New Delhi: Nadeem Bari Husain, a frail, wrinkly man, sits leisurely at his home in Seelampur, Delhi, as he slowly types “PVC pipe sellers in Delhi IndiaMART” into his son’s laptop. It’s a simple query, the kind he has relied on for years, scrolling through listings, calling numbers and shortlisting suppliers the old way.
His 26-year-old son had recently urged him to try ChatGPT, the popular generative artificial intelligence (AI) chatbot.
“Just type it like a Google search,” he told his aging father. “You won’t have to scroll for hours. You’ll get three-four options; pick one,” he said.
The results came back quickly. Neat, confident, helpful. Each with a direct link, not to IndiaMART, but straight to the seller’s own website. IndiaMART, the platform that had listed them, aggregated them and made them discoverable, was nowhere in the response.
Husain closed the laptop and made his calls. A new ritual, faster than before. No reason to think about what sat behind the answer.
A few hundred kilometres away, in a small industrial unit in Surat, Manoj Shah has been on IndiaMART for six years selling industrial textile machinery. He has never built a website. He never needed to. His IndiaMART listing was his storefront, and his only presence on the internet.
Lately, business has been quieter. “I have not changed my listing,” he said. “Earlier I used to get four-five calls a day. Now I get one or two.”
He cannot explain why. He has not heard of ChatGPT. He does not know what an AI agent is.
Several other small manufacturers and traders Mint spoke to said the same—inquiries have thinned, with no explanation about what has changed. IndiaMART has been their entire digital existence. But somewhere between a buyer’s question and an AI agent’s answer, their phones have begun to ring less often.
IndiaMART, India’s largest business-to-business (B2B) marketplace, was built on a single premise: helping buyers find sellers. The platform has enabled discovery for millions of micro, small and medium enterprises (MSMEs) for two decades. The company had around 42 million active registered buyers, over 124 million listed products and more than 8.3 million registered suppliers as of September, according to its disclosures.
But now, the ground is shifting. In part, IndiaMART has only itself to blame, given how it has been faulted year after year by an American federal agency for facilitating the sale of counterfeit goods, including electronics, apparel and, most worrying of all, pharmaceuticals.
With AI bots bypassing the platform altogether to avoid legal complications stemming from dodgy listings, IndiaMART faces a threat to its revenues. And that has led it to take OpenAI, the maker of ChatGPT, to court.
Piracy fallout
Internally, IndiaMART first noticed something was wrong last October. Their listings were not showing up in ChatGPT responses while competitors—other e-commerce and B2B platforms—continued to appear, two people directly aware of the matter told Mint.
The tech team initially looked inward—checking for blockages at their end. Finding none, they sent emails to customer support IDs at OpenAI’s Silicon Valley offices. The first few responses were automated, AI-generated replies, one of the people said. The first real human response came on 8 November. Ten days later, another response said IndiaMART’s links weren’t being surfaced because of a United States Trade Representative (USTR) caution.
The USTR, a federal agency that develops and enforces American trade policy, has listed IndiaMART as a “notorious market” since 2018 for allegedly facilitating the sale of counterfeit pharmaceuticals, electronics and apparel in large volumes, and failing to implement measures to counter piracy.
IndiaMART has made the non-binding nature of the USTR designation—it carries no legal force outside the US—as the central plank of its case in a petition before the Calcutta High Court.
But, ‘non-binding’ does not mean ‘irrelevant.’ Courts often allow companies to rely on credible third-party risk reports as part of internal compliance and safety frameworks, said Sohom Banerjee, senior research associate at CUTS International, a policy research and advocacy group.
“If OpenAI positions the USTR listing as one input among several in a structured risk assessment process, particularly in categories involving counterfeit pharmaceuticals or consumer harm, it may be seen as a reasonable prudential measure rather than an extraterritorial imposition of US authority,” he added.
