Every industry conference has a surface conversation and a real one.
The surface conversation at Restaurant Spaces 2026 was about AI, permitting delays, and construction costs. The real conversation, the one happening between sessions, underneath polite panel language, in the moments when someone said something they clearly meant, was about whether the restaurant industry is building the right things for the right reasons.
Spoiler: some of it is. A lot of it isn’t. Here’s what cut through.
Culture & Growth
Scale is a culture problem. It just shows up as a construction problem.
The rapid expansion panel could have spent an hour on cost-per-square-foot and timeline compression. Instead, the conversation kept gravitating toward the same uncomfortable truth: when brands grow fast, the thing that breaks first isn’t the supply chain or the permitting process. It’s the part of the company that made anyone care in the first place.
Paris Baguette is opening 120 locations in North America this year. Dutch Bros is targeting north of 180. Torchy’s is pushing into new markets where the brand is effectively unknown. The brands doing it well aren’t winning on better software or faster GCs. They’re winning by refusing to let an unknown person run an opening. Torchy’s plants experienced GMs and AGMs on-site two to three months before launch and keeps them there. Dutch Bros deploys culture teams from across the country to live inside a new shop for a month before doors open.
The logic is simple and almost never followed: you cannot export what you haven’t protected. One panelist put it plainly. If you get a new market wrong, you’re probably dead there for three to five years. The capital loss is visible. The brand damage is quiet and permanent.
“Once you lose your culture, your brand is done. The numbers will fall off at some point.” – Rapid Expansion & Brand Evolution panel
Design & Infrastructure
You’re not building a store. You’re signing a 10-year lease on a decision you made today. Gregory Zamfotis, founder and CEO of Gregorys Coffee, opened his session with a story that landed harder than any statistic: he planned ahead, added extra outlets everywhere, and then launched in-house baking, which needed a 220V line on the exact opposite end of the counter. The store he had “future-proofed” wasn’t actually future-proof at all. It was just optimized for his best guess.
The data behind why this matters is not subtle. Gregorys has operated 19 stores across 19 different Simon properties, including indoor malls, outdoor kiosks, inline stores, and a shipping container, all with the same brand, same product, and same experience across every footprint. What made that possible wasn’t a bigger budget. It was a design philosophy built for adaptability from the start.
- 96% of grocery stores now have self-checkout, up from near zero in 2006
- 83% of QSR traffic is now off-premise, up from 76% in 2019 (NRA)
- 40%+ of orders arrive before the customer walks through the door
“The question isn’t whether things will change. It’s whether you’ve designed for when they do.” – Gregory Zamfotis, Founder & CEO, Gregorys Coffee
The four principles Zamfotis laid out are unglamorous but specific: design for the menu you don’t have yet; design for the channel you haven’t launched yet; design for the format you haven’t tried yet; and above all, make the cost of change low. The real ROI of modular design, he argued, isn’t in what it costs to build. It’s in the renovations you never have to do.
AI & What’s Next
The AI conversation at every conference is running about two years behind. Here’s what’s actually happening.
Ron Galloway, futurist and author, framed his session with a provocation: most of the industry is still debating whether to use AI. The threshold that actually matters crossed on January 13, 2026, when two things happened simultaneously. The OpenClaw framework launched as an open-source autonomous agent capable of controlling browsers, terminals, files, and messaging apps without human intervention. And Anthropic released Claude Cowork, bringing agentic orchestration to non-developers for the first time.
The difference between what came before and what’s here now is not incremental. Pre-ChatGPT, AI was a linear process matrix: human-driven, single-step, slow. The ChatGPT era gave us assistive compounding, where the AI accelerates drafting but humans still orchestrate, manage, and execute every step. Agentic AI is autonomous orchestration: the system plans, adapts, and completes complex workflows independently as a unified virtual team member.
“I wake up in the morning and it’s done things I didn’t ask it to do, because it figured out that’s what needed to happen next.” – Ron Galloway, futurist and author
The restaurant development applications aren’t theoretical. Galloway’s deck laid out five friction points where agentic systems are already moving: site selection agents that analyze real estate, foot traffic, and behavioral data to build performance models before any human broker engages; brand consistency agents that continuously audit every location against guidelines; permitting agents that run parallel code-compliance scenarios before municipal submission, turning the drafting, submission, weeks wait, red flag, start over cycle into a continuous real-time loop; project management agents that read daily field reports, cross-reference BIM models, and flag conflicts before physical work begins; and supply chain agents that monitor thousands of procurement line items simultaneously and adjust in real time.
His ask for operators: don’t try to overhaul the entire lifecycle at once. Deploy one verifiable tool on a single high-friction phase. Permitting is the obvious first target. Watch what happens.
