Zepto Is Building India’s Urban Grocery Supply Chain Around Quick Commerce

Aadit Palicha described Zepto’s scale in physical, not just digital, terms: millions of deliveries per day, crores of monthly transacting users, millions of people ordering every day, and a workforce across the platform that includes delivery partners, pickers, packers, truck drivers, and operations roles. The consumer sees an app that can deliver groceries in about 10 minutes. Palicha framed the company underneath that interface as logistics, retail operations, supply chain, automation, and everyday staples. The 10-minute promise, in his account, is not merely an app feature; it is an infrastructure bet.

He rejected the idea that Zepto is fundamentally a software company. “If you remove all the software and the tech and the dark stores,” he said, “fundamentally we’re an atta, dal, chawal company.” Customers think of Zepto as the place they get flour, rice, vegetables, and milk, not as a grand technology company. That clarity, in his telling, shapes the operating system: software matters because it improves the physical movement of groceries.

The visible app sits on top of hundreds of operational choices inside the dark stores: how selection is designed, how the store is oriented, what infrastructure is installed, how replenishment algorithms work, how trucks move goods in the back end, and how tasks are assigned and managed. Palicha said Zepto has also built industrial-grade automation in its back-end supply chain to move products from suppliers such as Unilever, Procter & Gamble, and Coca-Cola through the network with fewer people and faster replenishment to dark stores. A meaningful part of the technology team, he added, now works on hardware and robotics.

The business logic he gave for that operational depth was direct: every rupee saved in the supply chain can either be passed to the customer or reinvested in better selection, better delivery times, or other improvements. Efficiency was not presented as separate from the customer promise; it was the mechanism for making that promise durable.

The clearest example was fruits and vegetables. Palicha said Zepto runs “one of India’s largest, or at this point the largest,” fruits and vegetables supply chains in the country. The company sources millions of units per week from farmers across India, including strawberries from Mahabaleshwar and produce from areas around Karnataka, Bangalore, and Hoskote. When Jared Friedman described Palicha, at 23, as the largest seller of fruits and vegetables in India, Palicha qualified the statement: “on the platform.”

Palicha still described the company as early. Grocery remains a large market, he said, with more efficiency to extract from the supply chain and the consumer side of the business. One newer growth area is advertising: brands such as Coca-Cola, Pepsi, and Nestlé bid on search keywords in the app. Palicha said the ads business was inconsequential two years earlier but now generates hundreds of millions in ARR.

Aadit Palicha said Zepto did not begin with the intention of building a company. He and co-founder Kaivalya Vohra were 17, back in Mumbai during COVID after expecting to study in California, and saw little point in “COVID college online learning.” They took a year off to work on a project.

The first problem was immediate: during the first wave of the pandemic, people in their neighborhoods could not reliably get groceries. Small mom-and-pop stores had labor issues. Larger delivery players offering next-day service were backlogged. Palicha and Vohra started a WhatsApp group and delivered groceries to neighbors in Mumbai themselves. That became an app called Kirana Kart.

The origin matters because Palicha’s explanation for Zepto’s later decisions depended on proximity to customers. The first users were not an abstract market segment. They were “the 30 aunties in our neighborhood,” people whose groceries the founders were physically carrying to their doors.

Jared Friedman recalled that Palicha had been admitted to Stanford and was two or three months from matriculating when he asked whether he should go or keep building the company. By then, Kirana Kart was already in motion and the founders were close enough to see signs of whether the business was working. Friedman joked that the only thing he did was convince Palicha not to go to Stanford; Palicha said Friedman, his YC group partner, did more than that.

Palicha resisted making the dropout decision sound heroic. His answer was tactical. They did not take the major risk until they had evidence that the product was working. Eight or nine months into the journey, he said, they saw early signs of product-market fit: retention, compounding, and scale. By the time they chose to commit, they were doing roughly 10,000 orders per day and had reached what he described as a 60–70 crore revenue or GMV run-rate scale. They also had investor interest and a term sheet.

His advice from that experience was not “drop out.” It was to have a real product, some product-market fit, and proof of concept before taking the leap.

Grocery delivery was not an undiscovered category in India when Zepto began. Jared Friedman noted that there were already multiple large players, including billion-dollar companies, operating at scale. Aadit Palicha’s explanation for entering anyway was that the founders were not reasoning from market maps. They were reasoning from customer dissatisfaction.

In 2020, Palicha said, India already had several grocery delivery models: delivery from existing stores, next-day delivery, two-day delivery, and other attempts to digitize and organize grocery. The opportunity had been pursued for years, including offline. But when Palicha and Vohra handed grocery bags to customers at their doors, they asked why customers used Kirana Kart, why they did not use other services, and why they did not simply shop themselves.

The answer was not that Kirana Kart had solved the problem. It had not. Customers complained that selection was incomplete, pricing was too high, and delivery times were not sufficiently controlled. Palicha reduced the demand to four variables: speed, quality, selection, and price. The question became how to improve all four at once.

That search produced the dark-store pivot. The first version was not a polished infrastructure plan. Palicha said he and Vohra took stock and put it in Vohra’s apartment. Friedman called it the first dark store; Palicha agreed. They then tried what he described as a “mini cardboard version of a warehouse” in Juhu, and eventually a proper warehouse.

The important observation was not a small gain. Palicha said the neighborhood served from the dark store produced three to four times the volume of the rest of the city for them. He described the effect as a “10x” improvement in customer response, because controlling inventory and operations let them control delivery times, quality, selection, and pricing.

