By Anup Mishra, Co-Founder, Finifi
India’s fast commerce industry is expected to reach GMV of $10 billion+ in 2025 and grow at about 40% per year till 2030. At the moment, Blinkit has 45-46% market share, Swiggy Instamart 25-27%, and Zepto 21-30%, depending on the city and the metric. These three players combined have over 90% of the consolidated market.
The reason why this market is interesting in the realm of AI is that the demand problem is very specific. Every time a consumer searches for “shampoo” or “oat milk” on any one of these platforms, an AI layer determines which brands are first, which brands are out of stock and which brands are included in the personalised recommendations feature.
However, the bigger shift that nobody is talking about is Assortment Control. Platforms now decide what they show to consumers, be it seasonal, festive or even daily purchases.
At Finifi, some of the biggest brands trade on these channels, and we have consistently seen the revenue split for Qcom / Ecom for even traditional brands increasing at an astounding rate. The revenue from the Qcom / Ecom channel is not rare and is around 15-20% for the biggest brands (5000+ Crore) as well.
What we’ve seen is that the brands with high availability scores, strong conversion data and clean operational metrics are rewarded with better placement while others miss out on this visibility.
A brand that is in stock with the forecasted demand will get better placement, thus more sales, more data, and even better forecasting. Many companies that have taken their AI use beyond separate pilots and are making it part of the fabric of their business are already seeing the impact. Stord’s 2026 State of AI report revealed that these firms are closing the income gap and driving a 40% increase in revenue and 30% in Customer Lifetime Value over their counterparts who are still using disjointed implementations.
Swiggy Instamart achieved an average delivery time of 13 minutes in fiscal 2025 by optimising its routes and boosting the number of dark stores. In certain cases, Blinkit has achieved sub-3-minute delivery in 2025, which can only happen when the AI layer that predicts demand has placed stock within a very short physical distance of potential customers.
Amazon has also ventured into the fast delivery segment in India with Amazon Now in September 2025, challenging the market with 10-15 minute delivery time. Competition for everyone to keep their algorithms as efficient as possible will only grow more intense.
There is a simple message for every brand selling through these channels: consumer buying habits are changing quicker than many brands are aware of, and the dark store operators are making decisions on SKUs and inventory allocation based on data which brands are benefiting from, not just contributing to.
AI is now considered the number one priority for 84% of these eCommerce companies. But priority and capability are two different things. Only 7% of organisations have truly scaled their AI initiatives, but they are not simply copying and pasting what they are already doing. They have changed the way they make decisions.
The Bottom Line:
Now, predicting consumer demand is an art. The margin for error has reduced to virtually zero with Q-commerce and e-commerce. Saturday night, no stock-out is possible. If the pricing is done wrong at a sale weekend, it’s a real loss of money. The distance between a brand that anticipates and a brand that is constantly reacting is growing, and AI is the game-changer.