Kerala’s deep-tech ambitions have found a new talking point. Netrasemi, founded by Jyothis Indirabhai, Sreejith Varma, and Deepa Geetha, a technology startup backed by Zoho, has launched its first artificial intelligence chip, the A2000, putting another Indian name into the difficult and highly specialised race to design AI silicon.
The chip has already been tested and declared production-ready by the company. For India’s semiconductor and AI ecosystem, the launch is significant because the A2000 is being positioned among the early AI and machine learning chips designed in India.
A chip built for devices that need intelligence on the edge
The A2000 is not aimed at large data centres or cloud-heavy AI workloads. Its focus is more grounded and immediate: compact, power-efficient edge devices.
That means the chip is designed for hardware such as smart cameras, edge AI boxes and intelligent video gateways, where AI processing has to happen closer to the device instead of being sent back and forth to the cloud.
Built on TSMC’s 12nm process, the A2000 brings together several processing units developed in-house by Netrasemi. These include engines for neural processing, computer vision, image handling and security.
This makes the chip relevant for use cases where devices need to process images, video and sensor inputs quickly, while keeping power consumption under control.
Trials already underway with customers
Netrasemi is currently running early trials of the A2000 with three customers in the surveillance and automotive sectors.
The company plans to begin mass production at TSMC’s Taiwan facility in 2027. A commercial launch is being targeted for mid-2027.
“We are currently working with several OEMs on early sample evaluations and co-development,” Netrasemi CEO Jyothis Indirabhai said.
The choice of sectors also shows where the company sees its early market. Surveillance systems and automotive applications both need faster on-device intelligence, especially in video-heavy environments where latency and power efficiency matter.
Netrasemi: Backed by Zoho, funded for the long semiconductor road
Chip design is a capital-intensive business, and Netrasemi has built its journey with institutional support.
The company has raised ₹125 crore in total funding from investors including Zoho and Unicorn India Ventures. It was also one of four startups selected for ₹15 crore support under the government’s Design Linked Incentive scheme in 2023.
The funding and policy support are important because semiconductor development usually involves long design, testing and fabrication cycles.
Unlike software products, chips take time to move from design to tape-out, sample testing, customer validation and commercial production.
For Netrasemi, the A2000 is not the only product in the pipeline.
More chips in development
The startup has also developed the R1000 AI/ML MCU, a second chip aimed at the IoT sensor market.
That chip was developed in collaboration with the College of Engineering Trivandrum under MeitY’s Chip-to-Startup programme.
Netrasemi is also working on the A4000, an advanced edge AI server chip supported by MeitY. The company expects the A4000 to be fabrication-ready by the second quarter of 2027.
Taken together, the A2000, R1000 and A4000 suggest that Netrasemi is trying to build a wider edge-AI chip portfolio rather than depending on a single product.
India has been pushing to strengthen its semiconductor design and manufacturing ecosystem, but chip design remains one of the toughest parts of the technology value chain. The A2000 launch matters because it shows an Indian startup moving beyond the idea stage and into customer trials with a production-ready AI chip.
The real test will come next: customer validation, manufacturing scale-up and commercial adoption.
For now, Netrasemi has placed itself in a space where India wants to be taken more seriously: AI hardware, designed locally, built for real-world devices and backed by both private capital and public semiconductor programmes.
Also Read: Vaibhav Suryavanshi’s Rise Is Bigger Than a Net-Worth Number
