Akashteer is a failed system. It clearly is not. The Indian Army is already extending it onto tracked vehicles called Carrier Air Defence Tracked (CADET) platforms, with the Ministry of Defence having issued an RFP for 83 such carriers in late April 2026. The aim is to keep Akashteer mobile with armoured formations across plains, deserts and high-altitude sectors up to 5,000 metres. That is sensible incremental work on a system that has already proved itself.
The argument is that Sindoor's drone-and-missile campaign was the lower bound of what India will face, not the upper bound. Pakistan, like Iran, can scale cheap drone saturation. Epic Fury demonstrated that thousand-drone waves are now a real operational scenario, that interceptor magazines drain faster than attackers run out of drones, and that the only software stack which holds against this threat is one that is modular, vendor-agnostic, AI-fused and mesh-networked. The threats and the defences will keep changing faster than any single procurement cycle can accommodate. An Akashteer-class architecture, however well executed, is built around a fixed mission and a fixed weapons inventory. A Lattice-class architecture is built to keep absorbing new ones.
The takeaway
Akashteer is what good Indian defence software looks like when the mission is well defined: indigenous, combat-proven, integrated with existing service systems, delivered in numbers. Lattice is what an open, AI-native, mesh-networked C2 platform looks like at scale, paired with a procurement vehicle that moves at the speed of the threat. The two represent different design philosophies, not different points on the same maturity curve.
The honest question for DRDO, BEL and the Indian Army is not whether to copy Lattice. It is whether the next generation of Akashteer, whatever it ends up being called, is being designed as a control system or as an operating system. The answer will decide how India fights its next thousand-drone day.