正文 · AI 翻译
IT Home reported on November 6 that India is accelerating the development of its domestic semiconductor industry, with plans to develop its own GPU chip in about 24 months while bringing the RISC-V-based SHAKTI and VEGA processors to the commercial market.
According to a November 4 local-time report by Indian technology outlet ETGovernment, S. Krishnan, secretary of India's Ministry of Electronics and Information Technology (MeitY), said in an interview that the government plans to support domestic GPU chip development under India Semiconductor Mission 2.0.
He said the work could not be completed overnight and was expected to take about 24 months, as India works toward the goal of having its own GPU. However, the GPU project is still in the early planning stage. Krishnan did not identify the companies, chip suppliers or manufacturers that would be responsible for development, or disclose technical details such as the chip architecture, performance specifications or manufacturing process.
Krishnan also outlined five foundational layers needed to develop India's AI industry: energy supply, accelerator chips, computing infrastructure and servers, AI models, and applications.
According to the report, Indian companies VVDN, Kaynes and Avalon already have server manufacturing capabilities. India's Centre for Development of Advanced Computing (C-DAC) has developed the Rudra series of servers, including products for AI workloads, and is helping local companies expand production through technology transfers.
In addition to the GPU, the Indian government plans to commercialize the domestic SHAKTI and VEGA processors. Both use the open RISC-V instruction set architecture and are CPU designs, separate from the GPU chip now being planned. Using RISC-V means domestic developers can design processors without relying on licenses for proprietary CPU instruction set architectures.
Both processor initiatives have been underway for years. The Indian Institute of Technology Madras (IIT Madras) released the SHAKTI software development kit in 2019, allowing developers to begin building related applications before the processor was commercialized. The Aries 3.0 development board, launched in 2024, uses C-DAC's VEGA ET301 processor.
The VEGA ET301 is a 32-bit chip with a maximum operating frequency of 100 MHz and 26 KB of SRAM. It is intended mainly for relatively simple applications such as sensors, wearable devices and the Internet of Things, rather than desktop PCs or intensive graphics computing tasks.
In terms of practical applications, the Indian government showcased several projects using SHAKTI and VEGA processors at the SEMICOM India event on September 26, including a SHAKTI-based autonomous water-surface cleaning robot, a digital telescope equipped with a VEGA processor, and a VEGA project for monitoring forest environments.
India's Ministry of Electronics and Information Technology also supports domestic chip design companies through the Design Linked Incentive (DLI) scheme. The ministry said 24 companies had received support under the scheme and subsequently secured venture capital or other private investment. Their products include microcontrollers, drone flight controllers, smart meter chips and closed-circuit television cameras.
In semiconductor manufacturing, India is also bringing in international companies to help build its domestic industry. According to information previously released by the Indian Prime Minister's Office, Tata Electronics has partnered with Dutch semiconductor equipment maker ASML to support India's first front-end semiconductor wafer fab in Dholera, Gujarat. Intel has also announced a strategic partnership with the Tata Group covering semiconductor manufacturing and packaging for the Indian market.
On AI's impact on the job market, Krishnan compared the current opportunity to the Y2K issue around 2000. He believes AI could create a similar growth opportunity for India's IT services industry, although the specific types of business would differ. Indian companies could help global firms integrate AI technology into existing business systems, providing services such as data security, systems integration and enterprise deployment.
A research report previously published by India's National Institution for Transforming India (NITI Aayog) estimated that the country's IT and IT-enabled services sector currently employs about 6 million people. As AI technology changes the industry's structure, about 2 million existing jobs could be eliminated, while the industry could also add about 4 million jobs. Krishnan believes enterprise AI deployment will become a new area of business growth for India's IT services industry.
Through India Semiconductor Mission 2.0, the Indian government hopes to gradually build a complete industrial ecosystem spanning chip design, wafer fabrication, advanced packaging, semiconductor materials, manufacturing equipment and research and development.
According to data previously released by the India Electronics and Semiconductor Association (IESA), India's semiconductor market is currently worth more than $50 billion and is expected to approach $100 billion by 2030 (IT Home note: approximately 668.252 billion yuan at current exchange rates).