India Must Move Beyond The ‘AI Application Edge’ Narrative And Build Foundational Strength

Should India focus on building its own foundational artificial intelligence (AI) models or should it prioritise creating practical AI applications? Abhishek Singh, then chief executive of the IndiaAI Mission, asked Microsoft chief Satya Nadella early last year.

At the time, the US and China were sprinting ahead in the AI race while India appeared to be caught between two competing instincts—build the base or build on top of it. As time passed, India accepted that it must invest in foundational models for strategic reasons even though its real edge lies in the application layer.

The tech ecosystem has pegged the application layer as the saviour of the country’s AI ambitions—the one layer of the stack where India has no structural disadvantage.

Infosys co-founder Nandan Nilekani has repeatedly argued that the country’s bigger opportunity lies in building AI solutions that solve real-world problems at scale.

“The Western focus on AI is on infrastructure, larger models, more chips, more data centres. But infrastructure eventually gets commoditised. The value shifts to the application layer, and India can become the AI use-case capital of the world,” he told a forum this year.

Similarly, Mahindra Group chairman Anand Mahindra made a case in a recent post on X, comparing AI to the smartphone. He stated that a remarkable technology becomes indispensable only through the apps and experiences layered on top of it.

It is the ecosystem, he argued, that determines who captures lasting value—not the chip or the model underneath. “That’s where I truly believe AI service companies have the edge. Not necessarily in owning the model, but in owning what sits above it,” Mahindra said.

Some, like Satish Viswanathan, author of a book on enterprise AI, point out that it makes more sense to focus on downstream services, given our traditional strength in application development, rather than core technologies: “There’s massive global investment in AI right now, but ROI [return on investment] is still missing, and that ROI has to be realised at the process level, which is exactly where India excels.”

And right now, says Kirthiga Reddy, CEO and co-founder of AI marketing platform OptimizeGEO.ai, is the right time to enter the applications market. “The first wave of budgets funded pilots because enterprises wanted to understand what was possible. The next wave will focus on what produces measurable value.”

Ultimately, enterprises will continue to invest only when AI moves from an interesting demonstration to a repeatable business outcome, points out Reddy.

This should be seen as a strategic play, says Krish Ramineni, co-founder of AI-powered meeting assistant Fireflies.ai. “Customers don’t wake up wanting a model. They want work to get done…The opportunity isn’t to copy what Silicon Valley builds for less, it’s to build globally important AI companies from India.”

Indeed, the money has kept pace with the narrative.

Nearly 80% of AI investment in India is flowing into application-layer start-ups in areas like enterprise software, fintech and health care, according to a report by venture-capital firm SenseAI Ventures.

But as we approach the fourth anniversary of the so-called ChatGPT moment (November 2022), it is time to ask a harder question: is the progress at the application layer keeping pace with the noise around it?

India’s Advantage

India has a long history when it comes to developing enterprise software. For three decades, India has dominated the export of IT services, thanks to its large pool of lower-cost IT workers. When the traditional software business model was disrupted by the emergence of software as a service (SaaS), the country also saw the rise of a new breed of start-ups such as Zoho.

Hence, many believe that the AI disruption will also open a new window to create global brands from India. Already, there are some enterprise-application pioneers that are making global impact.

One is health-care-imaging AI company Qure.ai, which has raised about $125mn to date. It logged a revenue of roughly $20mn in 2025 and is growing 60–70% a year, with the US alone adding about a quarter to its revenue pie, as reported by Reuters.

However, Indian AI application start-ups are nowhere close to generating the kind of revenue their US or Chinese counterparts command.

In the US, application-layer leaders like Perplexity has logged annualised revenues of over $450mn by early 2026, up from about $100mn a year ago, and bootstrapped image-gen-eration platform Midjourney has been running at over $300mn in annual revenue.

Venture-capital firm Accel’s partner Prayank Swaroop pointed out a related pattern to TechCrunch, an online newspaper. Reviewing over 4,000 applications for Accel and Google’s AI accelerator, he found roughly 70% were “wrappers” layering AI features onto existing models rather than reimagining workflows.

The first is low domestic willingness to pay: India’s digital economy has been built on free or heavily subsidised software for two decades, and even global giants like OpenAI have had to price products far below their US rates to gain traction locally.

The second is a comparatively shallow pool of late-stage capital. Roughly 71% of Indian AI start-ups are still at the seed stage, with only about 2% having reached Series C or beyond, highlights SenseAI Ventures.

