Work on developing the framework for the chip to power artificial intelligence systems is under way, and an official announcement is expected later this year, said three senior officials aware of the development. The government has set a tentative timeline of 2027 to make the chips at a domestic chip fabrication plant, they said on the condition of anonymity.
The ministry of electronics and information technology (Meity) has partnered with the Centre for Development of Advanced Computing (C-Dac) and the National e-Governance Division to design the chip from the ground-up, the officials said.
The chip will leverage C-Dac’s expertise in building processors based on the open-source ‘Risc-V’ (reduced instruction source computer) core architecture, they said. It is a set of openly accessible standards, which can be used by any party to build its own processor’s core architecture.
India’s quest to develop its own proprietary AI chip, crucial for running generative AI models that made tools such as ChatGPT possible, reflects concerns about lack of access to technology that will shape the future. According to a report by Gartner, over 70% of the world’s semiconductor revenue was earned by the US last year, giving it disproportionate power to control semiconductors that will power everything from appliances and cars to spacecraft.
Emails seeking responses from Meity and C-Dac did not receive responses until press time.
Also read | AI Mission should focus on core tech to avoid errors made in chips, smartphones
Still, developing this technology from the ground up typically requires years of research and development (R&D), and billions of dollars in investments to create the level of performance and efficiency of processors made by Nvidia, Advanced Micro Devices (AMD) and Intel.
“We are decades behind the US firms on R&D—and semiconductors need billions of dollars of large-scale R&D to build a proprietary chip,” said Jayanth Kolla, partner at technology consultancy firm Convergence Catalyst. “Also, India’s only announced chip fab right now can only make up to 28nm (nanometre) chips, which isn’t sophisticated enough for AI operations. This leaves a clear gap to be addressed.”
Nvidia’s AI chips are 5nm or smaller.
Even China took decades to develop its own technology. After US imposed sanctions against China during Donald Trump’s first tenure, Huawei, one of China’s largest electronics conglomerates, shifted most of its electronics devices to its custom HiSilicon chips—built on the company’s proprietary Kirin cores. However, HiSilicon was incorporated in 1991—over three decades ago.
Help from Big Tech
“The C-Dac Risc-V chip core will be the fundamental processing framework, while experts including from private Big Tech firms in the US and executives from the Taiwan Semiconductor Manufacturing Company (TSMC) foundry are consulting with Meity to help develop the same,” said one of the officials cited above.
The Centre’s push to develop these chips is to provide ample computing power for researchers in academia, as well as startups, to develop foundational artificial intelligence (AI) models.
Also read | Between chips order, Stargate and DeepSeek, Indian AI firms take a beating
“There is already a substantial amount of talent in chip design within India, and more engineers and researchers are being trained at TSMC, in partnership with some of India’s most premier educational institutes, to bring chip fabrication expertise to the country,” the second official said. “There is a three- to five-year roadmap in discussion within Meity, in order to bring the said chip to mass-scale production.”
Not a knee-jerk reaction
On 14 January, former US president Joe Biden signed an executive order curbing chip supplies to select nations. India was not included in the list of nations exempted from the order’s restrictions, but it wasn’t blacklisted either, unlike China. However, the move was seen as a threat to India’s growing technology sector. On 30 January, Mint reported that India’s AI-facing technology firms saw their share prices decline by nearly 50%, since the order was announced.
The same day, Union IT minister Ashwini Vaishnaw announced that the India AI Mission had confirmed the availability of 18,693 graphics processing units (GPUs)—processors that are used to power the training and development of foundational AI models.
The move is not a knee-jerk reaction to the US action, but is a long-term move that has been in development since before the unveiling of the India AI Mission’s contours, the officials said.
Also read | From chips to Stargate: Can India weather the storm of Trump’s AI policy changes?
“A semiconductor chip is close to becoming a fundamental commodity today. This means that no nation can continue to remain reliant on another nation for its chip supply—or will end up playing into the hands of established superpowers,” said the third official quoted earlier.
“India’s own chip infrastructure will look to soften the impact of this by creating a power-efficient computing chip that’s at par with the kind of processors required for cutting-edge AI development,” the official said.
Opportunity for India
According to Kolla of Convergence Catalyst, the US model showcases the government’s role both in funding private institutes to boost R&D, and to become a customer of an advanced technology early in its cycle. “This fosters a habit of risk-taking, which the US has leveraged to the fullest to become a leader in global technology today. The same has long been missing in India.”
Ashok Chandak, president of industry body India Electronics and Semiconductor Association (Iesa), also stressed on prompt action.
“Deep investments in R&D should start flowing into educational institutes and government-affiliated bodies in order to develop core technologies in semiconductor, AI and the electronics component ecosystem,” he said. “Failing to do so will mean that an executive order signed by the US can severely impact multiple industries in India, and how investors deem our market.”
Also read | Targeting global chip shifts more important than self-sufficiency: Chris Miller
Kolla cited the example of how Europe has faded away as a place for cutting-edge technologies even for consumer technologies. He said the US and China are headed in the same direction as the US-Soviet Union space war from the 1960s, with AI.
“For India, which is on the geopolitical edge, this opens up a considerable opportunity to become a key stakeholder in global technologies,” he said. “But a broad-based change of mindset in all parties involved is required, before we can speak about selling our fundamental technologies to the rest of the world.”
Bid for better fabs
A senior industry executive, who works closely with Meity and the India Semiconductor Mission, said the government is in talks with “key strategic partners” to bring “more sophisticated fabs” to India.
“The next phase of the India Semiconductor Mission will need to bring higher-value fabs to India, which is a tricky task since the global value chain is currently ruled by Taiwan—followed by the likes of Malaysia and Vietnam,” the executive said on the condition of anonymity. “But India plays a crucial role geopolitically thanks to its proximity with the US, its importance in the Asian subcontinent, and itself as the fastest growing large economy—which could play to our strength.”
And read | Govt says India’s semiconductor market to cross $100 billion by 2030
Meity is expected to announce the second phase of ISM in the coming weeks, Vaishnaw said on Saturday. Mint reported in August last year that an additional $15-20 billion in semiconductor fab incentives was on track to be sanctioned for it, subject to cabinet approval.
