India is pouring its power into catching up with the US and China on constructing foundational Synthetic Intelligence (AI) fashions. Authorities officers are discussing billion-dollar investments in knowledge centres and computing infrastructure. Lecturers are calling for sovereign giant language fashions (LLMs). Trade teams are lobbying for extra public-private collaborations in mannequin coaching. However all this focus is on yesterday’s downside.
The world doesn’t want extra foundational fashions. It already has dozens. Open-source options like DeepSeek, LLaMA, Qwen, and Mistral are freely accessible — and bettering at a breakneck tempo. DeepSeek not too long ago launched a mannequin that rivals GPT-4 in reasoning benchmarks. Qwen, developed by Alibaba, has turn into a top-tier mannequin for multilingual duties. India doesn’t must construct its personal from scratch. It could take these fashions and run with them. The true alternative isn’t in recreating what already exists. It’s in doing what India has at all times achieved greatest: Constructing on high of what’s already there.
That’s precisely how India grew to become a superpower in IT providers. It didn’t invent the microchip, but it surely constructed billion-dollar corporations that helped the world use microchips. It didn’t create working methods, but it surely grew to become the worldwide centre for enterprise software program improvement. India didn’t pioneer cloud computing, but it surely gave rise to SaaS (software program as a service) giants like Zoho and Freshworks. It didn’t create banking infrastructure, but it surely constructed UPI — the world’s most superior funds platform.
Now, the same alternative is rising in AI. The subsequent frontier shouldn’t be extra fashions— it’s platforms. In my e-book, From Incremental to Exponential, I defined how platforms grew to become Silicon Valley’s largest strategic benefit. They enabled trillion-dollar firms by permitting others to construct on high of their infrastructure— creating digital ecosystems that attracted builders, locked in customers, and fuelled unstoppable community results. India has the chance to duplicate this success within the AI period. platform gives instruments, attain, and scale. The perfect platforms turn into the spine of complete industries. And the subsequent wave of platforms shall be powered by AI.
After which, there are AI brokers. These are autonomous methods that may take a aim — “analyse Infosys’ inventory” or “plan a three-city trip”— and break it down into steps, execute these steps, and ship outcomes. They browse the online, write code, summarise paperwork, and work together with different software program. When constructed nicely, they behave like junior analysts or digital interns. When constructed brilliantly, they’ll turn into the infrastructure for future work — working methods for decision-making, technique, and execution.
Some early AI brokers, like China’s Manus, are already demonstrating what’s attainable. However the tempo of change is astonishing. Inside weeks of creating headlines, each Manus and DeepSeek are already being eclipsed by newer entrants and much more highly effective instruments. The AI area is shifting so quick that even OpenAI is scrambling to take care of its lead. Manus doesn’t personal a foundational mannequin. It wraps current fashions like Claude 3.5 and Qwen and combines them with sensible planning, orchestration, and gear integration. It integrates almost 30 instruments — from file readers to net browsers — and delivers outcomes, not simply responses. It’s been known as a “wrapper”, but it surely’s actually a brand new form of platform: One which transforms AI from passive assistant to lively problem-solver.
That is the place India has traditionally excelled. Its strengths lie in constructing on high of current methods — designing, deploying, and scaling platforms. Consider what Infosys, TCS, and Wipro did for IT. What Zoho and Freshworks did for SaaS, Adar Poonawalla’s Serum Institute did for vaccines — producing billions of doses not by inventing new medication, however by mastering large-scale manufacturing and distribution. India didn’t pioneer drug discovery, but it surely grew to become the ‘pharmacy of the world’ by perfecting the manufacturing of inexpensive generics. Solar Pharma and Cipla targeted on manufacturing and distribution at scale, whereas Biocon, underneath Kiran Mazumdar-Shaw, made insulin and biosimilars inexpensive by optimising current improvements.
India has the potential to do the identical in AI. It may construct AI brokers tailor-made to well being care, agriculture, logistics, schooling, and extra. It may develop orchestrators for companies, and public platforms that democratise AI entry for small enterprises and underserved communities — not simply in India, however globally. Think about a public AI stack, modelled after UPI, the place verified brokers can entry knowledge, execute duties, and collaborate securely throughout domains.
As an alternative of making an attempt to make amends for yesterday’s fashions, India can leapfrog into tomorrow’s platforms — and construct the identical kinds of trillion-dollar monopolies that Silicon Valley has. Keep in mind that Apple and Microsoft didn’t construct any foundational methods, they merely constructed nice options on high of what already existed — and they’re amongst probably the most precious firms on the earth.
Furthermore, the AI panorama is evolving so quickly that immediately’s leaders may be eclipsed in months. LLMs will quickly be yesterday’s information and AI brokers might dominate the subsequent wave — however one thing much more disruptive will comply with. Innovation is accelerating at an unprecedented tempo, and people who construct adaptable platforms — somewhat than static fashions — shall be greatest positioned to trip these exponential curves. By specializing in constructing the platforms and functions that others depend on, India couldn’t solely leapfrog the present wave however rise above the turbulence of fixed disruption.
Vivek Wadhwa is CEO, Vionix Biosciences. The views expressed are private
