Sweetgreen Sells Spyce to Wonder for $186.4M as Automation Strategy Reverses

I read Sweetgreen’s Q3 release twice on Wednesday afternoon — once for the headline, once for the part the headline was trying to bury. The first read was painful in the obvious ways: revenue $172.4M, down 0.6% year-over-year, same-store sales down 9.5%, a net loss of $36.1M against an adjusted EBITDA loss of $4.4M. Restaurant-level margin compressed to 13.1% from 20.1% a year ago. That’s a seven-point margin haircut in twelve months, the kind of move you usually see when a concept loses the plot.

The second read was the one that matters. Sweetgreen sold Spyce — the robotics company it acquired in 2021 and rebranded as the Infinite Kitchen — to Wonder Group for $186.4 million. The split is $100 million in cash and $86.4 million in Wonder Series C preferred stock. And Sweetgreen will keep using Infinite Kitchens via a commercial agreement with Wonder.

That is not how companies behave when an automation thesis is working.

The Spyce sale is the actual news

Let me say the contrarian part plainly: the SSS print is a six-month problem and the Spyce sale is a five-year decision. Sweetgreen bought Spyce in 2021 specifically to vertically integrate automation, and the strategic logic was always the same — own the robotics, own the IP, own the cost curve, become the only fast-casual operator who could make a bowl with zero crew labor at the make line. The Infinite Kitchen rollout was supposed to be the differentiator that pulled four-wall margins from the high teens into the mid-twenties as the system scaled.

Mark interpretation: by carving Spyce out and handing the equity to Wonder, Sweetgreen is telling the market it was the wrong company to build the robot. Wonder — Marc Lore’s vertically-integrated delivery-and-ghost-kitchen platform — has more concepts to amortize the R&D across, more capital to spend on the next-generation assembly machine, and more reason to push the technology into formats Sweetgreen will never operate. The commercial agreement keeps Sweetgreen as a customer. The strategic ownership transfers to somebody whose cost of capital and use cases make the math work.

This is the same move, structurally, that a lot of operator-tech roll-ups have been making the back half of this year: spin the hard tech to the platform that can scale it, keep a long-term offtake. A forthcoming May case study traces how the original Infinite Kitchen economics were supposed to play out at the unit level — worth reading alongside this print to see how the assumptions shifted.

The $100M cash piece matters too. Sweetgreen ended Q3 with a thinner cushion than it had in early 2024, and a $36.1M loss against an adjusted EBITDA loss of $4.4M means the GAAP-to-cash gap is wider than the operating numbers suggest. A hundred million dollars of cash dropped onto the balance sheet from a non-core asset sale is the kind of move you make when the next twelve months need a margin of safety the underlying business is not currently providing.

The 9.5% comp is a demand problem, not a weather problem

I want to be careful here, because comp deceleration in fast-casual is the kind of thing every operator wants to attribute to traffic mix, weather, calendar shifts, and consumer pullback. Sweetgreen’s release frames it broadly the same way. But strip the print apart: same-store sales down 9.5%, with traffic-and-mix down 11.7% and pricing up 2.2%. That is not a mix story. That is a guests-not-walking-in story, partially papered over by menu pricing that is itself adding to the friction.

Pricing 2.2% into a market where guests are pulling back is a defensible call when you have to protect the four-wall, but it accelerates the demand-elasticity problem in concepts where the value perception is already wobbling. Sweetgreen has been a fifteen-dollar bowl for a long time. At 11.7% traffic decline, the elasticity question stops being theoretical.

Digital is still 61.8% of revenue — that’s the part of the business that hasn’t broken. The order-ahead, app-loyalty, in-app upsell stack is doing its job, and frankly it’s the single best argument that the brand still has equity to work with. The problem is that a digital-heavy mix amplifies the labor model’s awkwardness when in-store throughput slows. The Infinite Kitchen was supposed to fix that. Now the Infinite Kitchen belongs to somebody else.

Management announced what it’s calling the “Sweet Growth Transformation Plan” on the earnings call, a phrase doing a lot of work. The substance, as I read the prepared remarks and the supplementary 10-Q filing, is three-pronged: simplify the menu, slow new-unit growth, redirect capital to remodels and digital. It is, in other words, the playbook every fast-casual concept runs when same-store sales fall off a cliff. Whether the playbook works depends on whether the brand is in a cyclical trough or a structural one.

I don’t know yet. Nobody does, on November 6, 2025, three days after the print. But I’ll tell you what I’d watch.

I’d watch Q1 traffic, because the Sweet Growth Transformation Plan needs a stabilization quarter before anybody believes the comp can recover. I’d watch what Wonder does with the Spyce IP, because if it shows up at a Wonder-branded location running double Sweetgreen’s throughput, the commercial-agreement-as-consolation-prize narrative gets uglier. And I’d watch the cash burn against that $100M cushion.

The robot got sold. The bowl still has to sell itself.

— Luca covers restaurants for TableTransfers. Tips: [email protected].