India can build any technology, says Prof. Ashok Jhunjhunwala; calls for backing startup risk-takers

Jhunjhunwala argues that India already has the talent to develop any technology, but lacks enough support for turning ideas into physical products. Technological self-reliance means backing young founders, changing engineering education, and directing innovation toward practical Indian problems. The goal is to move from importing products to creating them domestically and eventually leading globally.

He proposes removing half the existing curriculum to create space for hands-on building. He also recommends an AI strategy based on edge computing, where smaller models run on phones, laptops, and other local hardware. Consortia of scientists and startups would work on difficult civic challenges.

The current paradox is stark: Indian engineers create advanced software while India imports almost every physical product. Projects through ITEL illustrate the intended path, including transit systems moving people 15 kilometres in 20 minutes and automated waste separation. Supporting risk-taking founders could turn talent into self-reliance.

Jhunjhunwala sees a difference between established domestic companies and young startup founders. Established industry is traditionally risk-averse, which can discourage uncertain experiments and new physical products. Startups, by contrast, are willing to attempt difficult ideas despite the possibility of failure. That willingness matters because technological self-reliance requires more than technical skill; it requires sustained risk-taking.

The article connects this attitude to India’s local challenges. Young founders are being encouraged to build low-watt graphics processing units, automated waste-separation systems, and other practical technologies. They can work with scientists through consortia, test ideas, learn from failure, and adapt products to real conditions.

This does not mean every startup will succeed. It means founders may be more prepared to take the first steps that established companies avoid. Jhunjhunwala believes backing these entrepreneurs can help India stop importing solutions and develop global leadership by solving everyday problems at home.

Edge computing means processing data directly on local hardware, such as smartphones, laptops, or low-power workplace devices. Instead of depending mainly on enormous cloud models controlled by foreign conglomerates, AI can operate closer to where data is created. Jhunjhunwala sees this as a better strategic fit for India’s needs and resources.

Indian enterprises could retrain open-weight models with domestic datasets and deploy compact models on edge devices. Local processing can reduce dependence on distant cloud infrastructure and support AI inside manufacturing and business workflows. Incubators are also guiding founders to design custom, low-watt graphics processing units for fast computation on small devices.

The article contrasts this approach with chasing multi-trillion-parameter cloud models. India’s proposed focus is practical: how much AI can run on a mobile phone or laptop? If successful, this approach could spread useful AI across workplaces while improving efficiency without requiring the largest possible models.

The article says Indian enterprises could deploy compact AI models containing 5-billion to 80-billion parameters at the edge. These models would be smaller than the multi-trillion-parameter cloud systems India is urged not to chase. Their value would come from practical deployment, using domestic datasets and local devices to improve manufacturing and enterprise workflows.

The transport target is equally concrete. ITEL is using robotics and AI to design clean, high-efficiency urban transit systems. The stated ambition is to move commuters 15 kilometres in 20 minutes across congested cities such as Bengaluru. This links AI research directly to a measurable civic outcome.

These figures describe the scale of the proposed solutions, not results already achieved. The article presents them as development goals supported by incubators, scientists, and startups. Their broader importance is strategic: smaller local AI and applied robotics could address Indian conditions rather than copy foreign systems.

If local AI systems work as intended, Indian manufacturing and enterprises could become more efficient. Compact models running on nearby devices would support workflows without requiring India to depend entirely on giant foreign cloud systems. This could make advanced computing more practical for local operations and strengthen domestic technological capability.

Robotics and AI are also being aimed at two visible civic problems. ITEL is designing urban transit capable of moving commuters 15 kilometres in 20 minutes across congested cities like Bengaluru. At the same time, young founders are developing automated waste-separation technologies that identify and extract economic value from municipal garbage.

The article presents these as projects and ambitions, not completed nationwide outcomes. Their success could help India address traffic and waste while building products for its own conditions. Solving such everyday problems would also give startups experience, create locally useful technologies, and support the shift from importing foreign solutions to becoming a global technology leader.

Jhunjhunwala identifies an education problem behind India’s product gap. Engineering colleges often prioritise examination preparation and overloaded theoretical curricula instead of teaching students how to build. He argues that removing half the curriculum could create time for practical creation, peer learning, and continuous learning. This would help students turn technical knowledge into usable products.

Consortia add the missing bridge between ideas and difficult applications. ITEL brings academic scientists and startup founders together to work on civic challenges. Their projects include robotics for high-efficiency transit and automated systems for separating municipal waste. Scientists contribute research capabilities, while founders pursue iterative product development and accept failure as part of the process.

India already has talented young people, including in Tier II, III, and IV cities. The challenge is giving them opportunity, practical training, and institutional support. Combining hands-on education with consortia could help transform software expertise into physical products and locally relevant technologies.

A country can produce excellent software and advanced designs without developing the factories, products, and organisations needed to manufacture them. Jhunjhunwala describes this as India’s central paradox: engineers can develop sophisticated technologies, yet the country imports almost every physical product. Technical talent exists, but the pathway from knowledge to product is weak.

The article points to engineering education as one cause. Colleges remain overloaded with theory and examination preparation, while students receive too little practice building things. Established domestic industry is also traditionally risk-averse, making it less likely to fund uncertain product development. These factors separate technical capability from commercial manufacturing.

Jhunjhunwala’s solution is to teach continuous learning and practical creation, then support young founders who will take risks. Consortia can connect them with scientists and local problems. Building products for traffic, waste, and industry could gradually convert India’s existing technical strength into physical self-reliance.