Chips, Indic LLMs, flying taxis: Inside India’s tryst with tech independence

“Quantum technology is opening up new frontiers in computing, communications, cyber security with wide-spread applications,” she said during the year’s budget speech. “It is expected that lots of commercial applications would emerge from theoretical constructs which are developing in this area,” she added.

The budget allocated nearly a billion dollars to the National Quantum Mission.

Not just quantum computing, India wants to build its own semiconductor plants, develop local language models, and execute an artificial intelligence (AI) mission among others. They all form part of a grand plan to build a robust homegrown technology ecosystem, one that can shield her against geopolitically-motivated supply chain issues and other disruptions. That ecosystem, when built, will also help India build a clean energy future.

As India celebrates its 78th Independence Day, we assess the progress of some of these approaches.

5th leg of production

Over the last decade, the Indian government has expanded its digital public infrastructure (DPI) framework, earlier known as the ‘India Stack’. Today, it covers digital identities, payments, financial services, e-commerce, healthcare and transportation.

DPI is being treated as a ‘factor of production’—the others being land, labour, capital and entrepreneurship.

Most of us are familiar with some of its platforms. They include Aadhaar, the Unified Payment Interface (UPI), the Health Stack, the Open Network for Digital Commerce (ONDC), and FASTag, used at toll plazas.

Current statistics are promising: 95% of Indians have an Aadhaar number; UPI transactions reached 131 billion with a value of ₹200 trillion in 2023-24; UPI now makes up 80% of digital payments in India. ONDC likened to the ‘UPI of commerce’, recorded 21.5 million transactions in the fourth quarter of 2023-24. As of January 2024, about 300 million cards have been issued under the Centre’s flagship health insurance scheme, Ayushman Bharat Pradhan Mantri Jan Aarogya Yojana.

Going forward, the government plans to encourage the private sector to develop DPI apps in areas such as credit, agriculture, education, health, law and justice, logistics, and urban governance among others.

DPIs could significantly boost India’s economy, helping it reach a size of $8 trillion by 2030, Nasscom, an IT industry’s lobby body, and Arthur D Little International, a management consulting firm, stated in a report.

Meanwhile, the DPI framework can help other countries—in fact, it is already a global product. UPI is accepted in countries like UAE, Singapore and France. “DPI has the potential to not only reshape economies but also societies to be more equitable,” the World Bank had said. The UN has also released an interim ‘DPI Safeguards’ framework to ensure that DPI global implementation is “secure, inclusive, practical, and adaptable”.

The AI goal

Currently, the US and China lead in AI— together, they hold nearly 50% of the world’s 36,000 AI companies, according to KPMG International and ZGC Industry Institute. The UK, India, and Canada follow with 2,367, 2,080, and 1,515 AI companies, respectively. A report from Nasscom in June this year indicated a slightly higher figure for India at over 3,000 companies.

In terms of investment, the US has pumped in nearly $250 billion into 4,643 AI companies since 2013, while China has invested $95 billion in 1,337 startups. In comparison, India has contributed a mere $8 billion to 296 startups.

China-based inventors are also filing the highest number of GenAI patents, far outpacing inventors in the US, Republic of Korea, Japan and India, according to a recent World Intellectual Property Organisation (WIPO) report. China accounted for 38,000, or 70%, of the 54,000 GenAI patents filed in the decade through 2023, six times more than the US.

So, how could India catch up?

The government is investing ₹10,371.92 crore in its AI Mission, which may launch soon. This initiative includes establishing a domestic manufacturing base for graphics processing units (GPUs) through public-private partnerships and developing multi-modal domain-specific large language models to advance AI self-sufficiency.

Meanwhile, the government also plans to align the AI Mission with DPI principles to ensure fairness, interoperability, and transparency. But this approach may take some tweaking to achieve the desired results since AI is not a single technology but includes machine learning (ML), deep learning (a subset of ML), image recognition, computer vision, natural language processing (NLP), and now generative AI (GenAI).

Indic LLMs

Many of us have used ChatGPT, the chatbot developed by OpenAI. It is based on a large language model (LLM), or AI algorithms that use huge datasets to understand and generate content. Globally, LLMs are trained predominantly from English databases. A bunch of Indian companies have other ideas. They are developing regional and Indic small-and large language models and chatbots in India. Small language models use smaller datasets—they need fewer but more specific data, which is easier to get.

These companies include Krutrim by Bhavish Aggarwal, Tech Mahindra’s Indus project, AI4Bharat’s Airawat series, Sarvam AI’s OpenHathi series, CoRover.ai’s BharatGPT, and the Hanooman LLM series from SML India.

The IndiaAI Innovation Centre, part of IndiaAI mission approved by the cabinet earlier this year, seeks to encourage indigenous LLMs and domain-specific models. It is a crucial step since India is a laggard as of now—China leads with about 130 LLMs (40% of the global total) while the US holds 50%.

Nonetheless, training LLMs and multimodal language models requires quality datasets, a challenge given the lack of good quality digital data for many of India’s 22 official languages. To address this, the National Language Translation Mission’s Bhashini unit has invested $6-7 million in data collection, and Google India is funding Project Vaani (by IISc, Bangalore and ARTPARK) to gather speech data.

Rules for AI

India is a founding member and the lead chair (2023-24) of the Global Partnership on Artificial Intelligence, an international initiative to guide the responsible development and use of AI.

India, as of now, does not have a separate AI regulatory bill. But the country’s Digital Personal Data Protection (DPDP) Act, 2023, notified last year, has eased the stance on cross-border data transfer restrictions. Further, the proposed Digital India Act 2023, which aims to replace the Indian IT Act 2000, and also regulate AI, may undergo more iterations before being enacted.

