Pakistan's first YC-backed consumer startup doesn't list other people's food. It cooks its own, in kitchens with no dining room, and delivers it for less. Here is how the math works.
Order a pizza in Lahore and you might get it from a restaurant you will never find on a map. There is no sign, no dining room, no waiter folding napkins. There is a kitchen built for one purpose - to feed an app - and behind the app is a company called Byte. It sells fried chicken, biryani, pizza and wings under its own brand names, cooks all of it in-house, and charges less than the places with tables and chairs. That last part is not a coupon. It is the whole idea.
Byte was founded in 2020 by Safee Shah, who before this ran Uber's operations in Pakistan as its general manager. Ride-hailing and dark kitchens look like different businesses, but they rhyme: both are logistics problems dressed up as consumer apps, and both live or die on unit economics. Shah left one to build the other, and in the same year Byte became the first Pakistani consumer internet startup accepted into Y Combinator, joining the accelerator's Summer 2020 batch alongside the payments company Safepay.
To understand why Byte is built the way it is, start with the thing it decided not to be. Most food delivery you have used is a marketplace. A platform lists hundreds of restaurants, takes your order, sends a rider, and keeps a commission. In Pakistan that platform is largely Foodpanda, owned by Delivery Hero, with smaller players like Cheetay and Eat Mubarak around the edges. It is a good business for the platform. It is a harder business for the restaurant, which pays the commission, and often for the customer, who pays for the middle.
The catch with marketplaces is that everyone is selling the same shelf. If ten apps all list the same biryani house, none of them can make that biryani cheaper or better - they can only fight over delivery fees and discounts. Byte looked at that shelf and decided to stock its own.
There is a second problem underneath the first. In a market like Pakistan, a lot of people would order more often if the price were lower, but the restaurant plus marketplace plus delivery stack keeps the total stubbornly high. Every layer takes its margin, and the customer pays for all of them. Byte's answer is to collapse the layers into one company. When the same business writes the recipe, runs the kitchen and dispatches the rider, there is only one margin to earn, which leaves more room to lower the price and still stay in business.
Open the Byte app and you see what look like separate restaurants: Smackin Fried Chicken, Biryani Bowls, Filppin' Hot Pizza, and Just Wing It. They are all Byte. The brands are distinct storefronts, but behind them sits the same operation - the same dark kitchens, the same staff, the same supply chain. This is the quiet trick of the cloud-kitchen model. One physical operation can wear four faces on a phone screen, so a customer who wants wings and a customer who wants biryani are both, to Byte, the same delivery run from the same address.
Because those kitchens exist only to cook for delivery, they skip the expensive parts of a restaurant. No prime-location rent. No dining room to heat and clean. No front-of-house staff. What is left is food, packaging, and a rider. Strip out the storefront and the numbers change: a large Byte pizza has been listed around PKR 700, roughly $4.40, undercutting the established local and international chains it competes with.
Illustrative, not audited. The point Byte is making: costs a normal restaurant pays for a room full of tables get redirected toward cheaper food and delivery.
Byte's customers are everyday urban eaters in Lahore - people ordering dinner who care about price and speed more than about a brand name over a door. The company started in a limited set of Lahore neighborhoods, which is the sensible way to run a delivery business: a kitchen can only reach so far before the food gets cold and the rider gets expensive. Density first, map later.
The growth has been steady rather than loud. Third-party estimates put Byte's revenue run rate at roughly $4.5 million in mid-2024, climbing to about $7.4 million by the end of that year, while the team grew from around 30 people to about 50. Those are estimates, not audited filings, and worth reading as direction rather than gospel. But the direction is the interesting part: nearly doubling revenue in a year on a headcount of dozens, without a splashy funding round, is the signature of a business where the unit economics work.
Estimates from third-party startup trackers. Treat as approximate. The slope, not the decimal, is the story.
Byte is not really a software story, and Shah's background says so. He is a Fulbright Scholar with an MBA from Emory University and executive education from Harvard Business School, but the relevant line on his resume is Uber. Running a ride-hailing operation in Pakistan is a crash course in the exact problems Byte faces every night: fleets of people moving physical things across a city, priced tightly, at a speed customers can feel. Food is different cargo, but the discipline is the same.
That operational bent shows in how Byte launched - almost invisibly. For a while it had no public website at all, existing only as apps on the iOS and Android stores. A consumer company that hides its front door is unusual, but it fits a team that treats the kitchen and the delivery as the product and the marketing as a later problem.
It also fits the moment Byte was born into. In 2020, Pakistan's startup scene was starting to draw real attention from global investors, and being the first consumer internet company from the country to walk into Y Combinator was as much a signal to that ecosystem as it was a milestone for one company. The $150,000 that came with the batch was small money by Silicon Valley standards, but the stamp mattered: it told other founders in Lahore and Karachi that an accelerator known for software would back an operations-heavy, real-world business if the economics were sound. Byte spent the years since trying to prove the second half of that sentence.
Byte plays in two overlapping arenas. Against the marketplaces - Foodpanda and the rest - it competes on price and control, since it can undercut a listed restaurant that has to pay both rent and commission. Against a growing set of Pakistani cloud-kitchen operators like Pandakitchens, Hotpod, Eat Cloud and Lettus Kitchens, it competes as a peer, and the question there is who builds the best brands and the tightest operation. Byte's answer is vertical integration: own more of the stack than anyone else, and let that ownership show up as a lower price on the checkout screen.
There are real risks in the model too, and they are the mirror image of its strengths. Owning the kitchens means owning the fixed costs, the staffing, the food waste and the bad nights - a marketplace can add a thousand restaurants with a form; Byte has to build each one. The brands have to earn repeat orders on taste, not just price, or the whole efficiency argument collapses into a race to the bottom. And a delivery business that lives on density has to expand carefully, kitchen by kitchen, city by city. None of that is unique to Byte, but it is the homework that comes with choosing to be the restaurant instead of the listing.
The lesson worth taking from Byte is not "start a food company." It is narrower and more useful: when an entire industry has settled on being a marketplace, there is often room for someone to own the supply instead of the interface. Byte did that with dinner in Lahore, turned $150,000 of accelerator money into a real business, and proved that a "tech startup" can look a lot like a kitchen.