Court battle
IndiaMART opened its case against OpenAI on 27 December at the high court. The petition, a copy of which Mint has seen, noted that queries for IndiaMART sellers on ChatGPT were returning direct links to seller websites, bypassing the platform entirely, while the same query for competitor TradeIndia returned links back to TradeIndia’s webpage.
The company also pointed out that several other entities named in the same USTR reports for counterfeiting and piracy continued to appear in ChatGPT responses.
IndiaMART contended this constitutes trade libel through implied disparagement, dilution of its trademark and injurious falsehood. It argued that large AI services have become “intent-driven gateways” that shape what users see, giving them growing economic and strategic power. “Exclusion from such new discovery channels therefore has a catastrophic effect on a service,” it said.
The first hearing in December saw no representation from OpenAI. But at the second hearing, on 13 January, OpenAI’s lawyers were in court.
The ChatGPT maker’s argument mirrored its position in other Indian proceedings. It has contended that it has “no office or permanent establishment in India… The servers on which [ChatGPT] stores its data are similarly situated outside of India”, thereby asserting that Indian courts lack jurisdiction, according to legal filings accessed by Mint.
The Calcutta High Court noted that IndiaMART was being “selectively discriminated and unjustifiably excluded without any logic,” resulting in loss of goodwill, reputation and business. However, interim relief was denied.
The Competition Commission of India has previously flagged similar concerns—warning that AI systems can distort competition and market outcomes through automated decision-making, even when no human intent to harm exists.
Queries sent to IndiaMART and OpenAI were yet to elicit a response.
It is important to note that the regulatory architecture in India has not yet evolved to recognize a “right to be surfaced” inside an AI-generated response.
“As a result, when an AI system decides not to display a commercial entity, the issue does not fall within a dedicated AI inclusion framework. Instead, disputes are filtered through established principles of competitive neutrality, and reputational protection,” explained Banerjee.
Messy for the machine
The legal battle aside, IndiaMART faces integration challenges in the age of AI. Today, every online platform, including Amazon and Flipkart, is aiming to become the preferred answer inside a chatbot, Mint reported earlier. Swiggy now lets users order through ChatGPT, with Zomato recently following suit. Zepto is testing autonomous agents that complete purchases end-to-end without human input.
Business-to-consumer (B2C) platforms such as Swiggy and Zepto plug into AI as clean transaction rails: standardized catalogues, transparent pricing, deterministic payments, predefined fulfilment. Simply put, there is a reliable back end to receive orders placed through a bot.
But for IndiaMART, plugging into AI is not as seamless as it is for these platforms. The company operates on a different premise. Its core unit of value is a lead, a request for quotation, the beginning of a negotiation. Pricing is discussed, not listed. Credit terms, compliance documentation and supplier vetting are baked into every transaction.
“IndiaMART is a heterogeneous directory. It has fragmented suppliers, inconsistent metadata, and categories that include regulated and high-risk goods, with a litigation and enforcement history around counterfeits and pharma,” said Abhivardhan, president, Indian Society of Artificial Intelligence and Law, an industry forum.
“For an AI platform, that raises integration costs (normalization, quality, liability) and makes IndiaMART look less like a neat action provider and more like a noisy corpus that can create safety and compliance headaches,” he added.
The Central Consumer Protection Authority (CCPA), India’s consumer rights regulator, has issued notices to IndiaMART on at least two separate occasions. The first was for listing walkie-talkies without proper frequency disclosure, licensing information, or equipment type approval — a notice it shared with Amazon, Flipkart, and TradeIndia among others. More recently, the CCPA issued a second notice to IndiaMART for listing restricted wireless devices including drone jammers and GPS jammers without mandatory disclosures. The company said it is complying with the law.
The company has today become a directory of millions of suppliers, a far cry from its early days, when it had only a handful of listings. Ironically, IndiaMART traces its origins to founder Dinesh Agarwal’s experiences in the US, the home of the company that he has now taken to court.
The business model
In 1992, Agarwal was setting up information technology services company HCL’s offices on the US East Coast—close enough to Silicon Valley’s early stirrings to understand what was coming, far enough from India to see the gap clearly.
