The most digital companies in the world are opening coffee shops

The AI industry runs on GPUs, APIs and Discord servers. So why is an AI insurance startup valued at $4 billion signing a lease for a 24/7 café in Shoreditch? Corgi, the San Francisco insurtech that raised three rounds in eight weeks this summer (TechCrunch, July 2026), already runs two 24-hour cafés in San Francisco and Atlanta. Its London location on Great Eastern Street opens this month, with five more planned including New York (Sifted, July 2026). The pitch: give founders a place to work at 3am, and sell them AI liability insurance while they sip a “Brexspresso.” Is it working? The Mercury News reported in April that the San Francisco café was running at a loss with zero conversions to the insurance business (via Wikipedia). Investors funded three more rounds anyway. That tells you something about what the market believes physical presence is worth right now. AI companies are becoming event organizers Corgi is the extreme case. The pattern is everywhere. Anthropic held its first Code with Claude conference in May 2025 as a single-day event in San Francisco. One year later, it became an international tour: San Francisco on May 6, London on May 19, Tokyo on June 10, with a second SF day added because demand from independent developers exceeded capacity (Anthropic). OpenAI’s DevDay returns to San Francisco on September 29. ElevenLabs ran its Global Hackathon across 30 cities simultaneously last December and launched its own Summit. Lovable’s community events page lists hackathons from Barcelona to Bradford to Tbilisi, funded with credits and swag. Stripe, the company that made online payments invisible, now runs two event franchises: Stripe Sessions at Moscone Center in April, plus Stripe Tour, a global one-day roadshow hitting Paris, New York and other major cities. Even the investors backing these companies have become organizers. a16z presents Tech Week, a decentralized conference series across New York, San Francisco and Los Angeles that reached more than 740 events in New York alone in 2024 (Tech:NYC). The firm also runs a16z Build, an invite-only program of private dinners and retreats designed to connect early builders. A venture firm operating a citywide event franchise and a curated dinner circuit is a firm that treats community as an asset class. These are field marketing budgets that would have gone to paid social five years ago. When every feed is flooded with AI-generated content, a room full of verified humans becomes the scarce asset. The companies building the flood know this better than anyone. The money agrees While AI companies build community from scratch, institutional capital is buying live events at scale. Liberty Media completed its €4.2 billion acquisition of MotoGP in July 2025, adding it to a Formula One Group that also includes F1 and hospitality business Quint (Liberty Media). KKR acquired Superstruct Entertainment, operator of more than 80 festivals including Sziget, Sónar and Wacken Open Air, in a deal reported at €1.3 billion (Music Business Worldwide, June 2024). CVC joined as co-investor a few months later. And Ari Emanuel raised more than $2 billion from Apollo, RedBird and the Qatar Investment Authority to launch Mari, a holding company built to buy events: the Miami Open, the Madrid Open, Frieze, Barrett-Jackson (Bloomberg, October 2025). This week, Mari agreed to acquire ATG Entertainment, owner of 70 theaters across Broadway and the West End, in a deal reported at $6 billion (Axios, August 2026). “Live has only grown more powerful,” Emanuel said in the announcement. Read that list again. Sports, festivals, art fairs, theater. The smartest money in media is converging on one thesis: attention earned in person compounds in a way digital attention no longer does. What this means for founders Here is the contradiction worth sitting with. The companies automating knowledge work are the ones investing hardest in rooms, coffee and handshakes. They understand that when intelligence becomes a commodity, trust becomes the product. And trust still gets built face to face. For startup founders, the lesson is practical. Your customers, your investors and your future hires are recalibrating where they spend their scarce in-person time. The events that win their calendar slots will be smaller, more curated and more expensive to ignore. Where the two worlds meet If you work at an AI company or a scale-up that just discovered field marketing, here is the uncomfortable truth: the event industry has been perfecting this craft for decades. The people who run Web Summit, VivaTech or MWC have solved problems you are about to encounter, from audience acquisition costs to sponsor ROI to the logistics of moving 100,000 people through a venue. That is the room Sesame Summit puts you in. It is the conference of conferences: our annual gathering in Biarritz where leaders from Europe’s top event organizers meet the startups, investors and tech companies betting on IRL. Disclosure: I organize it, so read this with that in mind. But if the smartest money in media is paying billions for audiences that show up in person, spending two days with the people who build those audiences seems like a reasonable shortcut. If your company is doubling down on events this year, what would you want to learn from the organizers who have been doing this for 20 years?

