Nova Spectral AI Raises $2.25M Seed to Build Low-Cost SWIR Sensor Chips

By Grace Kim

Nova Spectral AI has raised ₹21.15 crore — roughly $2.25 million — in a seed round led by Shastra VC, with Equirus InnovateX Fund and a roster of angel investors joining. The startup plans to spend it on one thing first: a field-ready prototype within about 18 months.

The company builds short-wave infrared (SWIR) image sensor chips in India, using lead-free colloidal quantum dots and CMOS-compatible manufacturing processes. Its founders, Anish Priyadarshi and Sanjeev Kumar, are Indian Institute of Technology Kanpur alumni. The technology targets drones, defense systems, robotics, vehicles, smartphones and other physical AI applications.

How much cheaper can SWIR imaging get?

The startup's headline claim is a number: its approach could cut the cost of SWIR imaging by as much as 100 times compared with existing solutions. That matters because SWIR sensors today are expensive and depend largely on imported indium gallium arsenide (InGaAs) chips, which confines the technology to high-end defense and specialized industrial uses.

Nova Spectral AI's argument is straightforward. Cheaper sensors open bigger markets. Autonomous systems, machine vision, robotics, wearables, smartphones and automotive platforms all need better sensing to help physical AI understand and interact with its surroundings.

The physics does the heavy lifting. SWIR sensors let cameras operate where visible-light imaging breaks down — in smoke, haze and low light. Those conditions are exactly where drones and intelligence, surveillance and reconnaissance (ISR) platforms must perform, so the capability has a natural buyer base beyond consumer devices.

The takeaway: a 100x cost reduction, if it holds, would move SWIR imaging from a niche procurement line item into mass-market hardware bills of materials.

What does the money buy?

Two things, per the company: people and a prototype. Nova Spectral AI said it will use the seed funding to expand its team and develop a field-ready prototype within approximately 18 months. That timeline puts a working device — not just a lab demonstration — on the clock.

The round also lists early individual backers: Ashis Nayak, Avijeet Alagathi, Vasant Rao, Jivesh Madan, Shachi Jalote, Saurabh Gunwant, Krishna Jha, Sadhika Agarwal and Rituraj Chaudhari. Shastra VC led, with Equirus InnovateX Fund participating.

The takeaway: an 18-month prototype deadline gives investors and customers a concrete checkpoint for whether the quantum-dot approach delivers.

Why does this land in India's semiconductor push?

The financing arrives as India expands its domestic deep-tech, defense technology and semiconductor ecosystem. Nova Spectral AI is making its sensor chips in India, not importing InGaAs parts, which aligns squarely with that policy direction.

The company also frames its work around a broader shift toward physical AI — autonomous systems that increasingly depend on advanced sensors capable of operating reliably in real-world environments. Better sensing, in that framing, is the bottleneck for machines that must see through smoke, haze and darkness.

If the prototype lands on schedule, Nova Spectral AI will have an early answer to whether quantum-dot SWIR chips can genuinely undercut indium gallium arsenide at scale.