Design, Simulation, Operation: The Three Layers of Software Dependency That Leave Indian Industry Vulnerable to Foreign Chokepoints

At the Aeronautical Development Agency's design centre in Bangalore, engineers are working on India's most ambitious military aviation project: the Advanced Medium Combat Aircraft, a fifth-generation stealth fighter meant to place India alongside the United States, Russia, and China as one of only four nations capable of building such machines. The full-scale engineering model unveiled at Aero India 2025 drew crowds and national pride in equal measure, with first flight planned for 2028 and induction expected by 2035.

Look closer at those workstations, however, and a different picture emerges. The screens run CATIA, the computer-aided design software made by Dassault Systèmes of France, while the simulation tools testing the aircraft's aerodynamics, structural integrity, and radar signature come from Ansys of Pennsylvania and Siemens of Germany. When the aircraft eventually enters production, the factory floor will be orchestrated by industrial control systems from Honeywell, Rockwell, or Siemens—completing a dependency chain that runs from conception through validation to manufacturing.

Every layer of India's industrial capability depends on foreign software, which means that a single export control decision, a sanctions package, or a diplomatic rupture could freeze those screens and render India's most celebrated "indigenous" fighter undesignable, untestable, and unbuildable.

The easy part of sovereignty

On Monday (26 Jan), France announced that it would replace Microsoft Teams and Zoom across all government departments by 2027, switching to a domestically developed video-conferencing platform called Visio. "The aim is to end the use of non-European solutions and guarantee the security and confidentiality of public electronic communications," declared David Amiel, France's minister for civil service reform, noting that the platform—built by the Direction Interministérielle du Numérique—would be hosted on Outscale, a sovereign cloud provider owned by Dassault Systèmes and certified under France's stringent SecNumCloud 3.2 standard.

The symbolism is potent and the savings are real, amounting to roughly €1 million annually for every 100,000 users, but France's move also reveals the limits of sovereignty-by-substitution. Video conferencing is mature, commoditised technology for which open-source alternatives have existed for years, which makes replacing Zoom the easy part of any digital independence agenda.

What France is notably not replacing is CATIA—the same software that designs Rafale jets, Airbus components, and, as it happens, India's AMCA. Dassault Systèmes has grown into a €70 billion company precisely because France recognised decades ago that industrial design software was strategic infrastructure, not a procurement category. The same is true of Germany with Siemens, whose simulation and factory automation tools undergird European manufacturing from Stuttgart to Stockholm.

India's situation is the precise inverse: it has sovereignty in almost nothing, and the dependency runs far deeper than most observers realise, extending through three distinct layers of industrial capability, each controlled by a handful of foreign vendors and each representing a potential chokepoint that could be closed at any moment.

Three layers of vulnerability

India's industrial software dependency operates at three levels that together span the entire lifecycle of modern manufacturing. The first is the design layer, encompassing the CAD and CAE tools that bring new products into existence—aircraft, engines, automobiles, ships—without which nothing can be conceived in the first place. The second is the simulation layer, comprising the CFD, FEA, and process modelling tools that test designs virtually before physical prototypes are built, without which nothing can be validated against real-world physics. The third is the operations layer, consisting of the SCADA, DCS, and MES systems that run factories, refineries, and power plants, without which nothing can be produced safely or at scale.

Each of these layers is dominated by a small number of foreign vendors, predominantly American, French, and German, and each represents a chokepoint that could be closed by export controls, licensing disputes, or geopolitical rupture. Perhaps most striking is the near-total absence of these concerns from India's otherwise vocal debate about technological self-reliance, which has focused overwhelmingly on hardware—on fighter jets and aircraft carriers—while ignoring the invisible software substrate that makes such hardware possible.

The design layer: where products are born

A tour through India's defence-industrial complex—HAL's helicopter division in Bangalore, DRDO's missile laboratories in Hyderabad, Mazagon Dock's warship design centre in Mumbai—reveals the same software stack at every stop. CATIA and Siemens NX handle computer-aided design, SOLIDWORKS serves mid-range applications, and Autodesk tools dominate architecture and construction, not as peripheral aids but as the fundamental methodology of modern engineering. To design complex systems without these platforms is to attempt twenty-first-century engineering with twentieth-century tools—theoretically possible but practically uncompetitive.

The scale of dependence is substantial: India ranks as the second-largest market globally for both CATIA and Siemens NX, trailing only the United States in adoption. When HAL's Aircraft Research and Design Centre works on the AMCA, the Tejas Mk-2, or the Indian Multi-Role Helicopter, the geometry takes shape in Dassault software, and when Mazagon Dock designs the Navy's next-generation destroyers—with nineteen warships scheduled for commissioning in 2026 alone, the largest single-year fleet expansion in Indian naval history—the hull forms emerge from the same foreign platforms.

