Netrasemi Advances Indigenous Edge AI With 12-nm Vision and Surveillance SoC

Kerala-origin semiconductor startup Netrasemi is advancing India’s effort to build domestically designed processors for artificial intelligence at the edge, with the company showcasing its indigenous 12-nanometre AI System-on-Chip technology at SEMICON India 2026.

The startup’s semiconductor platform combines AI and machine-learning acceleration, vision processing and video-processing capabilities within a single chip architecture. It is being developed for applications including smart surveillance cameras, industrial vision systems, robotics, drones, mobility platforms and other connected devices that need to process artificial-intelligence workloads locally rather than continuously sending data to remote cloud servers.

The Press Information Bureau highlighted Netrasemi on 22 September 2026 among the Indian semiconductor technologies showcased at SEMICON India, placing the Kerala-founded company within the growing group of domestic startups attempting to move from semiconductor design services towards proprietary Indian chip products.

Kerala Startup Builds Chips for Edge AI

Netrasemi was founded in Thiruvananthapuram, Kerala, with the objective of developing semiconductor products specifically for Edge AI and intelligent vision applications.

Edge AI moves artificial-intelligence processing closer to the device generating the data. Instead of transmitting every video frame or sensor reading to a data centre, a processor embedded inside a camera, robot, drone or industrial system can perform much of the computation locally.

This approach can reduce latency and network-bandwidth requirements while allowing equipment to continue performing AI tasks when cloud connectivity is limited. Local processing can also reduce the amount of raw sensor or video data that must leave the device.

Netrasemi is designing its processors around these requirements, combining general-purpose computing with dedicated hardware for AI inference, computer vision and multimedia processing.

Indigenous SoC Taped Out on 12-nm Process

One of the most significant developments in the company’s programme is the tape-out of an indigenously designed AI SoC using a 12-nm semiconductor process.

Tape-out represents the stage at which the completed semiconductor design is released for fabrication. It follows extensive architecture development, circuit design, verification and physical implementation and precedes the arrival of manufactured silicon for validation.

For an Indian semiconductor startup, reaching tape-out is an important transition because it moves a processor from design files and simulation towards a physical semiconductor product.

Netrasemi’s 12-nm platform integrates dedicated hardware for AI workloads alongside vision and video-processing engines. This allows computationally demanding operations to be performed within the device rather than depending exclusively on general-purpose processors.

AI Accelerator Integrated Into the Chip

The processor architecture incorporates an AI/ML accelerator designed to execute neural-network inference efficiently at the edge.

Dedicated AI acceleration is increasingly important because workloads such as object detection, image classification, facial recognition, industrial inspection and autonomous navigation involve large numbers of mathematical operations. Running these entirely on conventional CPU cores can consume considerably more processing time and energy.

Specialised acceleration hardware allows frequently used AI operations to be executed more efficiently while leaving other processor resources available for system control and application software.

This architecture makes the Netrasemi platform relevant to devices where computing performance, power consumption and physical size must be carefully balanced.

Computer Vision and Video Processing Built Into the SoC

Netrasemi is also integrating dedicated vision-processing and video engines into its semiconductor architecture.

That combination is particularly suited to intelligent cameras. A conventional surveillance camera primarily captures and transmits video, while an AI-enabled camera can analyse images locally to detect objects, classify events or identify particular patterns.

The same underlying technology can be adapted for machine vision in factories, autonomous robots, drones, intelligent transport systems and smart infrastructure.

An industrial camera, for example, could inspect manufactured components and identify defects directly on a production line. A mobile robot could process camera feeds locally for navigation and obstacle avoidance, while a drone could analyse imagery during flight without continuously transmitting high-bandwidth video to a remote processing system.

Surveillance Emerges as an Important Application

One of the primary markets identified for Netrasemi’s technology is AI-enabled video surveillance.

India deploys large numbers of cameras across transport networks, industrial facilities, commercial infrastructure and public spaces. Many of the processors used inside these systems have traditionally been sourced from overseas semiconductor suppliers.

An Indian-designed vision SoC creates the possibility of building more of the underlying computing platform domestically. This includes not only the processor itself but also the software, AI models and application ecosystem operating around it.

The opportunity therefore extends beyond replacing an imported chip. Domestic semiconductor intellectual property can allow Indian equipment manufacturers to develop products around processors whose architecture and software environment are controlled within the country.

Applications Extend to Robotics and Drones

Netrasemi’s architecture is also relevant to India’s rapidly expanding robotics and unmanned-systems ecosystem.

Autonomous machines depend heavily on sensor processing. Cameras and other sensors continuously generate information that must be analysed before the machine can decide how to move, identify objects or respond to its environment.

Processing that information locally becomes particularly important for drones and mobile robots because communication with a remote data centre may be delayed, disrupted or unavailable.

An Edge AI processor can perform object detection, image processing and other inference workloads onboard the platform, allowing decisions to be taken with significantly less dependence on external computing infrastructure.

This creates potential applications across industrial robotics, warehouse automation, inspection systems, agricultural drones, infrastructure monitoring and autonomous mobility.

Netrasemi Represents India’s Fabless Semiconductor Push

Netrasemi’s development also reflects an important component of India’s semiconductor strategy: the growth of fabless Indian chip companies.

Building semiconductor fabrication plants requires enormous capital expenditure, but chip design companies can develop valuable processor architectures and intellectual property without owning the fabrication facility themselves. Once a design is completed, manufacturing can be contracted to an appropriate semiconductor foundry.

The resulting chip can nevertheless remain an Indian-designed semiconductor product, with its architecture, intellectual property and application ecosystem developed domestically.

This model has become central to the global semiconductor industry and creates an avenue for Indian startups to enter specialised processor markets while India’s domestic fabrication infrastructure develops in parallel.

From Semiconductor Design to Indian Chip Products

India already has a substantial semiconductor-design workforce, including engineers working for many of the world’s largest chip companies. The emerging challenge is converting that engineering capability into Indian-owned semiconductor products and intellectual property.

Netrasemi belongs to a new generation of startups attempting that transition.

Rather than operating primarily as a semiconductor engineering-services company, it is developing processors that can become commercial products for equipment manufacturers. Successful tape-out, silicon validation, software development and eventual volume production would allow the company to move progressively further along that path.

The distinction is important for India’s semiconductor ecosystem because domestically owned chip architectures can create longer-term intellectual property, engineering capability and product businesses within the country.

SEMICON India Brings the Startup Into National Focus

Netrasemi’s presence at SEMICON India 2026 gave the company’s technology greater visibility within India’s expanding semiconductor ecosystem.

The government’s subsequent recognition of the startup through the Press Information Bureau highlighted the progress being made by domestic companies developing semiconductor intellectual property alongside the much larger investments taking place in fabrication, assembly, testing, equipment and materials.

For the Kerala semiconductor startup, the 12-nm tape-out represents a particularly important engineering milestone. The next stages involve manufactured silicon, validation, software integration, customer evaluation and eventual commercial deployment.

Netrasemi is therefore moving beyond the idea of India merely providing semiconductor-design talent to global companies. Its programme represents an attempt to turn that talent into an Indian-designed Edge AI processor, with applications ranging from surveillance cameras and smart sensors to robots, drones and intelligent industrial machines.

---

References

Press Information Bureau, Government of India — Semiconductor technologies and innovations showcased at SEMICON India 2026, 22 September 2026.

Netrasemi — Official Edge AI semiconductor technology and product information.

India Semiconductor Mission — Design Linked Incentive and domestic semiconductor-design ecosystem information.

Ministry of Electronics and Information Technology — India semiconductor and chip-design programme documentation.

---