Sweetgreen Sold Its Robotics Company But Kept the Robots

The Automation Stays, the Company Goes

Sweetgreen spent years building one of restaurant automation's most ambitious systems. In December 2025, it sold the robotics company behind that technology to Wonder for $186.4 million—then kept right on deploying the robots.

The transaction offers a revealing look at how restaurant automation may mature. Sweetgreen acquired Spyce in 2021 for approximately $70 million and developed its Infinite Kitchen technology across a growing number of locations. By mid-2026, the system operated in 35 of 287 restaurants, with roughly half of the chain's 13 planned new locations for fiscal 2026 expected to include it.

But Sweetgreen decided it didn't need to employ the engineers, fund the development organization, and operate an automation company to benefit from the automation itself. Under the Wonder deal, Sweetgreen received $100 million cash, $86.4 million in Wonder preferred stock, and a perpetual, royalty-free license to use Infinite Kitchen in its branded facilities. Wonder now supplies new units, handles commissioning, provides support and maintenance, and bears the cost of continued platform development.

Sweetgreen became the customer instead of the manufacturer—and the robots kept working.

Why It Matters

This model could reshape how operators approach automation investment. Rather than choosing between building in-house or buying off-the-shelf, companies can own technology during its immature phase to control development, then transfer it to specialists once it proves viable. That shifts ongoing R&D costs and operational risk while preserving access to the automation. For an industry where most operators lack robotics expertise, it's a practical path to advanced technology without permanent infrastructure overhead.

Walmart Ran the Same Play

The pattern isn't unique to restaurants. Walmart worked with Alert Innovation on automated grocery fulfillment, acquired the company in 2022, then sold the operation to Symbotic in January 2025 for up to $550 million. Walmart remained an anchor customer, selecting Symbotic to develop next-generation systems in a deal Symbotic valued above $5 billion.

Both cases follow a similar arc: partner with an automation company, acquire it to accelerate development inside real operations, prove the technology works at scale, then spin out the automation business while retaining deployment rights. Own it when ownership accelerates innovation. Become the customer when specialization delivers more value.

Automation Becomes Infrastructure

The shift reflects automation's evolution from novelty to operational necessity. Early restaurant robots attracted attention because customers could see them—arms flipping burgers, machines making coffee behind glass. The next generation increasingly disappears into workflow.

Orders flow from apps directly into production systems. Equipment receives instructions, portions ingredients, controls cooking processes, and coordinates assembly. No humanoid required. What matters is how much of the workflow can be automated and how effectively those pieces communicate with point-of-sale systems, kitchen displays, and delivery platforms.

Counting robots becomes a poor proxy for measuring automation. A restaurant can automate meaningful operations without installing anything customers would recognize as a robot. Digital ordering, automated dispensing, sensor-driven cooking, software-coordinated production—each reduces labor, time, or error. The larger gains come when those technologies connect into integrated systems.

The Tradeoffs Are Real

Sweetgreen's deal involves clear dependencies. The company now relies on Wonder for supply, support, maintenance, and upgrades of technology increasingly central to operations. Wonder may also supply or license Infinite Kitchen to third parties, including potential Sweetgreen competitors, subject to contractual restrictions.

Sweetgreen traded exclusive ownership for capital, reduced development burden, and access to specialized automation expertise. For many operators, that may be the right exchange. Running restaurants and operating robotics companies require fundamentally different capabilities.

The question for other industries investing heavily in robotics—automotive manufacturers building humanoids, for example—is whether ownership remains strategically necessary once the technology matures. Hyundai owns Boston Dynamics and plans Atlas deployment in its factories by 2028. But if humanoids eventually become standardized industrial equipment, automakers might follow the Sweetgreen-Walmart path: incubate the technology, prove it works, then rely on specialized suppliers.

Integration, not novelty, becomes the competitive advantage. The winners may be companies whose automation disappears into workflow and executes thousands of transactions reliably—not those with the most impressive robot demonstrations.

These details were first reported by Automation Watch.