Picture this. A strategy director at a major exhibition calls with six weeks to go before the show. The brand new startup area has sold exactly one booth. The show runs on a multi-year cycle, so a failed launch means the whole concept probably gets cancelled before it gets a second chance. This is a composite of several conversations I’ve had this year, and the pattern is always the same. The organizer builds a startup area, assigns it to the existing sales team, waits, panics, then calls for help when the calendar has already decided the outcome. The diagnosis is simple: startup acquisition is a different business than exhibitor sales. Most organizers discover this too late. Here are the five reasons why. 1. They sell square meters to companies that buy outcomes A corporate exhibitor renews a booth the way it renews an insurance policy. There’s a budget line, a history, a floor plan discussion. The sales conversation is about location and dimensions. Startups have none of that. They buy pipeline, investor meetings, and proof that the show is worth their time. And their time is expensive: my rule of thumb is two full prep days for every event day, more if the team is small or the show is far. A founder deciding between your startup area and a customer roadshow is running an ROI calculation, and a rate card doesn’t answer it. A sales team trained on renewals and floor plans doesn’t speak this language. It’s nobody’s fault. It’s a different job. 2. They start the clock six months too late Startup areas usually get scoped after the main floor is sold. The launch lands a few months, sometimes a few weeks, before the show. Founders don’t work like that. They lock their event strategy two or three quarters ahead, because attending well requires prep: outreach, meeting scheduling, demo logistics, travel. A six-week sprint is competing against decisions that were made in the spring. The paradox is that organizers know this about their corporate exhibitors, who book 12 to 18 months out. Somehow the assumption becomes that startups, the most resource-constrained companies on the floor, can be converted on short notice. 3. They design the offer around what they can administer Here’s a real example, anonymized. One show’s main startup offer was a 60 percent discount, funded by a national grant. Great deal. One catch: only domestic startups qualified for it, at an international show. The offer wasn’t designed around the buyer. It was designed around available paperwork. The addressable pool shrank to a fraction of the relevant ecosystem, and everyone else got a full-price booth with no story attached. Startup offers that work are built the other way around: define which companies belong on that floor, then engineer the package (price, format, visibility, matchmaking) that makes their decision easy. Administration comes second. 4. They confuse margin kept with money made This one stings, because I’ve watched it happen twice this year. An organizer works with a partner on startup acquisition, hits targets, then decides to insource the next edition to keep the full margin. On a spreadsheet, it’s savings. In reality, the target gets missed, the area sits half-empty, and the organizer comes back mid-campaign asking for rescue. Some results are still possible at that point. The results a proper campaign would have delivered are gone. The full cost of insourcing shows up later: lost booth revenue, a weaker visitor experience in that zone, and a startup program that gets cancelled for “lack of demand.” Against that, the partner commission was the cheapest line on the P&L. 5. They run a program where they need a pipeline Startup acquisition compounds. Alumni come back. Competition applicants become exhibitors. Founders talk to each other, and a good experience at one edition sells the next one. None of that happens inside a one-off project. It requires a multi-year cadence: scouting, competitions, curated programs, follow-up between editions. Shows on two or four year cycles feel this the hardest, because a standalone approach means restarting from zero every single time, with a new team and no institutional memory. What compounding looks like JEC World, the composites industry show in Paris, is the counter-example, and yes, they’re our client, which is exactly the point. The startup work there is a bundle, built over multiple editions: a startup competition that lowers the barrier for first-time startup exhibitors, an Investor Day that brings capital to the floor and gives founders a concrete ROI reason to attend, and a startup village that gives them a curated home inside a very large show. Each piece feeds the others. Startups apply because clients & investors are there. Investors come because the startups are curated. And the ones that grow don’t disappear: they graduate into regular exhibitors. That’s the part most organizers miss. A startup exhibitor is just a first-time exhibitor. Treated well, they’re the cheapest exhibitor acquisition channel you’ll ever have. Treated as filler for a leftover corner of the floor plan, they don’t come back, and neither do the ones watching. The question for organizers If you run a show with a startup area, ask yourself one thing: is it a strategy or a floor plan decision? If the honest answer is the second one, here’s my prediction. The area launches late, gets staffed by a team hired to sell something else, underperforms, and quietly disappears from the next edition. The internal conclusion will be “startups don’t work for our show.” The real conclusion is that the approach didn’t. Startups work fine. They’re just customers who need to be sold to like startups. Disclosure: Sesamers sells startup acquisition and curation services to event organizers. JEC World is a client. Read accordingly.

