How AI is cutting U.S. grocery waste by 25% with $34M in new funding

U.S. grocery stores discard approximately four million tons of food each year—primarily fresh items like strawberries, beef, and other perishables. The challenge stems from store managers struggling to predict exact inventory needs, often relying on outdated methods like printed spreadsheets and manual estimates.

Until recently, most inventory planning was done by hand. Now, AI-powered tools from Afresh, a San Francisco-based startup, are helping grocers cut food waste by up to 25%. Today, the company announced $34 million in new funding, co-led by Just Climate and High Sage Ventures, to accelerate its expansion.

From MBA project to industry solution

A decade ago, Afresh cofounders Matt Schwartz and Nathan Fenner were MBA students at Stanford. While researching food waste, they visited grocery stores and observed produce managers using rudimentary tools to estimate orders. Packaged foods were tracked digitally, but fresh produce and meat relied on guesswork and basic spreadsheets.

“It was ultimately a pen and paper process,” Schwartz recalls. Determined to solve the problem, they began developing software to more accurately predict inventory needs—a complex task given the variables involved.

Why fresh food waste is hard to track

Fresh food waste is difficult to measure due to several factors:

• Produce loses weight as it dehydrates, skewing inventory counts.
• Self-checkout systems may misclassify items, such as organic vs. non-organic apples.
• Spoiled or discarded items, from raspberries to salmon fillets, are often inaccurately logged.

Afresh’s software addresses these challenges by analyzing vast datasets from each grocer—sometimes hundreds of billions of transactions—to account for pricing, promotions, shipping origins, and perishability. Its deep learning models forecast demand using additional factors like food stamp distribution schedules and weather patterns. An optimization algorithm then recommends precise order quantities.

The system continuously refines its predictions as it learns from new data. Implementation typically starts with a pilot in 10 to 20 stores within a chain, with performance compared to a control group of non-participating stores.

“We typically see a 20% to 25% reduction in shrink when we go live with our system.”
— Matt Schwartz, Co-founder and CEO, Afresh

Afresh’s technology is now deployed across more than 12,500 grocery store departments nationwide, including locations at Safeway and Albertsons.

Beyond waste reduction: Smart store operations

Afresh’s tools offer additional benefits beyond cutting waste. For example, the software has identified that some produce displays are unnecessarily large, prompting stores to resize them or use “dummy” displays to create the illusion of abundance with less actual fruit. Grocers can also repurpose nearly spoiled produce into prepared foods, such as turning avocados into guacamole. The company recently launched a new tool to help delis forecast demand for prepared foods more accurately.

These improvements create ripple effects throughout the supply chain. Schwartz explains, “When you clean up store ordering, it makes it easier for distribution centers to buy the right amount. Then, ideally, if DCs are buying the right amount, it reduces waste at the source.”