The automotive sector, though less strategically sensitive, follows identical patterns: Tata Motors, Mahindra, and Maruti Suzuki all run their design operations on CATIA, with Tata Technologies serving as a certified Dassault and Siemens partner across India. The vehicles roll off assembly lines in Pune and Chennai, but the designs that define them originate in software developed in Vélizy-Villacoublay and Plano.

Product lifecycle management—the enterprise systems that track designs from initial conception through manufacturing and into long-term maintenance—exhibits the same dependency, with Siemens Teamcenter dominating Indian aerospace and defence while PTC Windchill and Dassault ENOVIA serve supplementary roles. No indigenous PLM alternative of comparable capability exists, leaving the entire knowledge base of Indian industrial design resident on foreign platforms.

India does have one modest homegrown success in this domain: ActCAD, a 2D/3D CAD platform developed by Jytra Technology Solutions in Andhra Pradesh and marketed as an AutoCAD alternative at roughly eighty percent cost savings. The company claims 30,000 users across 103 countries, which represents genuine commercial achievement, but ActCAD remains a niche product for basic drafting that does not begin to compete with the parametric, simulation-integrated, enterprise-scale platforms required to design fighter jets and automobiles. The gap between ActCAD and CATIA is not one of degree but of kind—the difference between a competent tool and an industrial ecosystem.

The simulation layer: where physics meets code

Designing an aircraft or an automobile is only the beginning of the engineering process, because before any metal is cut, the design must be tested—virtually—against every conceivable stress and operating condition. Will the wing survive a 9G combat turn? Will the radar cross-section evade detection at operationally relevant angles? Will the engine intake maintain stable airflow at Mach 2? Will the heat shield protect astronauts during the violent deceleration of atmospheric re-entry? These questions are answered not primarily in wind tunnels, which remain expensive and slow, but in simulation software that models the governing physics computationally—and here India's dependency approaches totality.

The simulation layer fragments into multiple technical disciplines, each dominated by foreign specialists with decades of accumulated expertise. The larger of these disciplines—and by a considerable margin—is structural analysis, where finite element methods predict how components respond to mechanical forces, thermal loads, and vibration. Here, four vendors corner eighty to ninety percent of the Indian market: MSC Nastran (owned by Sweden's Hexagon), Ansys Mechanical, Dassault's Abaqus, and Altair OptiStruct. Indian aerospace and automotive engineers rely on these tools to validate everything from aircraft landing gear to automotive crash structures, and the domestic alternatives are essentially non-existent.

Computational fluid dynamics, though a far smaller market—perhaps one-twentieth the size of structural analysis—proves equally concentrated and arguably more strategically sensitive. The field itself fragments by application, with different tools dominating different speed regimes and industrial sectors. For low-speed aerodynamics and aerothermal simulation, the kind that matters for automotive design and cooling systems, Dassault's PowerFLOW, Ansys Fluent, and Siemens Star-CCM+ together command roughly eighty percent of Indian usage. High-speed aerodynamics—the transonic and supersonic flows relevant to missiles, fighters, and re-entry vehicles—runs predominantly on Ansys Fluent across most defence laboratories, with CFD++ appearing in some facilities and an IISc-developed code finding limited use in select research environments.

More specialised applications show similar patterns of foreign dominance. Gas turbine research, critical for both aircraft engines and power generation, relies on Siemens' GE Flo Simulator, Ansys Fluent, and Cadence Turbo. Combustion simulation, which models the violent chemistry inside engines and industrial burners, runs on Siemens Star-CCM+ for more than half of Indian applications, with Ansys Fluent capturing another twenty to thirty percent. In each niche, the same handful of American and European vendors appear, their tools so deeply embedded in engineering workflows that switching would require not just software replacement but wholesale retraining of technical staff.

Electromagnetic simulation, which proves critical for radar design, antenna optimisation, and stealth engineering, depends on Ansys HFSS and Dassault's CST software, with Entuple Technologies serving as Ansys's designated partner for Indian aerospace and defence and working directly with both ISRO and DRDO. When Indian engineers design the radomes that protect sensitive radar equipment or optimise antenna placement for minimal interference, they do so entirely within American software ecosystems.