The event calendar goes quiet in July and August. Deadlines don’t. This article assumes you’re an early stage founder, pre-seed or seed at most, without a marketing team. You are the events team. The next eight weeks are the only window of the year where you can work on your startup event strategy instead of running it. Here’s how to use them, roughly in order of urgency. 1. Search for calls for speakers Most Q1 and Q2 2027 conferences select their speakers in autumn, which means applications open now. SXSW PanelPicker is the obvious one and since it closes on July 26th they always lack submissions from Europe. But every major event runs some version of it, usually buried three clicks deep on their website. Before you apply anywhere, build a speaker one-pager: your topic, three talking points, a short bio, one decent photo, and links to any previous talk. Program teams review hundreds of proposals. Make theirs easy. 2. Apply to startup competitions Autumn competitions open their calls in summer: One warning from someone who reads hundreds of these applications every year: judges can tell when ChatGPT wrote your answers. Roughly 80% of the applications I review show obvious AI usage, and the low-effort ones go straight to the no pile. Use AI to structure your thinking if you want. Write the answers yourself. Sophie wrote a full breakdown of how startup competitions work from the organizer’s side. Read it before you apply. Knowing what organizers optimize for changes how you write. 3. Apply to your country’s delegation for major events CES, MWC, Web Summit: most countries send an official startup delegation, and the selection happens months ahead. For CES only: Netherlands, France, Hong Kong, etc. A quick LinkedIn search gives you tons of results.  Delegations get you a subsidized booth, press attention you’d hardly get alone, and a cohort of founders going through the same thing. The application effort is low compared to what you get. The catch is timing: CES delegations typically close applications in September. 4. Get feedback from founders who exhibited in your industry Summer is the one season when people answer cold messages. Search LinkedIn for founders who exhibited at the last edition of the trade show you’re considering using its hashtag. Ask for 15 minutes. Ask three questions: what did it cost in total, how many qualified conversations did they have, and would they do it again. Three of these calls will teach you more than most post-event reports the organizer publishes.  That’s what we learned interviewing ReSnack founders. 5. Run a pitch practice session with your peers, and moderate it Get five founder friends on a call or around a table. Everyone pitches, everyone gives feedback. You moderate. The pitching part is obvious. The moderating part is the underrated one: keeping time, asking follow-ups, managing the room. That’s a skill you’ll need on every panel you ever join, and nobody teaches it. As Lubomila Jordanova told us on the Selected podcast, small formats with harsh feedback are where you learn to hold an audience. 6. Volunteer at a startup event Unglamorous advice, and one of the best access you’ll ever get. Volunteers see how the machine works from the inside: how speakers get booked, how the VIP room operates, who actually makes decisions. You’ll meet the organizing team, and organizing teams remember people who showed up to work. An obvious one is Slush where 1,800 volunteers come together to produce one of the best startup events on earth: https://slush.org/audience/volunteers  7. Plan a side event for the back-to-office season Every ecosystem has a September event where everyone reappears. For example FDDay in Paris. Don’t compete with the main program. Host a breakfast before it opens or drinks after it closes, 20 to 30 people, one clear theme. Side events cost a fraction of a booth and put you in the host position instead of the badge-wearing position. Start planning now: venues and calendars fill up faster than you’d expect for the first week of September. 8. Budget your 2027 event strategy Nobody wants to open a spreadsheet in July. Do it anyway, because budget season at your company happens whether you participate or not. The mistake founders make is counting the ticket and the flight and stopping there. Every event day requires two preparation days: outreach before, follow-up after. That’s the 2:1 rule, and it changes the math on which events deserve a slot at all.  Pick a maximum of 5 events for 2027. Assign each one a job: sales, hiring, fundraising, or press. If an event has no job, it has no budget line. 9. Check if your summer festival has a business track A growing number of music festivals run pitching sessions or networking programs alongside the main stage. Tomorrowland even hosts a dedicated event around impact & social innovation: Love Tomorrow Summit.  Is the deal flow serious? Sometimes. Is it the most pleasant place you’ll pitch all year? Definitely. If you’re on holiday near one anyway, the marginal cost is a badge upgrade. 10. Rest The circuit restarts in September and doesn’t stop until Christmas. Slush alone will take a week out of your life, and that’s before the follow-up emails. Founders treat rest as a productivity hack, which slightly misses the point. Take actual time off. Turn off the notifications. The events will still be there in September, and so will everyone else, looking exhausted already. Don’t be them. Photo credit: Anik Labreigne on Unsplash + Gemini