Aakar Khanna and Arjun Chaliha (@arjunjchaliha) launched Marble on Thursday, pitching an AI operating system that handles inventory, purchasing, invoices, food preparation and labor planning for restaurants.
Chaliha introduced Marble as an "autonomous back of house" in a July 30th post on X. The New York company, previously called Truffle, is part of Y Combinator's Summer 2026 batch and has two employees: Khanna, its CEO, and Chaliha, its CTO.
The founders met as University of Michigan freshmen nearly a decade ago. Khanna grew up managing his family's multi-unit restaurant business before covering restaurants and food and beverage as a Goldman Sachs investment banking analyst. He later worked in business operations at trucking-finance company AtoB and led strategy and operations at healthcare startup Kouper.
Chaliha studied computer science at Michigan and later led AI and machine-learning projects for Bloomberg's electronic trading operation. His first founder project began in college, when he helped build Peer Minutes, a marketplace connecting students with peer tutors at per-minute rates. The project competed in the university's Michigan Business Challenge.
How Marble works
Marble starts with the phone already in a restaurant manager's pocket. Operators can use its mobile application to photograph stock or count items aloud. Marble says its computer-vision and voice systems record inventory 90% faster than manual counts, with 99% accuracy.
The resulting inventory data feeds machine-learning models that Marble says forecast item-level demand with 97% accuracy. Software agents then use those forecasts to place orders, generate preparation sheets, draft employee schedules, track waste, reconcile invoices and analyze menu economics. The models can also consider reservations, weather, holidays, local events and news when estimating demand.
Those accuracy and performance figures are company-reported. They are central to Marble's sales case because restaurant software has to convert messy, location-specific information into orders and schedules that managers can trust during daily service.
Marble says it is operating across multi-unit restaurant groups in the United States and Canada. Its Y Combinator launch page lists A&W Restaurants, The Kati Roll Company, Sophie's Cuban, Kokomo NYC, HolyShred and Dead Letter No. 9 among its customers.
According to Marble, customers have reduced food waste by 40%, eliminated stockouts and saved managers more than 30 labor hours each week. Marble has not tied those results to specific customers in its public launch materials, so the figures remain Marble's account of early deployments.
A crowded back office
The product is entering a restaurant-software category with established vendors and a growing number of AI-focused challengers. MarketMan sells inventory, purchasing and recipe-costing software starting at $199 per month. Stockcount offers voice inventory, invoice scanning and an AI assistant starting at $19 per month. MarginEdge combines inventory, invoice processing, ordering and cost reporting, while Crunchtime covers inventory, scheduling, kitchen management and operational tasks for large restaurant groups.
Marble is making a broader product bet. Instead of selling inventory counting as an isolated time-saving feature, Khanna and Chaliha want one system to capture restaurant data, forecast demand and execute the work that follows. That scope gives Marble several entry points into an account, but it also puts the two-person company against vendors already connected to point-of-sale systems, accounting packages and suppliers.
The economic pitch is direct. The National Restaurant Association estimates that food and labor each consume about one-third of a typical restaurant's sales, leaving a pre-tax profit margin of roughly 5%. The association estimates that average restaurant expenses rose 36% between 2019 and 2026. During the first half of 2026, 33% of operators surveyed said their restaurants were unprofitable.
That pressure makes back-office automation easier to justify, provided the savings survive daily restaurant conditions. Marble's test will be whether its agents can handle differences among menus, suppliers, storage layouts and staffing rules without adding another system for managers to maintain.
YC backs a vertical AI play
Y Combinator lists Marble as an active Summer 2026 company, with partner Brad Flora assigned as its primary partner. Khanna also thanked Flora and YC partner Christina Gilbert when he announced the acceptance in June.
Under Y Combinator's standard deal, accepted companies receive $500,000. YC invests $125,000 for 7% and another $375,000 through an uncapped safe carrying a most-favored-nation provision.
The investment gives Khanna and Chaliha room to prove that multimodal AI can structure the physical details that restaurant software has historically relied on employees to enter. Marble's advantage will come from the operational data accumulated after each inventory count, correction and order. Its customers will decide whether that learning loop is useful enough to replace the clipboards and point tools already running the kitchen.