“I don’t assume your classification of India within the second tier is appropriate. India is clearly within the first.” he stated
Vaishnaw stated the rationale for India truly being within the first group is “there are 5 layers within the AI structure. The appliance layer, the mannequin layer, the chip layer, the infra layer, and the power layer. We’re engaged on all of the 5 layers, making excellent progress in all of the 5 layers.”
“On the appliance layer, we’ll most likely be the most important provider of providers to the world, go to an enterprise, perceive the enterprise of enterprise, perceive the working of that enterprise, and supply that service utilizing AI functions. That is going to be the most important issue of success or profitable deployment of AI, as a result of that is the place ROI comes from,” he stated.
Vaishnaw additionally stated that constructing extraordinarily massive AI fashions alone doesn’t give international locations actual power. “ROI does not come from creating a really massive mannequin. Ninety-five % of the work can occur with fashions that are 20 billion or 50 billion parameters. We’re making a bouquet of such fashions. We have already got. We have already got a bouquet of such fashions, which at the moment are being deployed in a number of sectors to extend the productiveness, to extend the effectivity, to extend the efficient use of know-how,” he stated.
Vaishnaw’s pushback at Davos highlights a elementary divergence in perspective. Conventional international establishments usually equate AI management with frontier mannequin scale, hyperscale compute focus and the dominance of a handful of huge know-how corporations. India’s argument, in contrast, is that AI energy should be evaluated throughout all the worth chain, together with diffusion, software and financial influence. Vaishnaw’s five-layer framework gives a coherent rationalization for why India’s AI trajectory seems totally different however no much less vital.
Vaishnaw’s articulation of AI improvement throughout 5 layers — the appliance, mannequin, chip, infrastructure, and power layers — gives a helpful analytical construction to know India’s method. Relatively than chasing symbolic management by trillion-parameter fashions alone, India is trying to construct balanced functionality throughout every layer, making certain that advances on the high of the stack are supported by resilient foundations under.This layered roadmap mirrors India’s earlier digital public infrastructure technique, the place success got here not from proprietary dominance however from scalable, interoperable techniques that may very well be broadly adopted. In AI, the identical philosophy is being utilized: breadth of adoption issues as a lot as technical novelty.
Additionally Learn | Davos 2026: World curiosity grows in India’s method to AI, says Vaishnaw
Software layer: India’s core benefit
Vaishnaw’s strongest declare lies within the software layer, the place India’s long-standing strengths in IT providers, enterprise integration and course of understanding converge with AI. Indian corporations are deeply embedded within the operational workflows of world enterprises, giving them a bonus in translating AI capabilities into measurable productiveness features.
By emphasising that return on funding doesn’t come from constructing ever-larger fashions however from deploying fit-for-purpose options, Vaishnaw makes a realistic financial argument. Most real-world enterprise use-cases akin to automation, forecasting, optimisation and determination help, don’t require frontier-scale fashions. Fashions within the 20 to 50 billion parameter vary, which India is actively creating and deploying, are sometimes ample and more cost effective.
If AI adoption is finally judged by how deeply it improves enterprise outcomes and public service supply, India’s give attention to functions might show extra sturdy than a race for headline-grabbing mannequin dimension.
Mannequin layer: A bouquet method
India’s mannequin technique, as described by Vaishnaw, is intentionally pluralistic. As a substitute of betting on a single nationwide massive language mannequin, the nation is constructing a “bouquet” of medium-scale fashions tailor-made to particular domains and sectors. This method aligns with India’s variety of languages, industries and regulatory contexts.
Such a method additionally reduces danger. It avoids overconcentration of assets and permits innovation to emerge from startups, analysis establishments and public–non-public collaborations. Whereas these fashions could not dominate international discourse, their deployment throughout a number of sectors enhances India’s AI penetration, a metric the place worldwide assessments already rank the nation among the many international leaders.
Additionally Learn | India to continue to grow 6-8% in actual phrases in subsequent 5 years: Ashwini Vaishnaw at Davos
Infrastructure and expertise
Vaishnaw’s acknowledgment of compute as certainly one of India’s greatest challenges lends credibility to his broader argument. AI management is constrained not simply by expertise or concepts however by entry to inexpensive, scalable computing energy. India’s response — a public–non-public partnership mannequin that has empanelled round 38,000 GPUs right into a shared nationwide compute platform — is a big coverage innovation.
By subsidising entry to this infrastructure and providing compute at roughly one-third of prevailing international prices, India is reducing entry boundaries for college students, startups and researchers. This method prioritises diffusion over exclusivity, making certain that AI functionality doesn’t stay locked inside a small elite of well-funded actors. Whereas the size nonetheless lags behind the hyperscalers of the US and China, the intent is to maximise financial influence per unit of compute moderately than absolute capability.
Whereas India could excel in diffusion, it nonetheless faces gaps in deep analysis, semiconductor self-reliance and foundational breakthroughs. The absence of domestically produced cutting-edge AI chips and restricted presence on the very frontier of mannequin analysis stay constraints that can not be ignored. Nevertheless, India has launched emga initiatives for chip manufacturing that are anticipated to point out leads to only a few years.
Worldwide rankings that place India second globally in AI expertise help Vaishnaw’s competition that the nation belongs within the high tier. India’s massive pool of engineers, researchers and practitioners creates a virtuous cycle of adoption and innovation. But retaining high expertise stays a problem, as international competitors for AI experience intensifies and frontier analysis continues to gravitate towards better-funded ecosystems.
The power layer
Together with the power layer within the AI structure is especially notable. AI techniques are energy-intensive, and future progress will depend upon the provision of dependable, inexpensive energy. India’s determination to deal with power as a core a part of its AI roadmap displays an understanding that digital ambitions can’t be separated from bodily infrastructure.
Nevertheless, this layer additionally represents a vulnerability. Energy availability, grid stability and the environmental footprint of information centres stay structural challenges. With out sustained funding in clear and scalable power, AI growth dangers operating into exhausting bodily limits, particularly as compute demand accelerates.
A special definition of AI management
Vaishnaw’s pushback in opposition to the IMF’s classification finally rests on redefining what it means to be an AI energy. If management is measured by who builds the biggest fashions or controls probably the most compute, India could certainly seem to lag behind the first-tier group which incorporates the US and China. But when management is judged by preparedness, penetration, expertise and actual financial influence, India’s case for being within the “first group” turns into way more persuasive.
The five-layer roadmap reveals a rustic trying to combine AI into the material of its financial system moderately than treating it as a standalone technological contest. The problem forward might be to maintain momentum throughout all layers concurrently. Success will rely not simply on coverage imaginative and prescient however on execution, coordination and the power to beat persistent constraints in compute, power and superior analysis.
In that context, Vaishnaw’s argument is smart not as a result of it denies India’s limitations, however as a result of it frames them inside a coherent technique geared toward long-term, inclusive AI progress moderately than symbolic dominance.
