India Wants Its AI Talent Back; But What’s The Incentive?

For decades, India has exported some of its brightest minds to the world’s leading research labs. Researchers of Indian origin have helped shape breakthroughs at OpenAI, Google DeepMind, Anthropic and Meta, contributing to technologies that are now redefining economies and industries. Yet very few of these breakthroughs are happening in India.

The government’s Prime Minister Research Chair (PMRC) scheme seeks to change just that. The initiative aims to attract accomplished Indian-origin researchers, scientists and technologists from global institutions to work with premier universities and national laboratories in India. The scheme covers thirteen strategic sectors, including AI, semiconductors and quantum computing, and offers research grants, infrastructure support and institutional backing.

While India recognises that getting its talent back will solidify its position on the AI world map, the launch of PMRC also raises a larger question: why would the world’s best AI researchers choose India over Silicon Valley? Let’s try to understand in this edition of The AI Shift.

Can India Give Researchers A Reason To Return?

According to the cofounder of Bharat1, Umakant Soni, researchers today want to work on the toughest problems in the world. 

History testifies that the greatest scientific talent has gravitated towards the biggest missions. The Manhattan Project reshaped physics. The Apollo programme transformed aerospace research. CERN became a magnet for scientists because it pursued questions no other institution could answer. And the list goes on.

If AI research is to flourish in India, the same kind of breakthrough is needed. Today, much of India’s AI ecosystem is concentrated around applications, enterprise deployments and model fine-tuning. Valuable businesses are being built, but very few efforts are attempting to redefine the frontiers of AI.

So, we must ask: can India build entirely new forms of AI systems? Can we create world-class models that understand physical environments? Can it solve problems unique to a country of 1.4 Bn people and, in doing so, produce technologies relevant to the world?

Without such moonshot ambitions, India risks remaining a consumer of breakthroughs developed elsewhere.

The PMRC scheme creates pathways for researchers to come back, but pathways alone aren’t sufficient. Experts believe researchers also need a compelling scientific reason to return.

A senior scientist at a leading AI lab in Silicon Valley is unlikely to move countries simply because funding is available. The greater attraction is the chance to work on a problem that could fundamentally advance the field.

Frontier Research Needs More Than Talent

India’s challenge today is not producing AI talent but creating an environment where it can thrive and conduct world-class research.

Bhaskarjit Sarmah, the head of AI research for financial services at Domyn, an enterprise AI venture, breaks this challenge further. He said the first hurdle towards this endeavour is compute. 

Training and experimenting with frontier AI systems increasingly depend on access to massive computing resources, high-end GPUs and supercomputing infrastructure. Without these resources, researchers often find themselves limited to incremental work instead of pursuing ambitious scientific questions. This infrastructure gap is one reason why many assume that America does all the research while India executes.

The second challenge is collaboration. 

Leading research ecosystems across the United States and Europe are deeply interconnected. Universities work closely with industry through joint laboratories, sponsored projects and shared infrastructure. Researchers routinely move between academia and industry while continuing to pursue frontier work. Sarmah believes that India’s industry-academia interface remains far from that level of maturity. A stronger bridge between startups, enterprises and research institutions could significantly improve access to infrastructure and create larger pools of ambitious research problems.

Then there is the issue of incentives. According to Levels.fyi, an independent portal tracking tech compensation, an AI researcher in the US earns a median total compensation of $178,920 annually. The same role in India pays roughly $30,782 a year. That is nearly a six-fold difference. 

“We should pay professors on par with leading AI labs,” said Paras Chopra, the founder of AI lab Lossfunk.

Research is rarely driven by money alone. But compensation influences where talent chooses to build long-term careers. A country attempting to attract the world’s best researchers cannot ignore the economics of talent.

Why India’s AI Ambitions Need A Mission-Mode Approach

India may never match OpenAI, Google or xAI dollar for dollar. But do we really need to? 

The emergence of companies such as DeepSeek and Mistral has shown that relatively small, highly capable teams can produce globally significant breakthroughs.

The real question is whether India can assemble and empower such teams. The country has already demonstrated its ability to execute large-scale national technology missions. Aadhaar and UPI succeeded because they had a clear objective, long-term commitment and coordination across institutions.

AI needs a similar playbook. A mission that combines compute infrastructure, talent recruitment, academia-industry collaboration and a handful of ambitious scientific goals could significantly improve India’s chances of becoming a serious destination for frontier AI research.

The PMRC scheme is therefore an important beginning. But attracting AI researchers back to India is not merely about creating fellowships. It is about creating an ecosystem in which the world’s best minds believe the next big breakthrough can happen.

