THIRUVANANTHAPURAM: Netrasemi, a Thiruvananthapuram edge AI chip startup, now lists five system-on-chip products on its official product page, led by the A2000, which it describes as a mid-range AI SoC. The company's own news page says the A2000 has achieved successful silicon bring-up and is being readied for commercial production, with the target set for 2027.

Netrasemi designs chips that run AI models on the device itself, rather than sending data to a cloud server. Its pitch is aimed at surveillance cameras, robotics, drones, smart city systems and factory equipment, where decisions have to be made in real time and often without a reliable internet link.

The Netra chip lineup

The product page names the NETRA R1000, a real-time edge microcontroller marked as taped out; the C1000, a connected vision SoC; the C2000, an entry AI accelerator; the A2000 mid-range AI SoC; and the A4000, an advanced AI SoC. Detailed specifications for each are not posted publicly, and the company asks interested partners to get in touch.

Netrasemi's news page groups several updates on the A2000 from May and June 2026. Those summaries describe it as India's first indigenous AI system-on-chip, production-ready, built on a 12 nm process, and say early customer trials have begun with clients in the surveillance and automotive sectors. These are the company's own descriptions of its milestones.

On its technology page, Netrasemi says it focuses on application performance per watt rather than headline AI numbers. A real surveillance camera, it argues, has to run computer vision, image signal processing, video compression and neural processing together, so the chip has to balance all of them. The company says it has a patented graph fabric hardware acceleration design that streams data between accelerator blocks with minimal CPU involvement, and gives the example of a drone surveillance system running more than ten AI models in parallel on a cost-effective SoC.

Chiplets, founders and government support

Netrasemi also says it has a UCIe-based die-to-die interconnect, which would let it combine several chips in one package. It says this could extend its design to neural chips of up to 100 TOPS and to hybrid application processors. The company adds that it is looking for partners on that work.

The about page names Jyothis Indirabhai as CEO and co-founder, Deepa Geetha as COO and co-founder, and Sreejith Varma as CTO and co-founder. "We not only make System-on-Chips, but own the IPs which goes inside," the CEO is quoted as saying. The company's address is near the College of Engineering Trivandrum at Sreekariyam.

Netrasemi is among the chip design firms supported under the government's Design Linked Incentive scheme, and its news page records the Union electronics minister's meeting with DLI companies. A Press Information Bureau release dated 27 January 2026 said the scheme then supported 24 startups, which had completed 16 tape-outs, with six chips fabricated at nodes as advanced as 12 nm.

Kerala does not often appear on India's semiconductor map, and that is part of why Netrasemi gets attention. The state's tech push also includes the IndiaAI centre of excellence in Thiruvananthapuram. For the wider chip policy picture, see our report on the domestic chip mandate for smart meters, and for on-device AI software, our story on ShepHertz's offline AI models.

The hard part comes next. Moving from silicon bring-up to volume production means locking foundry capacity, packaging and testing, and winning design slots in cameras and vehicles that may already use imported chips. Pricing and power numbers for the A2000 will decide how many of those slots it gets.

Edge AI chips are worth explaining for general readers. Instead of streaming video from a camera to a distant server for analysis, an edge chip runs the AI model inside the camera or the drone itself. That saves bandwidth, cuts delay and keeps sensitive footage on the device, which is why defence, traffic and factory systems are interested. The trade-off is that the chip has to deliver enough computing power within a tight power and cost budget, and that is exactly the problem Netrasemi says its architecture is designed to solve.

Netrasemi lists info@netrasemi.com for business enquiries and hr@netrasemi.com for careers.