The pivot also changed the company’s model and name. Kirana Kart was the delivery-from-existing-stores version; Zepto became the controlled dark-store infrastructure version. Palicha said many observers thought the move was wrong at the time. His explanation was that critics were reasoning backward from what seemed financially or operationally viable. The founders were asking what customers most wanted, then working backward to the system that could provide it.

Aadit Palicha’s operating philosophy was that the starting point of a business should be the customer, not the current limits of the business. Others in the category, he said, often began with the model they believed was viable and then gave customers the best experience that model could support. Zepto began by asking for the most extreme customer experience imaginable.

For Zepto, that answer became 10-minute grocery delivery. Palicha connected the thinking to Brian Chesky’s “5-star, 7-star” concept: imagine an unbelievable customer experience first, then work backward to make it possible at scale.

The operational economics followed from that demand. If customers loved the experience, demand could exceed forecasts. If demand concentrated in a mini-warehouse, throughput rose. If throughput rose in the same physical footprint, costs fell. Palicha’s claim was that these advantages were hard to forecast in advance because they depended on how strongly customers pulled the product.

He put the point bluntly: it is impossible to build a financially viable, profitable business without customer delight. In his formulation, “customer delight is where financial value starts.”

That was also how he explained the early speed of iteration. Palicha and Vohra were doing deliveries themselves, testing changes, and changing the model based on what they saw. Palicha said that account skips over many failures; he remembered at one point in the YC batch being one of the worst-performing companies at getting tasks done. But he said their youth and naivety helped because they did not yet know all the reasons the business should be hard.

COVID lockdowns created an unusually insulated environment. The noise of startups, VCs, blog posts, and industry conventional wisdom fell away. The only question was why the 30 or 40 people around them were not ordering more. If a customer was unhappy, Palicha said, “it was right in front of our face.” Friedman connected that to the YC advice to make 100 users love the product rather than trying to get thousands of indifferent users. Palicha agreed: that focus revealed that Kirana Kart was not good enough and forced the pivot.

Aadit Palicha described Zepto’s long-term ambition as building urban grocery infrastructure for India. Major countries, in his view, have their own retail, logistics, and supply chain infrastructure for day-to-day goods. Consumers in the US and Europe buy through more organized systems, with lower costs and more sophisticated quality control. Zepto’s intended role is to be an organizing force in India’s grocery supply chain.

The nearer-term goal he named was to become a large grocer in the country over the next four or five years across the top 40 or 50 cities. If Zepto can do that, he said, it could become India’s homegrown version of Amazon or another major ecommerce platform — not by copying a foreign model, but by making hyperlocal ecommerce work in a format suited to India.

Palicha said he does not think anyone has really cracked this hyperlocal ecommerce format anywhere in the world, but he believes it works particularly well in India. If Zepto does crack it, he said, it could produce a large outcome and help other businesses grow with it.

One such effect is already visible in consumer brands. Palicha said Zepto and other quick-commerce platforms have become distribution platforms for startups building their own consumer brands. He compared the possibility to the role played in the US by retailers such as Walmart and Costco decades earlier, where distribution infrastructure helped create large consumer-brand outcomes. His hope is that Zepto can help create “10 more large outcomes” in day-to-day consumer brands.

He also tied the company’s growth to employment. Scaling the business, in his view, creates more jobs on the ground. He said people see Y Combinator and its investors as a positive force partly because they help create jobs in India. Zepto, in his framing, is building a small part of India’s infrastructure while creating employment.

Although Zepto began before the current AI boom, Aadit Palicha said AI and machine learning are now embedded in supply chain operations, advertising, and internal work.

On the supply chain side, he described a shift from teams spending days manually forecasting supply-chain needs to machine learning systems forecasting millions of units per day without humans in the loop. The result, he said, is greater agility, faster dispatch, more throughput in the same physical spaces, and better forecast accuracy.

On the consumer and monetization side, the largest benefit he named was advertising. Zepto’s search advertising stack includes AI-led products for brands, including keyword generation and models that predict which keywords will produce the best return on ad spend. Palicha said brands increase spending when they can see that marketing dollars are being deployed more efficiently. He credited those tools with a large bump in advertising revenue.

AI has also changed internal operations. Palicha said Zepto’s corporate office can now do more with the same headcount. He said the company has cut almost all software spend to zero, reduced managed-services spend, and automated many manual tasks and inefficiencies. Walking into the Zepto office now, he said, one would see many more automated tools and almost no outsourced tools.

The engineering organization itself is substantial but, in Palicha’s view, still startup-like. He said Zepto has about 500 engineers, plus another 150 people across data science, analytics, product management, and design. The company is hiring, he said, and still sees itself as an early-stage startup rather than a large, slow company.

Running Zepto has required Aadit Palicha to learn across grocery, cold storage, trucks, warehouses, sorting machines, hardware, software, and the management of a very large workforce. Palicha again avoided the heroic version of the story. Compared with companies such as Uber, DoorDash, and Airbnb, he said, Zepto still feels small and has a long way to go.

His explanation for leveling up was the people around him. He credited YC group partners, including Jared Friedman, Tim, and Surbhi, and then emphasized Zepto’s management team: the CFO, chief operating officer, CTO, head of growth and marketing, chief business officer, and product leadership. Many had decades more experience but were still hungry and joined early because they believed in the vision.

Working shoulder to shoulder with experienced leaders forced Palicha and Vohra to learn quickly. His method was simple: ask basic questions, even at the cost of looking uninformed. He still asks the finance team how an accounting policy works and the operations team how a hardware-software system works.

That sentence captured the version of founder development Palicha offered: not innate mastery, but a willingness to look foolish, ask direct questions, absorb expertise from stronger operators, and keep raising one’s own level as the company’s complexity increases.