Top Stories From India & Around The World

• Pramaana Labs Raises $27 Mn: The startup has secured $27 Mn in a seed round led by Khosla Ventures to build a verification layer for AI systems in regulated sectors such as tax, law and healthcare. The startup plans to advance its formal reasoning and AI prover models.
• Gnani.ai Unveils Prisma v2.5: Gnani.ai has launched Prisma v2.5, a speech-to-text model spanning 12 languages and trained on 14 Mn hours of Indic speech data. The startup claims the model delivers higher accuracy and lower latency for multilingual, real-world Indian voice environments.
• IFC Backs Sify’s AI Data Centre Push: International Finance Corporation has committed $371 Mn to Sify Technologies’ data centre arm to build two AI-ready facilities in Navi Mumbai and Chennai. The centres will add 103 MW of capacity and support India’s growing demand for cloud and AI infrastructure.
• AI Upgrade For Cybercrime Helpline: Home Minister Amit Shah has directed officials to revamp the National Cybercrime Helpline 1930 with AI and multilingual support to improve cyber fraud reporting, accelerate investigations, and streamline grievances related to bank accounts frozen during fraud probes.

The Weekly Buzz: Coinbase Bets On AI Financial Advisors

A conversation around AI-driven wealth management gained traction last week after crypto exchange Coinbase unveiled an SEC-registered AI advisor to offer personalised investment recommendations.

The announcement stood out because of the regulatory groundwork involved. According to Coinbase’s chief legal officer Paul Grewal, the company created and registered an entity with the US Securities and Exchange Commission that is authorised to provide advisory services. The AI Advisor is non-discretionary, meaning it can suggest portfolio actions, but users must approve every trade themselves.

The product begins by asking users about their financial goals and risk appetite before analysing their existing portfolios. Its recommendations are generated using large language models guided by inputs from financial professionals, combined with the user’s portfolio data and risk profile.

The launch sparked broader discussions about the future of financial advice and whether AI could make personalised investing guidance more affordable and accessible. It also highlighted the growing role of AI in highly regulated industries, where systems are increasingly moving beyond information retrieval and into decision-support functions that can directly influence user outcomes.

The larger takeaway is that AI is steadily entering domains traditionally dominated by human experts. If products like Coinbase’s AI advisor gain traction, they could reshape wealth management by lowering the cost of financial advice while forcing regulators and financial institutions to rethink how AI-powered advisory services should be governed.

Startup In The Spotlight: Paygent

As enterprises increasingly deploy AI agents across customer support, sales and business workflows, many are discovering a new challenge: understanding the true cost of running these systems and whether they are profitable. Bengaluru-based Paygent is betting that financial visibility and pricing intelligence will become a critical layer of infrastructure for the agentic AI economy.

Founded in 2025 by Prakash Kushwaha and Aditya Sonkar, Paygent is building a financial operations platform for companies developing AI agents. The startup emerged from the founders’ observation that AI-native businesses often struggle to predict infrastructure costs, track margins and price their products effectively because AI agent costs vary depending on usage and model consumption.

The startup’s platform enables AI companies to track costs, revenues and profitability at a customer and agent level. By integrating with cloud providers and LLM services, Paygent gives businesses visibility into how much it costs to serve individual customers, the margins they generate and the effectiveness of different pricing strategies.

Paygent currently focuses on three areas: financial visibility for AI agents, experimentation with pricing models and an intelligence layer that predicts future costs and recommends model optimisation strategies. The company is also building capabilities to help businesses route certain workloads to more efficient AI models, reducing inference costs without compromising performance.

The startup is initially targeting voice AI companies, where usage patterns can fluctuate significantly and make pricing particularly complex. It currently has three paying customers and operates on a subscription model based on the number of events or interactions processed through its platform.

Operating in the rapidly expanding agentic AI ecosystem, Paygent sees itself as an infrastructure layer for AI-native companies. The startup plans to eventually build a full-stack platform that combines billing infrastructure, analytics and subscription management capabilities for businesses deploying AI agents, while also pursuing fresh funding to accelerate product development and expand into the US market.

Prompt Of The Week

What prompts and hacks are CTOs, CEOs and cofounders using these days to streamline their work? 

Here’s the prompt used by Ganesh Sahai, CTO at Nagarro, to rapidly turn customer ideas into demo-ready prototypes and significantly reduce the time between concept and execution.

“Here’s a concept we discussed with a customer: 

[describe the idea, the user, and the problem it solves].

Turn this into a working prototype plan I can demo in 48 hours.

• Define the single core user journey to build.

• Recommend the simplest tech stack to fake everything else.

• Create realistic sample data to make it feel alive.

• Write a demo script that walks the customer through the ‘aha’ moment in under five minutes.

Then generate the starter code for the core journey.“

Editor’s Note: Some prompts may need to be adjusted by users for best results or may not work as intended for certain users.

[Edited by Shishir Prasher]
[Creatives by Varshita Srivastava]

The post India Wants Its AI Talent Back; But What’s The Incentive? appeared first on Inc42 Media.