Netrasemi, a Kerala-based startup backed by Zoho and Unicorn India Ventures, has launched its first...
The AMW Read
First production-ready edge AI chip from an Indian startup with government backing; updates the AI infrastructure segment player map and silicon substrate landscape.
Netrasemi, a Kerala-based startup backed by Zoho and Unicorn India Ventures, has launched its first artificial intelligence chip, the A2000, a production-ready edge AI processor built on TSMC's 12nm process. The chip integrates in-house neural processing, computer vision, image handling, and security engines, targeting smart cameras, edge AI boxes, and intelligent video gateways. The company is running early trials with three customers in surveillance and automotive sectors, plans mass production at TSMC's Taiwan facility in 2027, and targets commercial launch in mid-2027. Netrasemi has raised ₹125 crore (~$15M) total funding and received ₹15 crore (~$1.8M) under India's Design Linked Incentive scheme. Additional chips in development include the R1000 AI/ML MCU for IoT and the A4000 edge AI server chip, expected fabrication-ready by Q2 2027.
Why it matters: Netrasemi represents a genuine entrant into the high-difficulty AI silicon design space from India, a country with minimal presence in the global chip design arena despite ambitious semiconductor policy push. The A2000 is not a data-center competitor to NVIDIA or AMD but a pragmatic edge-AI play targeting latency-sensitive, power-constrained use cases. This mirrors the pattern of several Asian sovereign-AI chip efforts that seek to build domestic capability without directly challenging hyperscaler GPU dominance. The company's selection under India's DLI scheme and backing by Zoho signal state-private alignment to incubate indigenous chip design talent. The real test — customer validation, yield ramp, and commercial scale — lies ahead, but Netrasemi has moved beyond the PPT stage to silicon-in-hand trials.
Expert take: Netrasemi's trajectory updates the Indian semiconductor startup landscape. Unlike software-first AI companies that can iterate monthly, chip startups face 18-36 month design-to-revenue cycles, making the ₹125 crore raise relatively lean for a tape-out and customer sampling program. The company's deliberate focus on edge inference over training avoids directly competing with established players in the high-margin data-center segment. Success will depend on whether the A2000's on-device vision and video processing delivers competitive latency/power ratios against offerings from Ambarella, Hailo, and Synaptics in the edge AI camera market. The A4000 server chip, if fabricated on schedule, will test whether Netrasemi can move up the performance stack without external foundry constraints.