Robot kitchens.

Not yet at chain scale: the machines work, but rollouts keep getting cut. Sellers say fully automated by 2030. The tape says otherwise. Sweetgreen once planned robot kitchens in half of 40 new stores a year, now nearly half of 15, with sales falling for 2 years. Circus cut its 2026 revenue forecast by 90% in July and pushed the rollout to 2027. About 340 robot-run restaurants exist worldwide. Wonder's own deck burns $2.7 billion by 2029. Zume's pizza robots already died. The machine works. Running a chain on it is the hard part.

Because a robot hour costs about $3 and a human hour $26. The kitchen is the most routine workplace in the service economy. US restaurants pay 15.8 million people on $1.5 trillion of sales, fast food alone $331 billion on 4.1 million workers, and staff are hard to find at $15 an hour, $20 in California. A robot hour costs about $3, a human hour $26. The Infinite Kitchen machine makes 500 bowls an hour, a human 45. 10x faster, 1/8 the cost: that is why the money is moving.

Marc Lore's Wonder, $9 billion, bought Sweetgreen's Infinite Kitchen, Grubhub and Blue Apron, and cut 150 jobs in September 2026 to fund robots and an IPO. Sweetgreen runs 35 robot makelines that save over 7 points of labor cost and ring up 10% higher tickets. Chipotle and White Castle run robots. Circus SE in Munich promised the first fully robotic chain, 500 kebab robots for Mangal, up to 2,400, then cut its 2026 revenue forecast by 90% and now feeds the Ukrainian army.

I mean the industrial chains, McDonald's, KFC, Burger King and their clones on every continent: burgers dripping fat, fried chicken, nuggets, fries and buckets of sugar drinks, engineered to be craved, not to feed you. Not speedy delivery, not robot kitchens, not your local Asian kitchen or a fresh bowl of pho: much of Asia eats faster and healthier than the West. The problem is not speed. The problem is what is in the bag.