IIT Delhi Develops India's Ist Indigenously Designed Micro‑GPU

IIT Delhi has achieved a major milestone by developing India’s first indigenously designed micro‑GPU, a compact programmable graphics processor aimed at affordable embedded systems such as e‑rickshaw dashboards, industrial displays, and educational devices. This breakthrough reduces India’s dependence on imported GPUs and opens pathways for indigenous chip design.

Notably, the key difference between a GPU and a micro-GPU is that a conventional GPU is a high‑performance graphics processor designed for intensive parallel workloads like gaming, AI, and scientific computing, while a micro‑GPU is a compact, low‑power version optimized for basic graphics tasks in embedded systems.

GPUs dominate in AI, gaming, and scientific computing, but they are costly and power‑hungry. While, micro‑GPUs fill the gap for low‑cost, socially useful devices, enabling India’s push for indigenous chip design and reducing import dependence. IIT Delhi’s prototype micro‑GPU shows how embedded systems can benefit from programmable graphics without needing full‑scale GPUs

Key Highlights

• Developed by: IIT Delhi’s Electrical Engineering Department, led by Prof. Jayadeva, Prof. Kaushik Saha, and M.Tech students Nammi Akash and M. Ravi Teja.
• Platform: Demonstrated on a Spartan‑7 FPGA using a custom floating‑point GPU engine written in Register Transfer Language (RTL).
• Significance: First working, demonstrable indigenous micro‑GPU from an Indian university.
• Applications: Industrial control displays, Human‑machine interfaces, E‑rickshaw dashboards, Fishing boat navigation terminals, Educational e‑book readers.

Technical Details

Feature	Details
Architecture	Scalable, programmable graphics processor IP
Implementation	RTL mapped to Spartan‑7 FPGA
Future Roadmap	8–16 core vector‑style GPU with optimized compiler & toolchain
ASIC Migration	Planned proof‑of‑concept at 65nm process node for economic viability
Goal	Affordable, socially useful computing platforms bridging the digital divide

Why It Matters for India

• Reduces import dependence: Currently, all GPUs used in India are imported.
• Supports indigenous hardware ecosystem: Aligns with India’s semiconductor mission and DLI program.
• Educational impact: Provides a platform for students and researchers to experiment with programmable graphics hardware.
• Economic viability: Mature nodes like 65nm can make indigenous GPUs cost‑effective for mass deployment.

Possible Challenges & Trade‑offs

• Performance limitations: Micro‑GPUs are not designed for high‑end AI/ML workloads; they target basic graphics needs.
• Scalability risks: Moving from FPGA prototypes to ASIC chips requires significant funding and industry collaboration.
• Global competition: India must compete with established GPU giants like NVIDIA and AMD, who dominate advanced GPU markets.

India’s indigenous GPU ecosystem is accelerating — beyond IIT Delhi’s micro‑GPU, the government and academia are driving multiple initiatives under the IndiaAI Mission and Semicon India Programme to expand domestic GPU design, manufacturing, and compute capacity.

Government announced addition of 20,000 GPUs to India’s compute infrastructure in February 2026.Builds on existing 38,000 GPUs made available at subsidized rates (₹65/hour) for startups, academia, and public institutions.The government aims to democratize access to high‑end compute for AI research and innovation.