While Silicon Valley and Beijing race to claim dominance in the generative AI era, India—long considered the back office of the global tech industry—is facing a sobering reality check. Despite its massive pool of engineering talent, the country is reportedly struggling to produce the kind of deep-tech AI startups capable of competing on a global scale.
V. Balakrishnan, the chairman of Exfinity Venture Partners and former CFO of Infosys, recently highlighted this growing gap. According to the veteran investor, the Indian startup ecosystem is currently 'short' on high-quality AI companies that go beyond simple applications to solve fundamental technological problems.
The Quantity versus Quality Gap
On paper, India's tech scene looks more vibrant than ever. The country boasts thousands of new ventures every year and has the third-largest startup ecosystem in the world. However, many of these companies are classified as 'AI startups' only in the most superficial sense.
Balakrishnan’s critique centers on the fact that many founders are building what the industry calls 'thin wrappers.' These are applications that simply put a slightly different user interface on top of existing models like OpenAI’s GPT-4 or Google’s Gemini. While these tools can be useful, they do not own the underlying technology and lack a 'moat'—a competitive advantage that prevents others from easily copying them.
For a startup to be truly 'strong' in the eyes of institutional investors, it needs to be doing something unique at the architectural level. This might include developing proprietary large language models (LLMs), creating hardware-efficient AI, or applying AI to specialized datasets in sectors like healthcare or agriculture in ways that others haven't yet mastered.
Why This Matters
- Economic Sovereignty: If India relies entirely on foreign AI models, it risks becoming a mere consumer in a market where the rules and costs are set by US and Chinese firms.
- VC Sentiment: Venture capital is becoming more discerning. Investors are moving away from 'copycat' models toward startups with defensible intellectual property.
- Global Competitiveness: To transition from a service-led economy to a product-led one, India must own the core technologies defining the next decade.
The Infrastructure and Capital Hurdles
One reason for the shortage is the sheer cost of building foundational AI. Training a world-class model requires thousands of high-end GPUs and millions of dollars in compute power. For many Indian startups, accessing this level of infrastructure is prohibitively expensive.