ETMarkets Smart Talk | Retail traders lost ₹1.8 lakh crore in F&O: Can algo trading change the game? Rakesh Pujara explains

India’s retail derivatives market has expanded rapidly, but the numbers behind that growth are sobering. More than 90% of individual F&O traders lost money between FY22 and FY24, with cumulative retail losses estimated at around ₹1.8 lakh crore. At the same time, algorithmic trading has become deeply embedded in the broader market, accounting for a significant share of both cash and derivatives activity.

So, could automation be the missing piece for retail traders? Rakesh Pujara, Founder and Managing Partner, Compounding Wealth Advisors LLP, believes the biggest shift is not that retail traders are suddenly becoming more sophisticated, but that tools once available primarily to institutions are now becoming accessible to individuals. APIs, AI-assisted strategy builders, backtesting platforms and cheaper infrastructure are lowering the barriers to algorithmic trading.

In an interaction with ETMarkets, Pujara explains why retail algo adoption could rise sharply from current levels, how AI is changing the way strategies are built and executed, and whether automation can help address the discipline, emotional and execution challenges that have historically hurt retail F&O traders. Edited Excerpts -

Q) For decades, algorithmic trading was largely the domain of institutions and sophisticated traders. Why is this suddenly becoming accessible—and attractive—to the Indian retail trader?

Three things converged at once. First, infrastructure that used to sit exclusively with institutions and prop desks is now available over an API from any broker. 5Paisa, Zerodha, Upstox, Angel One and others opened up direct API access years ago, and NSE formalized this further with its retail algo framework that came into effect on August 2025, following SEBI's February circular aimed at regulating grey areas in retail algo trading. That's a big shift from a decade ago, when algo access effectively meant colocation servers and institutional-grade capital.

Second, AI has cut both the skill and the time required to build something that works. A retail trader no longer needs to write execution logic from scratch or understand market microstructure in depth. Low-code platforms, AI-assisted strategy builders and backtesting tools have compressed what used to be a multi-month build into something that can be prototyped in days.

Third, and this doesn't get said enough, the cost of infrastructure has simply collapsed. Cloud compute, market data feeds and even colocation-adjacent speeds are available at a fraction of what they cost institutions ten years ago. Put those three together, and what was once a closed club has a much lower door now.

Q) Retail participation in derivatives has exploded over the last few years. Is AI and automation emerging because traders are becoming more sophisticated—or because manual trading is simply becoming too difficult?

I'd put more weight on accessibility than on either sophistication or difficulty. AI is reshaping how work gets done across every industry right now, and trading is not an exception. The retail trader doesn't need to become a quant to use a rules-based system. That drop in required skill, more than anything else, is what's driving the current wave toward automation.

That said, I wouldn't dismiss the pain-point angle either. Manual execution across multiple strikes and expiries genuinely is harder than it used to be, especially with weekly expiries and the volume of instruments retail traders now track. Those frictions are real and they do push people toward automation. But if I had to rank the two forces, easier access to the tools is doing more of the work than traders getting smarter or markets getting harder to trade by hand.

Q) What is the biggest pain point that AI and automation are solving for a retail trader—lack of time, lack of discipline, lack of data analysis or emotional decision-making?

Honestly, I think the question assumes the bottleneck was one of those four things. It wasn't. The bigger constraint for most retail traders was that automation itself was out of reach, not that they lacked discipline or time in the abstract. You can want to remove emotion from your trades and still have no practical way to do it if building or accessing a system requires capital, coding skill and infrastructure you don't have.

Once that access barrier comes down, the other pain points do get addressed too. Discipline improves because a system executes without hesitation. Time gets freed up because you're not watching five charts simultaneously. Data analysis improves because backtesting is now something a retail platform offers out of the box, not something you needed a Bloomberg terminal for. So all four matter, but they were symptoms of the same underlying problem: automation wasn't accessible. Now it is, and the rest follows from that.

Q) How large is the AI and algorithmic trading opportunity in India today, particularly on the retail side? Are we talking about a niche community of sophisticated traders or the beginning of a potentially massive new industry?

The numbers suggest this is well past the niche stage on the overall market side. Algorithmic trading accounted for 57% of equity cash trades and 70% of derivatives trades in India as of the last financial year, and by some measures algo-linked activity including colocation and DMA makes up around 90% of orders placed on the NSE, even after accounting for a large share of those getting cancelled. That already puts India in the same range as more mature global markets.

But that headline number is misleading if you're specifically asking about retail. SEBI's 2023 data showed only about 13% of retail traders were using algorithms, compared with 97% of foreign portfolio investors and 96% of proprietary trading desks. That gap, not the overall percentage, is where I'd point anyone asking about the size of the opportunity. Retail is nowhere near where institutions already are, and the infrastructure and regulatory groundwork with the new NSE retail algo rulebook is being laid specifically to close that gap.

There's also a harder number behind why this matters: more than 90% of individual F&O traders lost money between FY22 and FY24, with retail losses totalling roughly ₹1.8 lakh crore over that period. That's not a footnote, that's the actual case for automation on the retail side. A market where the vast majority of manual retail participants are losing money, sitting next to institutions that are almost universally automated, is not a stable equilibrium. I'd expect the retail algo share to climb meaningfully from that 13% base over the next few years.

And longer term, I think it's a fair bet that we eventually reach a market structure where a large share of retail flow is also algo-driven, meaning algos are increasingly trading against other algos rather than against manual retail orders. We're not there yet, but the direction of travel is fairly clear.

(Disclaimer: Recommendations, suggestions, views, and opinions given by experts are their own. These do not represent the views of the Economic Times)