Returns On AI Investments Will Begin To Emerge Steadily As Adoption Deepens: NASSCOM AI Chief Ankit Bose

Chennai: AI is redesigning roles across sectors and the real story is workforce transition, not collapse. There may be short-term friction during transitions, but in the long run, India has a significant opportunity, said Ankit Bose, head, NASSCOM AI in an interview with Financial Chronicle. For India, the opportunity lies not in replicating frontier models built elsewhere, but in building inclusive, scalable, and human-centric AI systems that address national priorities while contributing to global progress.

Q: AI is evolving at an extraordinary pace. How would you describe the progress AI technologies have made globally?

AI is moving faster than most people anticipated. Over the last two to three years, advances in foundation models, infrastructure and applications have been phenomenal. We are now seeing trillion-parameter models alongside highly capable smaller language models.

Countries across the world are investing heavily in compute infrastructure, specialized chips, and data centres. Enterprises have moved beyond experimentation — they are adopting AI, generative AI, and agentic AI at scale. Many are transitioning from pilot projects to production environments and actively measuring productivity gains and ROI.

Take software development as an example — code generation is increasingly AI-assisted, and decision cycles are becoming significantly shorter. Alongside this rapid adoption, there is growing emphasis on building a strong trust layer. When AI is deployed at scale, ensuring trust, safety and reliability becomes critical.

Q: Trillions of dollars are being invested globally in AI and frontier technologies. Is there a risk of a dot-com-like boom-and-bust cycle?

AI infrastructure is highly capital-intensive. Investments in GPUs and data centres are substantial upfront expenditures. However, the returns are staggered over several years.

This is not a conventional sector where incremental investment leads to immediate incremental profits. Instead, we are witnessing heavy initial investment followed by distributed returns. For this reason, I do not foresee a repeat of the early-2000s dot-com bust. Returns will begin to emerge steadily as adoption deepens.

Q: Where does India stand in AI innovation and adoption compared to major global players? Are we more consumers than creators?

The US and China are primarily focused on frontier models and massive compute capabilities. Europe is leading in regulation, trust, and governance frameworks. Israel is particularly strong in deep-tech startup innovation.

India’s differentiation lies in its use-case-driven approach, frugality, inclusivity at scale, and relevance for the Global South.

We are not building trillion-parameter frontier models at this stage. Instead, we are developing smaller, focused models trained on Indic data and cultural contexts. Multilingual AI for Indic languages is a major priority.

Our focus areas include population-scale solutions in agriculture, education, and healthcare. The approach is human-centric and designed to address national priorities. Simultaneously, we are building massive capacity in talent and deployment capabilities, because AI implementation requires skilled professionals across sectors.

Q: What are the major challenges India faces in scaling AI?

There are three major challenges:

India’s linguistic diversity makes high-quality, annotated multilingual datasets essential. Much of this data does not yet exist at the required scale, so we are building it from scratch through government, academia, and ecosystem initiatives.

India currently has around 30,000 GPUs. By comparison, some global companies individually operate over 300,000 GPUs. Expanding compute capacity is therefore a critical priority.

AI readiness must extend beyond STEM professionals. Non-STEM talent must also be equipped to use AI productively in their roles. At the same time, responsible AI development and deployment must remain central as startups and SMEs scale AI solutions.

Q: There are concerns about job displacement, especially in IT services, BPO, and KPO sectors. Do you see a serious threat?

I would frame this as job transformation rather than job displacement.

AI is redesigning roles across sectors by automating repetitive tasks and augmenting workflows. The real story is workforce transition, not collapse. India has a highly scalable and reskillable workforce. In IT services, BPOs, and KPOs, AI is augmenting process flows, risk controls, customer journeys, and domain functions. India is already a key global hub for AI deployment, and that role will only grow.

There may be short-term friction during transitions, but in the long run, India has a significant opportunity.

Q: How prepared are non-technology sectors for AI adoption?

In 2024, NASSCOM launched an initiative focused on non-tech companies. We reached around 500 CXOs across 11–12 cities to build awareness and readiness.

There is growing awareness and foundational capacity across sectors like healthcare, manufacturing, and agriculture. However, moving from awareness to deep adoption requires identifying the right use cases aligned with sector-specific needs. That journey is ongoing.

Q: What opportunities lie ahead for India in AI, particularly in talent?

A 2024 study showed India had demand for approximately 600,000 AI/ML engineers, against a supply of around 450,000. Since then, supply has grown significantly — by an estimated 50–70% — due to concerted efforts by enterprises, government, and academia.

However, continuous upskilling is essential. In software development alone, India employs nearly six million people across the development lifecycle. Tools and processes are evolving every few months. Keeping this workforce future-ready is a massive opportunity.

On the academic front, engineering curricula must become more applied and industry-linked. While India has produced over 50,000 AI-related research papers in recent years, stronger industry collaboration and commercialization of research will be crucial.

Beyond tech, AI is transforming marketing, media, manufacturing, and quality assurance workflows. Across industries, core processes are being redesigned.

Q: What should be the government’s role in this transformation?

The government must act as an ecosystem enabler. The IndiaAI Mission, with a ₹10,000 crore allocation, has already catalysed foundation model initiatives and expanded compute capacity.

There have also been strong efforts in skilling through AICTE and other sector bodies. However, given the pace of change, policies must remain adaptive.

The ultimate goal is not to compete with the US or China, but to leverage AI to improve citizens’ lives — expanding access to education, healthcare, and public services.

Q: What progress has been made through recent AI summits?

The first AI summit in Bletchley focused on AI risks and harm. The following year in France emphasized action. Now, the conversation has matured toward impact.

India’s hosting of the AI summit showcased the maturity of its ecosystem — startups, academia, research, and scalable deployment models. The summit declaration, signed by over 80 countries and international organizations, underscored a shared commitment to human-centric AI. There is clear intent for collaboration, and follow-up engagements are expected.

Q: With major investments in data centres and compute infrastructure, what role will the AI ecosystem play going forward?

Data centres are foundational to AI compute capacity. Beyond NVIDIA GPUs, there are discussions around diversified chip ecosystems, including AMD and other architectures.

These resources will enable India and Global South-specific model development, AI deployment across priority sectors, fundamental research and long-term strategic goals, including movement toward advanced AI capabilities. India must build self-sustaining AI capacity over time.

Q: Data centres are energy-intensive. Could this pose a sustainability challenge?

Data centres are indeed power hungry. The key question is the energy source. India is expanding its renewable energy footprint, including solar and other sustainable sources. At the same time, improvements in data centre efficiency — including cooling technologies and better Power Usage Effectiveness (PUE) — are critical. As we scale AI infrastructure, sustainability must remain central to planning.