Sweetgreen CEO Jonathan Neman on 18 years building healthy fast food: automation, seed oils, and the real estate game

Summary

Sweetgreen, founded in 2007 by Jonathan Neman and two Georgetown co-founders, has spent 18 years building toward a model that now sits at nearly 300 company-owned locations with roughly a quarter of revenue concentrated in New York's ~50-restaurant cluster. The company raised its initial $300,000 from 50 investors averaging $5,000 each, structured at the individual restaurant level before being rolled up after the third location. The first VC check came from Steve Case's Revolution, anchored partly on a technology thesis that put Sweetgreen among the earliest restaurant chains to offer mobile ordering.

Automation and the Infinite Kitchen

Sweetgreen's automation push, branded the Infinite Kitchen, can assemble 500 bowls per hour with consistent portioning. The underlying technology came via an acquisition made roughly four years ago of Spyce Kitchen, an MIT-founded startup that had built capable robotics but couldn't find buyers. Sweetgreen commercialized the platform and now runs the Infinite Kitchen in 30 restaurants. Ten days before this recording, it sold Spyce to Wonder, Mark Lore's company, for approximately $186 million, retaining full access to the technology and the right to benefit from future economies of scale as Wonder expands machine production.

The economics driving automation are straightforward: minimum wage in West Hollywood sits at $22/hour, and Sweetgreen pays $24–$25/hour across parts of LA. Beyond labor cost, the Infinite Kitchen allows the chain to serve twice as many customers per hour versus manual assembly. Neman is explicit that the technology doesn't fully replace human involvement — a finishing station handles hand-mixing and plating to preserve a craft feel. As of November 18, 2025, the company opened its first drive-through location featuring an Infinite Kitchen, combining automated speed with the drive-through format.

Why restaurant automation has largely failed as a standalone business comes down to three structural problems Neman identifies: workflows are highly restaurant-specific, most chains are franchised so capex decisions sit with franchisees who have little incentive to invest, and the capital intensity of the hardware must be justified by sufficient labor displacement. Sweetgreen's fully company-owned model removes the franchisee alignment problem entirely.

Seed Oils and Reading the Consumer Curve Early

Sweetgreen eliminated seed oils approximately two years ago, pre-dating the RFK-driven national conversation on the topic. The decision was pushed by a small but vocal segment of wellness-oriented customers in LA and New York, not by broad survey data — Neman is pointed that surveys optimize for the middle of the bell curve and are a poor tool for identifying durable trend signals. His CFO reportedly had no idea what seed oils were when the proposal was raised internally.

The strategic logic is that lasting health trends share a specific property: they are simultaneously healthier and better-tasting. Neman contrasts this with gluten-free bread, nut milks, or sugar-free substitutes — all health-motivated trade-downs on taste. Removing seed oils in favor of olive oil meets both criteria, which is why he views it as structurally different from flash-in-pan diet trends like paleo.

Real Estate, Unit Economics, and Scaling Logic

Restaurants are fundamentally a real estate and people business, not a food business — a distinction Neman draws explicitly. Sweetgreen's site selection model blends psychographic and demographic data, mobile movement data, proximity to gyms, and street-side factors like sun exposure, with mandatory human walk-throughs to assess brand fit. Early New York expansion deliberately targeted Nolita and Williamsburg over Midtown to signal brand identity.

New restaurant construction now costs $1.3–$1.4 million per location before Infinite Kitchen hardware is added. For operators evaluating restaurant concepts, Neman identifies two metrics that matter most: unit economics (build cost versus payback period) and addressable store count. He flags Columbus, Ohio as the standard test market for scalable fast-casual concepts — more representative of national consumer behavior than New York, which skews the data.

On delivery, Sweetgreen uses DoorDash as a white-label logistics provider powering native delivery inside its own app, while also participating in DoorDash's marketplace. Menu prices on third-party platforms run roughly 20% higher than on the Sweetgreen app directly. The dynamic mirrors hotel-OTA relationships: distribution reach comes at the cost of customer data ownership and direct marketing access.

On the frontier, Sweetgreen is an early pilot partner with Zipline for drone delivery to suburban markets, where backyard drop zones make the logistics viable. Ground-based autonomous delivery via robots like Starship remains a longer-term proposition given density and access challenges in urban environments.