More Stories

Informist, Wednesday, Sept. 9, 2026

 

--Angel One Kenghe: Herding risk in algo trading can move mkt via big volume

--CONTEXT: Angel One group CEO Kenghe's comments at Global Fintech Fest

--Angel One Kenghe:Liquidity drying up when needed key risk for algo traders

--SEBI member: Changes by AI model to algo strategies a monitoring challenge

--CONTEXT: SEBI Whole-Time Member Varshney's comments at Global Fintech Fest

 

MUMBAI – Dynamic changes by artificial intelligence models to algorithmic strategies deployed by market participants pose a big monitoring challenge for the regulator, Securities and Exchange Board of India's Whole-Time Member Kamlesh Chandra Varshney said. Speaking at the Global Fintech Fest 2026, Varshney wondered aloud whether the use of AI in algorithmic trading has the ability to cause a market crash similar to the one in 2010.

 

Varshney was referring to the May 2010 flash crash in the US equities market following triggering of a large volume of automated selling orders. The market recovered quickly, however. Ambarish Kenghe, group chief executive officer of Angel One Ltd. said at the same session that retail traders should be aware that a key risk in algorithmic trading was herding taking place and moving the market in one direction with large volumes.

 

Liquidity dries up at times in the equity market and if liquidity is needed for algorithmic orders to get executed then this can create a problem for traders, Kenghe said. Backtesting from historical data, a useful AI tool used by algorithmic traders, also has the risk that it may show good results of certain algorithmic strategies but these may not be successful when applied in a future live trading environment, Kenghe said. Some traders tend to "overfit the curve" when evaluating the backtesting results, he said.

 

Retail traders using algorithmic trading tools should keep these risks in mind, according to Kenghe. But as far as regulators are concerned, the risks from randomised orders from algorithmic strategies may be overstated since the randomisation often leads to deterministic outcomes, he said.

 

Algorithmic trading is in the limelight now for retail investors but it has been in use in the raw form for over 10 years, according to Shrini Viswanath, co-founder of Upstox Securities Pvt. Ltd. He said that around 12 years ago, a retail customer came to them and said he was "reverse engineering a web platform and found a way to place an order automatically, I am getting the data through Yahoo Finance, and I'm doing my own algorithmic training".

 

The evolution happening in retail algorithmic trading is just the start, according to Viswanath. "With the advent of AI, you don't even need to know how to code. You can actually describe what you want to run, what your thesis is on an investment strategy, and then backtest it and see," he said.

 

SEBI's Varshney highlighted the need to differentiate white and black boxes in algorithmic strategies. "White box is simple execution where the logic is transparent. But in black box the logic is not transparent, it is a proprietary strategy of the person who is writing that software," he said.

 

SEBI has mandated that only registered research analysts can come out with a black box algorithm strategy, according to Varshney. But a retail algorithmic trader will find it hard to trust a black box strategy because he or she wouldn't know how it was created, he said.  End

 

Reported by Rajesh Gajra

Edited by Rajeev Pai

 

For users of real-time market data terminals, Informist news is available exclusively on the NSE Cogencis WorkStation.

 

Cogencis news is now Informist news. This follows the acquisition of Cogencis Information Services Ltd. by NSE Data & Analytics Ltd., a 100% subsidiary of the National Stock Exchange of India Ltd. As a part of the transaction, the news department of Cogencis has been sold to Informist Media Pvt. Ltd.

 

Informist Media Tel +91 (22) 6985-4000

Send comments to feedback@informistmedia.com

 

© Informist Media Pvt. Ltd. 2026. All rights reserved.