Process simulation, which models the chemical engineering of refineries and petrochemical plants, runs on AspenTech's Aspen HYSYS, Honeywell's UniSim (which has served the oil and gas industry for over three decades), and AVEVA's various platforms. Every refinery expansion at Reliance Jamnagar, every petrochemical project at Indian Oil's Paradip facility, every fertiliser plant upgrade across the country—all are designed first in foreign simulation tools before a single foundation is poured.

There is, however, one significant Indian exception that proves instructive about what sustained investment can achieve. In May 2024, ISRO announced PraVaHa, an indigenous CFD solver developed over many years at the Vikram Sarabhai Space Centre specifically for aerodynamic simulations of launch vehicles and re-entry capsules. The software was used extensively in the Gaganyaan human spaceflight programme, and ISRO has stated that it will "soon replace most CFD simulations currently carried out using commercial software," with plans to make the tool available to DRDO and academic institutions.

PraVaHa matters because it demonstrates that indigenous simulation capability can indeed be built when strategic priority aligns with patient, long-term investment. But it also illustrates the daunting scale of the broader problem: ISRO spent years developing a single solver for a single discipline—fluid dynamics—tailored to its own specific applications in aerospace. Structural analysis, thermal simulation, electromagnetic modelling, and the multi-physics coupling that integrates all these domains would each require equivalent sustained effort, meaning that the simulation layer presents not one challenge but dozens, each demanding specialised expertise, extensive validation, and years of iterative development.

The operations layer: where industry runs

Design and simulation determine what can be built, but a third layer of software dependency governs what can actually be operated—the industrial control systems that run India's power grid, oil refineries, steel plants, chemical facilities, and pharmaceutical factories. These systems, which monitor thousands of sensors and coordinate complex processes in real time, are overwhelmingly foreign in origin.

SCADA systems—Supervisory Control and Data Acquisition—provide the nervous system for infrastructure at national scale, and Power Grid Corporation of India runs its operations on GE Energy's SCADA/EMS platforms, which cover both Regional and State Load Dispatch Centres. Legacy systems from Siemens and ABB, some installed as far back as the 1980s, continue to operate in portions of the transmission network, creating a patchwork of foreign dependencies accumulated over decades.

Distributed Control Systems, which manage continuous processes in refineries and chemical plants where precise coordination is essential for both efficiency and safety, show similar patterns: BPCL's Mumbai Refinery operates on Yokogawa CENTUM controllers with Yaskawa PLCs, Indian Oil's Gujarat Refinery runs on Honeywell SCADA, and ONGC's offshore platforms rely on systems integrated by Larsen & Toubro but built around foreign DCS and Emergency Shutdown components that Indian integrators assemble rather than design.

The programmable logic controllers that serve as the fundamental building blocks of industrial automation—controlling everything from steel rolling mills to pharmaceutical packaging lines—come predominantly from Siemens (whose SIMATIC line dominates globally), Rockwell Automation (whose Allen-Bradley brand is ubiquitous in American-influenced facilities), and Mitsubishi (whose MELSEC controllers serve Japanese-affiliated operations).

Manufacturing Execution Systems, which coordinate the complex choreography of modern factory floors, are likewise dominated by foreign vendors: Siemens Opcenter, Rockwell FactoryTalk, and SAP Digital Manufacturing command the high end of the market, and while Indian MES providers such as Autosys, AIP, and DMeX have emerged to serve smaller-scale deployments, mission-critical manufacturing remains firmly in foreign hands.

The cybersecurity implications of this dependency are both documented and disturbing. After the Stuxnet attack in 2010—when sophisticated malware destroyed Iranian uranium centrifuges by targeting Siemens industrial controllers—India was discovered to be the third-most-affected country globally, with infected SCADA systems identified in power plants and oil pipelines. CERT-In continues to report hundreds of SCADA-targeted attacks annually, and during the military operations of May 2025, multiple advanced persistent threat groups launched approximately 1.5 million cyberattacks over just three days, with malware specifically designed to target operational technology networks at GAIL and ONGC pipeline facilities, Narmada Water Authority infrastructure, and major steel plants.

Indigenous alternatives have begun to emerge in limited domains: C-DAC Bengaluru has developed SCADA systems for certain specialised applications, and Bharat Electronics Limited delivered an AI-integrated "Super SCADA" system for Delhi Metro that represents genuine capability. A 2021 Ministry of Power committee formally acknowledged the need to reduce dependence on foreign OEMs under the Atmanirbhar Bharat framework, but the same committee noted candidly that India currently lacks indigenous capacity to manufacture the servers, workstations, routers, and switches that such systems require—making hardware sovereignty at least as challenging as software sovereignty, and likely more so.

