According to Zerodha co-founder Nikhil Kamath's latest analysis, founders leading investment pitches with 'we use AI' are risking losing investor attention. As reported by NDTV Profit, Kamath stated that 'the number of investment decks I get leading with 'we use AI' is ridiculous' and has reached a point where it automatically makes his eyes roll. He has 'almost stopped looking at them' and emphasized that 'AI is just table stakes now' and that bragging about using AI is like boasting about taking a shower. Kamath warned that opening a pitch with AI is 'a surefire way to get ignored, not just by us, but by most serious VCs'.
According to Zerodha co-founder Nikhil Kamath's analysis on X, the economics of data centres have fundamentally changed. As reported by NDTV Profit, Kamath argues that the building is now barely a rounding error compared to the silicon inside it. His breakdown shows that of every ₹100 spent on a large AI data centre, approximately ₹40 goes to chips alone, which is more than the combined cost of building, power, cooling, networking and land. Land accounts for barely ₹1 of that ₹100, highlighting the dramatic shift in value proposition. The analysis cites data from Bernstein and Epoch AI estimates to support these calculations.
As reported by NDTV Profit, Kamath's analysis reveals starkly different approaches between companies and governments regarding AI infrastructure ownership. Global hyperscalers own only around 60% of their data centre capacity, with the remaining 40% leased because owning sinks returns. However, governments prioritize jurisdiction over hardware ownership, as exemplified by India's AI programme deploying more than 38,000 GPUs under agreements that keep hardware in private hands but data within Indian jurisdiction. Kamath's framework emphasizes that 'rent what depreciates, own what endures' applies differently to companies versus countries, with governments needing jurisdiction rather than hardware ownership for strategic control.
Kamath's criticism points to a broader problem as AI tools have democratized the ability to generate professional-looking investment decks. As reported by NDTV Profit, while this has lowered the barrier to producing polished presentations, it has created a problem where startups increasingly risk sounding identical. Kamath noted that 'especially now that AI has democratised the ability to generate 'nice-looking' decks, the last thing you want to do is say the exact same generic things as every other founder using the exact same tools'. His comments highlight the risk of confusing the use of widely available technology with genuine business differentiation, as investors increasingly look beyond AI adoption to how effectively startups apply the technology to build sustainable businesses.
For founders, Kamath's analysis suggests that the focus may need to shift from AI usage to fundamental business questions. As reported by NDTV Profit, he emphasized that founders need to realize that 'the harder question isn't can AI do this? — It's why this needs to exist, who actually needs it, and why they would keep using it?'. The consensus among founders is that while AI enables capabilities, the real value lies in proprietary data, rebuilt workflows, and unique distribution channels that survive model upgrades. Kamath's framework suggests that 'not every task needs the same machine, or environment', with the infrastructure landscape evolving beyond traditional real estate considerations.