Get the chip

Over the past decade, China’s Big Fund has raised hundreds of billions of dollars and acquired stakes in dozens of microelectronics companies, and is pumping in about $47.5 billion into a third investment fund. The US, which has sanctions against China’s semiconductor sector, is investing $50 billion to boost its own semiconductor manufacturing capabilities.

Geopolitics and the need for self-sufficiency have also made India think fabs, a capital-intensive affair.

India plans to build three semiconductor facilities worth almost ₹1.25 trillion. These include a fab by Tata Electronics with Taiwan’s PSMC in Gujarat; an outsourced semiconductor assembly and test (OSAT) facility in Assam, also by Tata Electronics; and an OSAT facility by CG Power in partnership with Renesas.

The goal is to reduce electronic imports and keep pace with countries such as Taiwan, China, and the US. India’s new projects, though, will currently produce chips of 28 nanometers (nm) and above. Modern chips in smartphones, computers, and servers are produced using 5-10 nanometer nodes. Building a 5 nm production line costs around $5.4 billion and takes two-three years to reach full capacity, with a break-even point potentially beyond a decade. The technology can become outdated by then.

That said, only 2% of global chip capacity uses nodes below 10 nm. Therefore, India is focusing on the 28-65 nm range for its wafer fabs, which requires lower investment and offers a longer shelf life. These investments should provide India with the building blocks for local microchip-making while bringing in the critical semiconductor value chain—design, fabrication, assembly, testing, marking and packaging.

Quantum dreams

India’s National Quantum Mission was announced on 19 April 2023. With a budget of a little over ₹6,000 crore over eight years, it aims to develop 50 to 1,000 qubits of quantum computing hardware, 2,000 kilometres of quantum communications network, and foster a domestic ecosystem of quantum research.

Unlike the bits that traditional computers use, quantum computers use quantum bits or qubits that can process the ones and zeroes simultaneously, allowing them to process a lot more information.

India plans to develop a quantum computer with about 50 qubits by 2026. It also has a quantum simulator platform built by the IISc, Bengaluru, the Indian Institute of Technology (IIT)-Roorkee, and the Pune-based Centre for Development of Advanced Computing (C-DAC). Using computing resources from C-DAC’s high-performance computers like PARAM Shavak and PARAM Siddhi, users can carry out quantum simulation. The Tata Institute of Fundamental Research has also built a 5-qubit quantum computer.

Meanwhile, Indian IT services firms, too, are exploring a range of use cases. Infosys is partnering with Amazon Web Services to develop quantum capabilities. Tech Mahindra’s Makers Lab has launched a quantum centre of excellence, QNxT, in Finland. Tata Consultancy Services (TCS) is focusing on quantum algorithms for optimization, ML, image processing, and other applications. HCL is exploring transport, logistics, finance, and security use cases, while Zensar is targeting drug discovery, genomic analysis and fraud detection.

“What sets India apart is its focus on real-world applications, specifically tailored to address the country’s unique challenges,” said Akshay Khanna, managing partner at Avasant, a consulting firm. “Instead of advancing technology purely for innovation’s sake, India is strategically leveraging AI and quantum technologies to tackle complex issues that resonate with its vast and diverse population,” he added.

Khanna cites the example of agriculture, where AI is being utilized to enhance crop management, optimize resource allocation, and streamline supply chains, “all while considering the varied climate and soil conditions across the country”.

According to a Nasscom-Avasant report, quantum technologies could add $280–310 billion to the Indian economy by 2030.

Fast and high

Ola Electric’s successful stock market listing in India indicates that there’s an investor appetite for India’s green economy story.

In March, the government lowered electric vehicle (EV) import taxes on certain models subject to carmakers committing to invest at least $500 million and start domestic manufacturing within three years. But India is not stopping there when it comes to transportation tech.

The Indian government is planning to run the country’s first bullet train from Ahmedabad to Mumbai in August 2026. India is also testing hyperloop technology—a low-pressure tube in which magnetically-levitated pod-like vehicles can ferry cargo and passengers, sometimes at speeds in excess of 1,000 kmph.

Several startups and universities are working on hyperloop. Examples include TuTr Hyperloop, incubated at IIT Madras. There is also Quintrans Hyperloop, which is based in Pune.

India, meanwhile, is also getting ready for passenger eVTOLs, or electric vertical take-off and landing aircraft. The ePlane Company, which is building an electric flying taxi, again incubated at IIT Madras, is a case in point.

Green India

Coal still accounts for over 70% of India’s electricity generation, and renewable energy sources contribute a mere 10%. India also ranks as the third-largest emitter of greenhouse gases, following China and the US.

The good news is that renewable energy capacity has crossed the 180 gigawatt (GW) mark, making India the 4th largest player globally in both wind and solar power, according to an InvestIndia.gov report.

The government, meanwhile, is targeting 500 GW of non-fossil energy by 2030 and net-zero carbon emissions by 2070. Favourable policies and the establishment of Ultra Mega Renewable Energy Parks (land developed with common infrastructure facilities like transmission infrastructure, road, water, drainage, communication network etc.), will only help this cause.

A production linked incentive (PLI) scheme aimed at bolstering local manufacturing, and the launch of the National Green Hydrogen Mission will further boost this effort. Green hydrogen, which uses renewable energy sources like wind or solar power to split water into hydrogen and oxygen (a process called electrolysis), generates no carbon emissions during production. It will therefore result in a clean and emission-free fuel.

“Green hydrogen has enormous potential to decarbonize several sectors, reduce carbon emissions and achieve energy independence. The production of green hydrogen using renewable energy sources like solar, wind, and hydropower is sustainable and environmentally friendly, making it an attractive option for the transition to a low-carbon future,” the ministry of new and renewable energy states on its website.