The third is a structural cost disadvantage in inference and data. India generates close to a fifth of the world’s data but hosts less than 3% of global data-centre capacity, pushing up the cost of running AI applications at scale domestically.

The Culture Factor

Cultural factors are also limiting India’s ability to leverage its IT talent to realise the true potential of AI-based applications in the enterprise segment.

Barring a handful of exceptions, its best-known enterprise technology companies are cost-arbitrage businesses—highly efficient at executing someone else’s specification, not at owning a product risk of their own.

These businesses reward predictability, tight margins and the disciplined execution of a known scope. However, application-led businesses reward the opposite: the risk of building something that might not work, price for value over cost and sit with years of uncertain product-market fit before revenue follows.

This riskier form requires both a mentality and skill set that is different from what prevails in India’s traditional IT sector.

While the outsourcing industry did help India develop deep pools of expertise, the era-defining business instincts may actually be standing in the way of developing successful applications and products businesses.

“We have not accepted risk-taking as a culture...a country cannot prosper without risk-takers,” Kunal Shah, the founder of fintech companies such as Cred and Freecharge, has argued in the past, noting that founders are routinely pushed towards profitability before they have finished scaling.

Besides a higher risk appetite, application also requires completely new skills compared with services.

Sunitha Viswanathan, a partner at Kae Capital, an early-stage fund, points out that companies now have to focus on things such as building moats around their businesses through proprietary data, deep workflow integration and real switching costs. It is not enough, she says, to have “a better interface sitting on top of someone else’s model”.

Secondly, the traditional Indian playbook—coming up with more affordable versions of successful applications from the West by leveraging lower-cost workers—may not work in the future.

“The bigger opportunity is to build products that are not just cheaper, but better suited to complex and underserved markets,” says Viswanathan.

She points out that start-ups must also invest in product depth, customer trust, brand and global go-to-market capabilities. “That is how a cost advantage becomes a lasting competitive advantage.”

Companies that make this pivot successfully, such as Qure.ai, which now sells into over 3,000 healthcare sites globally, are able to command premium, Western-style revenue multiples that would be far harder to justify on domestic pricing alone.

The Other Side

The AI transition is also about protecting India’s massive IT industry and the associated employment. Over the past three decades, India’s IT majors have built up billions of dollars of labour-arbitrage revenue, most of which is still billed to their clients on the basis of the number of employees deployed on the project.

This is now under threat as AI increasingly automates these jobs.

A new wave of AI-native application start-ups is going after these enterprise budgets, betting that some of the outsourced work can now be delivered as software rather than headcount.

However, those at the helm of the IT industry see it differently. Nilekani is one of those who see this as an opportunity. He points out that enterprises have only started scratching the surface of what AI has to offer, and Indian IT companies have a large business opportunity helping their clients delve deeper.

IT veteran and Tata Sons’ chairman N Chandrasekaran has argued that the edge lies in the contextual understanding of enterprise processes, and that “AI will expand that role much further”.

For now, Indian IT majors have taken several steps in this direction, offering services like AI chatbots. Still, the transition is not going to be easy, says author Viswanathan.

“Are you pushing your product onto clients without understanding their real needs? Are you repackaging the same offering under a new label?” he asks. “A CEO of an IT/AI services company should take a deliberate pause to rethink from first principles. Is it a money problem? No. A people problem? No. It’s a strategy problem,” he adds. However, he notes that the quarterly reporting pressure makes this kind of pause difficult to sustain for anyone but privately held companies.

As AI makes its entry into enterprises, the Indian IT sector faces both a grave threat and a tremendous opportunity. Sectors like quick commerce have proven that Indian founders have the risk-taking ability and vision to build category-defining bets that have no peers in the West.

The more pertinent question today may be: can India’s IT-services establishment and the capital that funds it bring that same appetite to enterprise AI applications, where sales cycles are longer and the payoff is far less visible than in consumer apps?

India’s application edge rests on decades of enterprise-software expertise, deep process knowledge and a few global success stories. Turning that potential into results will take more than cheaper talent.

Founders will need to build products with proprietary data, workflow depth and global reach. Investors will need to back them through the long, uncertain stretch before revenue arrives. And an IT-services establishment built on predictability will have to learn to live with risk.

Until then, “AI use-case capital of the world” is an ambition, not an achievement.

The next few years will show whether India’s application layer becomes the advantage its champions say it is, or remains a story told more loudly than it is proven.