The chokepoint beneath all the others

Beneath these three visible layers of dependency lies a fourth that is less apparent but more fundamental still: the software used to design semiconductors themselves. Every chip in every industrial controller, every processor in every simulation workstation, every integrated circuit in every radar and missile guidance system—all are designed using Electronic Design Automation tools made by just three American companies that together constitute an effective oligopoly over the foundational infrastructure of the digital age.

Synopsys and Cadence together control roughly sixty-five to seventy percent of the global EDA market, with Siemens EDA (formerly Mentor Graphics) holding another thirteen percent, and no meaningful alternatives exist at the scale and capability required for advanced chip design. When engineers at Intel's Bangalore design centre, AMD's Hyderabad facility, or any of India's emerging fabless semiconductor startups sit down to design an integrated circuit, they open Synopsys Design Compiler for logic synthesis, Cadence Innovus for place-and-route, and Siemens Calibre for physical verification—American tools, American licences, American export control jurisdiction.

India houses approximately twenty percent of the world's semiconductor design engineers, and its Global Capability Centres produce some 3,000 chip designs annually, making it an essential node in the global semiconductor value chain. The irony embedded in this position is structural and profound: India designs chips for the world but cannot design chips without the world's permission, because the tools that enable this work remain entirely foreign.

The vulnerability inherent in this arrangement surfaced briefly but dramatically in May 2025, when the United States threatened to extend EDA export restrictions to China and every chip design house on the planet was forced to recalculate its risk exposure within a matter of days. The restrictions were ultimately lifted within weeks—reportedly as part of a diplomatic exchange involving Chinese rare earth shipments—but the episode demonstrated with uncomfortable clarity how rapidly the fundamental infrastructure of design can be weaponised for geopolitical purposes. India currently sits in Tier 2 of the US export control framework, meaning continued access to EDA tools is likely but explicitly conditional on American policy preferences. "Likely but conditional" may be many things, but it is not sovereignty.

Indigenous processor development has made progress—IIT Madras's SHAKTI project has produced working RISC-V chips, and C-DAC's VEGA and DHRUV64 processors serve industrial automation applications—but every one of these ostensibly indigenous processors was designed using Synopsys, Cadence, or Siemens EDA tools. India has demonstrated the capability to build processors; it has not begun to build the software required to design them.

What India has actually built

The picture that emerges from this survey is sobering, but it is not uniformly bleak, because India has in fact achieved genuine technological sovereignty in several domains—and the pattern of these successes offers important lessons about what works.

In telecommunications equipment, sustained government commitment combined with industrial policy has produced remarkable results: Tejas Networks, now a Tata Group subsidiary, has completed the world's largest single-vendor indigenous 4G/5G deployment—more than 100,000 cell sites for BSNL—making India one of only five nations globally with fully homegrown mobile network technology. HFCL deployed India's first indigenous consumer routers under the BharatNet programme in January 2026, and the Production Linked Incentive scheme has catalysed the establishment of 61 manufacturing units that together produce telecom equipment worth over ₹57,000 crore while exporting to 70 countries.

In geospatial technology, India has built genuine world-class capability through a combination of public investment and private entrepreneurship: ISRO's Bhuvan platform now serves 150,000 daily users with high-resolution satellite imagery, while MapmyIndia has captured approximately ninety percent of the domestic GPS navigation market and provides location intelligence services to Amazon, Uber, and Apple. The 2021 Geospatial Policy deregulation removed longstanding bureaucratic obstacles and has since enabled 375 GIS startups to emerge.

The migration of 1.2 million central government email accounts to Zoho—including sensitive users at the Prime Minister's Office—represents the largest indigenous enterprise software deployment of its kind anywhere in the world. And India's digital public infrastructure achievements—UPI for payments, Aadhaar for identity, DigiLocker for document management—demonstrate world-leading innovation that other countries now seek to emulate.

Defence hardware indigenisation has likewise accelerated substantially, with DRDO's recently delivered Electronic Pilot Simulator for AMCA and Tejas Mk-2 development claiming eighty-five percent indigenous content. The Tejas light combat aircraft, the INS Vikrant aircraft carrier, and the BrahMos cruise missile system all demonstrate that Indian industry can manufacture sophisticated military hardware to competitive standards.

The pattern that emerges from these successes is consistent and instructive: India achieves technological sovereignty where strategic priority aligns with sustained investment over decades, whether in space technology, telecommunications, digital payments, or defence hardware. It fails—or rather, has not yet seriously attempted—to build capability where the challenge is invisible to policymakers and the public, as with the industrial software that designs rather than the products that result from design. It is time to change this.

(The author wishes to thank Krishnan CV for his inputs)