Cargill Taps AI to Get More Meat Off Bone as Cattle Herd Dives

Cargill Inc. is turning to artificial intelligence to ensure it gets more beef from its processing plants as the U.S. cattle herd dwindles to the lowest level in seven decades.

The world’s largest agricultural commodities trader is using proprietary AI-assisted camera technology to provide teams at its meat plants with real-time feedback on how much meat is being left on the bone, according to CEO Brian Sikes. The technology, which is helping the company improve productivity, was first tested at a Friona, Texas, facility and is now being rolled out more broadly.

“If we can improve yields across the industries 1%, that’s over 200 million pounds of food — meat that ends up on tables and in bellies and not in rendering,” Sikes said Oct. 23 at the World Food Prize Foundation’s Borlaug Dialogue in Des Moines, Iowa. He described how workers are alerted via a system that uses red, yellow and green smiley faces “that gives them instant feedback on what’s being left on the bone that could come off.”

The move comes as cattle prices soared to unprecedented levels this year due to a severe shortage, driving up the cost of beef for consumers and squeezing profits for meatpackers, including Cargill and rivals JBS NV and Tyson Foods. Surging costs caught the attention of President Donald Trump, who this week released a plan to encourage ranching and boost domestic beef production. The administration also plans to import more beef from Argentina to cool prices.

Cargill ranks No. 21 on the TT sector list of the top agriculture and food processing carriers, while Tyson and JBS rank Nos. 1 and 12, respectively on that list. Also, Tyson ranks No. 9 on the TT Top 100 private carriers list with JBS at No. 86. Cargill is unranked.

Cargill is also using AI technology to train teams and to ensure safety in jobs that often require repetitive motion, Sikes said. Further, it is using AI in its ocean transportation business to help predict things like time and moisture levels.