In August 1995, he resigned, returned home and decided to build an online marketplace for exporters and importers. When the government would not give him access to trade directories, he improvised, printing free listing forms and mailing them.
On 1 April 1996, IndiaMART went live, at a time when India barely had any internet users. His cousin Brijesh Agrawal joined him, and together they built what was, at the time, an almost absurd idea: an online directory connecting Indian suppliers to buyers in a country where most businesses had never seen a computer. In the early days, they would print buyer inquiries and fax them to suppliers.
The model was simple: free listings for suppliers, paid subscriptions for those who wanted more visibility and more leads. The more buyers came, the more valuable a paid listing became. It was a classic marketplace flywheel, built for the unglamorous world of PVC pipes, industrial machinery and bulk chemicals rather than consumer goods.
Soon IndiaMART became the go-to platform for millions of MSMEs lacking any other digital presence. Chinese rival Alibaba had effectively retreated from the Indian market.
IndiaMART went public in 2019, raising ₹475 crore at a valuation of ₹3,748 crore, with the initial public offering subscribed over 36 times. The company listed at ₹973 in July 2019 and hit an all-time high of ₹9,780 in February 2021, a nearly tenfold rise in under two years, riding the pandemic-era digital commerce boom. Since then, the stock has lost more than 55% from its split-adjusted peak. The company had executed a 2-for-1 share split in June 2023.
Last January, Japanese brokerage Nomura downgraded its rating for IndiaMART from ‘neutral’ to ‘reduce,’ slashing the target price from ₹3,150 to ₹1,900, citing shrinking paying subscribers, weak customer additions and high churn. Around the same time, Nuvama cut its target from ₹2,500 to ₹1,970, maintaining a ‘reduce’ rating. Motilal Oswal was the outlier, reaffirming a ‘buy’ rating with a revised target of ₹2,600. By June 2025, however, Nuvama flipped to ‘buy,’ raising its target to ₹3,800 and calling it a new demand upcycle. The stock currently trades at around ₹2,183.
Bot or buyer?
IndiaMART’s revenue and profit numbers indicate that it is still recording steady growth. The Noida-based company’s revenue from operations grew 13% year-on-year to ₹402 crore in the third quarter (Q3) of fiscal year 2026 (FY26). Net profit came in at ₹188 crore, though this was significantly inflated by ₹135 crore in other income, of which ₹82 crore was a fair value gain on its strategic investment in Baldor Technologies.
However, that same quarter, IndiaMART stopped sharing traffic data entirely, citing the rise of agentic AI and the proliferation of chatbots, which had made it increasingly difficult to tell bot-driven traffic from human traffic.
“Currently, bot traffic is coming from everywhere. There are now LLM (large language model) bots, search engine bots, new browser crawler bots and agentic bots. Traffic is moving erratically so we removed it,” Agarwal told analysts during this quarter’s earnings call.
All search-dependent B2B directories are seeing noisy, bot-inflated traffic, said Abhivardhan, “but B2B marketplaces feel it earlier—advertisers and sellers care about verified leads, not raw visits. The credibility cost of publishing polluted numbers is higher.”
IndiaMART’s unique business inquiries—the platform’s core measure of buyer-seller connections—grew just 4% year-on-year in Q3 FY26, a sharp deceleration from the 12-17% growth the company had posted in the previous quarters.
Active buyers declined 2% year-on-year to 42 million. Around 1,000 paying suppliers exited the platform during the December quarter, and the total number of paying suppliers fell to 221,000.
Agarwal attributed the slowdown to seasonal factors and lower advertising spends during the festive season.
But, as things stand, the visibility IndiaMART built its business on is becoming harder to measure. The demand it once intermediated is being quietly delegated to machines it has no relationship with. B2C platforms can afford to be inside the AI. B2B may have no choice but to fight it, and IndiaMART, almost accidentally, has fired the first salvo.