Marc Lore’s “infinite bowl” makes 500 an hour. Humans top out at 45

Marc Lore is pitching an “infinite bowl-making machine” and backing it with a simple comparison: Wonder’s hardware can make 500 salads, Tex-Mex, and poke bowls in an hour, while Lore says a human worker is capped at roughly 30 to 45 bowls an hour. He didn’t hide the punchline either. Speaking at the 25th annual Fortune Brainstorm Tech conference in Aspen on Tuesday, the Wonder founder and CEO described a system that builds a bowl to the specs from an online delivery app order, down to the personalized macros people are tracking.

The core claim is operational. Wonder’s machine spins each bowl on a turntable while ingredients drop into place based on that order. Lore says the resulting bowls have “no errors,” meaning customers get exactly what they ordered, at the throughput that makes staffing look radically different from traditional fast-casual. He also points to the tech’s rollout: Sweetgreen, which sold the “infinite bowl” technology to Wonder (an acquisition Wonder made), already runs it across 32 locations, and it will land in its first Wonder kitchen next month.

This is not just a gimmick about robots making lunch. It is an attempt to turn food assembly into a predictable, repeatable manufacturing process where the bottleneck shifts away from labor and toward inputs, logistics, and kitchen throughput. Lore frames Wonder as a “vertically integrated food platform” that owns 26 restaurant brands, including a Bobby Flay steakhouse, and includes multiple delivery options. He says Wonder also owns and manages the kitchens and handles delivery after buying GrubHub in a deal valued at $650 million that closed in 2025.

That vertical integration matters for one big reason: it changes where margin has to survive. Lore argues that combining all the different brands in a single kitchen creates a single profit pool. In his telling, prices can be less expensive because the margins do not need to support both the restaurants and the delivery companies separately. It is a direct counter to the common fast-casual scaling challenge: building locations that can support themselves even when delivery demand, third-party fees, and staffing costs fluctuate.

Wonder’s pitch also leans into geography. Lore says by combining multiple brands into one kitchen, Wonder can serve areas and regions that do not have the population numbers to support larger fast-casual chains like Chipotle or Cava. In other words, the machine-driven kitchen is supposed to make unit economics more tolerant of smaller markets. The operational logic is straightforward. If a kitchen can run 26 restaurants with far fewer late-night staff, the model can potentially keep locations open longer and still make sense.

Lore gave a concrete staffing example tied to those late hours. He says they can stay open until 2 a.m. in the suburbs because they can operate all 26 restaurants with three people late night. In his description, one human staffer answers the hotline, another handles finishing the dishes, and the third works the handoff to delivery drivers. That is a dramatic reframing of the “restaurant job” ecosystem. Instead of a labor model built around line cooks and assembly labor scaling with demand, Wonder’s model suggests more of the workflow can be handled by automation, with human time focused on exception handling, finishing work, and delivery handoffs.

The product roadmap extends the automation beyond bowls. Lore says an “infinite sauce machine” can spin up 500 sauces an hour from 152 raw ingredients. He also says an “infinite beverage machine” is slated for next year. The implication is that Wonder wants the kitchen to behave more like a configurable production line than a set of bespoke stations. And if you connect that to his broader ambitions, it points to something bigger than throughput. Lore wants Wonder to have an “indelible impact” on the public company landscape and is pursuing an IPO.

He told “The Aisle” founder Jason Del Rey that he expects to be ready to go public early next year, although he cautions that the market will likely dictate the timing of any potential public offering. That matters because an IPO investor’s question is not “can robots make bowls,” it is “does the model scale faster than competitors’ unit-economics decay.” Lore’s answer is a mixture of automation speed, error-free assembly, vertically integrated distribution, and a new way to create concepts.

Wonder is also building a tool he calls Wonder Create. Through a feature he described, anyone can describe a concept in an AI prompt like, “build me a fast-casual Mexican restaurant for Gen Z, for people that love cycling.” Lore says Wonder will then output a branded restaurant concept with its own name, menu, pricing, photos, and nutrition information built on Wonder’s automation in about two minutes. He says users can push their concepts live for $10 a month and calls it “Think Shopify on steroids.” In practice, this would mean the company can test and roll out new offerings without the traditional lead time and overhead of launching a restaurant the old way.

For executives watching from adjacent sectors, the second-order effect is that this is a convergence bet: automation, content and concept generation, and delivery economics pulling into one profit pool. If Wonder can actually sustain “no errors” at 500 bowls per hour while controlling delivery costs, the competitive pressure shifts to operators who still treat staffing and throughput as the main levers. Meanwhile, investors and boards evaluating IPO-bound companies may find themselves underwriting a different risk profile: less about finding enough hourly workers, more about scaling kitchen hardware, supply inputs, and the integrated delivery engine. Lore’s whole thesis is that AI can’t disrupt Wonder’s moat because the moat is operational and system-level, not just marketing. Whether the market rewards that thesis is the open question, but the direction is clear: food as a platform, kitchens as factories, and labor as the variable you try to shrink.