Reducing Food Waste with AI in Commercial Kitchens

Restaurants and campus dining halls track what they buy and what they sell, then fill the gap with guesswork. That gap—missing data on what actually gets cooked, plated and tossed—drives most food waste. Metafoodx CEO Fengmin Gong explains why he borrowed a playbook from cybersecurity’s “shift-left” movement: tackle the problem as early as possible, not after waste has hit the bin.

Gong’s team mixes computer-vision scanners and lightweight sensors that record every tray as it leaves the cooler, hits the grill or returns unsold. The result is a live ledger of consumption broken down by ingredient, recipe, hour and day of the week. Once operators can see that flow, overproduction shrinks; prep teams switch to smaller, just-in-time batches; purchasing managers order what guests actually eat, not what spreadsheets estimate.

The data also feeds staffing and menu choices. With clear peaks and lulls, kitchens can align hourly workers with real demand, swap unpopular dishes for high-turnover ones or repurpose the same chicken across several recipes to avoid spoilage. Universities—where thousands of students descend in unpredictable waves—were the first proving ground, but resorts and quick-service chains are already adopting the same dashboard.

Gong pushes back on criticism that less surplus means fewer donations. Redistribution programs often struggle with logistics anyway, he notes, and prevention beats last-minute rescue. Accurate, time-stamped inventories can still flag what’s safely edible for charities, while slicing the larger waste pile at its source.

Why would a former Palo Alto Networks co-founder move from firewalls to food? Scale and impact. U.S. kitchens discard hundreds of billions of dollars’ worth of food each year—not to mention the embedded energy and carbon. If AI can close the visibility gap, Gong argues, operators save money, staff work smarter and the planet gets a breather—all before anyone starts scraping plates.