Permitting
Permitting has gotten 400% slower since 1950. And the gap between good and bad operators is widening.
Andreas Rotenberg, Co-Founder and COO of Pulley, brought the data that turned a vague industry complaint into a specific, quantified crisis. Average permit approval time for a ground-up restaurant has gone from 4 months in 1950 to 18 months today, a 400% increase. At $6,000 in daily foregone revenue per location, that extra year of permitting costs a restaurant developer roughly $2.2 million per location. Multiply that across an aggressive opening pipeline and the numbers get uncomfortable fast.
- +1 yr added to average permitting timeline vs. 1950 baseline
- $6K in daily revenue lost per location while waiting to open
- $2.2M in lost revenue per year of permitting delay, per location
But Rotenberg’s more useful insight wasn’t the average. It was the variance. The top 20% of operators are getting through permitting in 12 months. The bottom 20% are taking 24. That 12-month gap is a structural competitive advantage, and it’s almost entirely within an operator’s control. His framework: permitting is rules-based (the rules are written and knowable, even if they’re multiplying), decentralized (50 states, 20,000 jurisdictions, 300,000 individual reviewers, all operating with their own incentives), and turn-based (every extra round of comments multiplies the total timeline).
“Time to approval equals round duration times rounds of review. They control the round duration. You control the rounds of review.” – Andreas Rotenberg, Co-Founder & COO, Pulley
His five rules for operators who want to move from the bottom 20% toward the top: make an informed forecast (most teams make zero effort to predict approval time); keep it to one round of comments by making a complete submission and responding to every comment thoroughly; establish a single point of ownership across the portfolio; create a repeatable process rather than bringing in one-off consultants; and build a feedback loop by reviewing comments quarterly and updating your prototype to pre-address corrections before you ever receive them. As Rotenberg put it at the close of his session: you can do all five yourself, or you can hire someone who already has. Either way, the work has to get done.
Brand Strategy
Cava‘s real competitive advantage isn’t Mediterranean food. It’s the conviction that hospitality belongs in fast casual.
Jeff Gaul, Chief Development Officer at Cava, didn’t spend much time on the numbers, though the $1.2B in revenue, 24%+ restaurant-level margins, and $3M average unit volumes are real and worth noting. He spent his time on the question underneath: why does any of that happen?
The answer traces back to Cava’s founding. Three Greek immigrants who ran a white-tablecloth restaurant and then figured out how to scale the philosophy, not just the menu. Two years ago, when Gaul joined, Cava’s physical stores had drifted from that origin. A lot of dark plywood, stainless steel, nothing that said Mediterranean. Project Soul was the correction: an architect brought in from Greece, sun-washed colors, a twelve-foot olive tree as a signature centerpiece, soft seating, upholstered chairs, and lighting that changes at dinner. Not details. A point of view.
“We always want to run the very best shift, every single time. Hospitality isn’t a department. It’s the whole operation.” – Jeff Gaul, Chief Development Officer, Cava
The Love Button is the operational version of that philosophy: every team member, on every shift, can comp a meal with no approval required. A first-time guest who looks nervous? Welcome to the table. A regular who’s clearly having a bad day? Lunch is on us. It’s a hospitality philosophy made into a daily habit, and Cava is doing it at 460 restaurants while opening 82 more this year.
The refresh program, “Sun Kist,” executed overnight over roughly three weeks to avoid revenue loss, signals something important: this isn’t just a new-location strategy. It’s a commitment to the idea that every existing guest deserves the same experience the new ones get. The older stores that don’t look and feel like current Cava aren’t just aesthetically dated. They’re brand drift made visible.
The thread connecting all of it, the culture risk at scale, the infrastructure trap, the agentic shift, Cava’s hospitality conviction, the permitting data, is the same question the best operators were asking out loud in Aventura: are we building for where our customers are now, or for where they’re going? The brands with honest, specific answers to that question are the ones worth watching.
Closing Thoughts
Restaurant Spaces revealed so much to us regarding the ongoing discussions around scaling brands, how consumers are dining, and the roles Ai will play in the immediate and near future. The real opportunities are clear – and it’s not Ai or some tech platform. Real restaurant growth is centered around finding what breaks as brands grow and scale, and finding ways to innovate, solve, and amend that breakage. The tension isn’t the tech, it’s growing a business intentionally, putting the guest experience first, while thoughtfully integrating new technology and capabilities into the greater brand vision at scale.
For the brands that are ready to scale, develop, and build experiences with intention, the market will look less like a gap, and more like a division between those who get it, and those are are on their way.
If you’re a brand ready to take that jump toward scale and experience design with intention, let’s connect.
