I meet Will O’Brien at a party in San Francisco. He’s wearing a two-piece suit. Sixty seconds into the conversation, when he finds out that I’m Catholic, he throws open his suit jacket to reveal the shirt underneath: white linen emblazoned belt-to-collar with the Virgin Mary.

Will is also carrying around at least a dozen small plastic rosaries wrapped in clear plastic packaging, the kind of thing you’d hand out at a school fair. Just moments after he unveils his trademark Vatican attire, he is unsheathing one of these rosaries, and is jovially forcing it over the head of my non-denominational friend. A baptism, of sorts. Welcome to San Francisco.

At sixty seconds into our interaction, Will showed me Our Lady of Guadalupe. At ninety seconds, he rosary’d my friend. Now, at five minutes, he’s pulling out his phone to show me a list of his favorite conspiracy theories. UFOs are big. So is the JFK assassination, and telepathy. He also tells me about St. Joseph of Cupertino, a levitating saint local to the San Francisco Bay. TradCath though I claim to be, this was the first I was hearing of the Flying Friar.

Halfway through the list, Will is loudly summoned across the room by another group. This dynamic continues throughout the night. Everyone, it seems, wants to be in the radius of this ebullient Irish cowboy. This is, to me, unsurprising. Will possessed in abundance the thing that San Francisco tech people so desperately lack: the ability to have a good time.

Subscribe now

Robots, underwater

Will was not the sort of person you’d expect to meet at an event like this. Will broke into that dimly-lit room of San Francisco denizens with the explosive energy of Elon’s Southaven datacenter or a benevolent, orange-mustachio’d Genghis Khan entering the battlefield. His huge smile and raucous laugh cut through the introverted, self-conscious comportment and flat, transactional parlance of the SF Tech Guy like the Point Bonita lighthouse cuts through the fog rolling over the Golden Gate.

Will had arrived in San Francisco just a few months earlier, at the helm of the fledgeling startup, Ulysses. Ulysses, the company, like Will O’Brien himself, is an outlier in the SF tech scene. The startup world can easily seem like a monoculture: hordes of bit-based enterprises battling ruthlessly to differentiate. Ulysses, however, doesn’t have to try that hard to seem different. In a world where everything is digital, Ulysses is refreshingly material.

The technical term for what Ulysses builds is “autonomous vehicles” for “scalable operations on the surface and in the subsea.” The nontechnical term: ocean robots.

Will, with his big laugh, his embroidered shirt, his cowboy boots, and his list of conspiracy theories, seems almost like a caricature of himself; he is entirely larger than life. But don’t be fooled by his gregarious demeanor: Will O’Brien is building something serious.

The first mission of Ulysses’ fleet of autonomous vehicles is remarkably noble. There is a specific type of seagrass which grows on the coasts and is astonishingly effective at carbon fixation: they store carbon up to 35 times faster than a tropical rainforest. In yet another entry in the endless litany of enviro-pocalypse nightmare statistics, these precious grasses have been dying at a rate of 7% per year, or two football fields every hour. Replanting these grasses by hand via human divers is about as effective as trying to cool your data center with a box fan. For the Ulysses fleet, it’s a piece of cake. One can imagine a sort of Pixar scenario: schools of wide-eyed little ocean robots venturing to the depths of the California coast, befriending the woebegotten seagrasses and their brightly-hued flora and fauna companions, the entire entourage bopping along to a soundtrack composed by Pharrell Williams.

But Ulysses vehicles are not merely Thunberg apparatuses: they’re a bit less Nemo, a bit more shark. A Ulysses autonomous vehicle looks a bit like some kind of compact, oddly-symmetric onyx jet plane: a scaled-down version of the ones in Avatar. Although they’ve started off with ecological restoration, they are swiftly branching out to use the vehicles for commercial and defense purposes. They are, their website states, “building the operating system for the ocean.”

The ocean is, according to Ulysses, humanity’s final frontier. It’s easy to forget that the ocean covers 71% of the earth’s surface. Over 80% of it remains completely unexplored. The deepest parts of the ocean — known as “Hadal depths” — are more difficult to access than the moon. This frontier has largely been ignored by the tech world, busy mining for crypto and building new electronic juicers. Ulysses intends to change that.

Since launching Ulysses in 2023, O’Brien and his three co-founders have scaled from five employees to eighteen. They’ve moved their enterprise from Dublin to San Francisco. They’ve expanded from ecological restoration to commercial and defense applications. This year, they’ll launch their biggest robot ever: Leviathan.

The lair of the ocean cowboys

I visited Ulysses HQ in the late afternoon on a Wednesday. From the outside, tucked away on a side street in SoMa with a sleek but nondescript gray exterior, Ulysses could have been one of any number of the hordes of AI-adjacent startups springing up in the fecund marshes of the mid-2020s Silicon Valley VC bonanza. But as soon as you step through the door, you know something is different: you’ve found yourself in the lair of the ocean cowboys.

The first thing you notice is the twelve-foot shark suspended via wire from the ceiling. The Ulysses headquarters is graced by not one, but two of these massive ceiling sharks, suspended in the air like aquatic zeppelins. There is also an entire school of fish swimming along the ceiling between the rows of desks and the primary meeting room. “We hung those ourselves,” Will tells me proudly.

There’s also a portrait of Steve Irwin, the great Australian zookeeper and conservationist, a cartouche with the official Ulysses motto (“Frontier Spirit”), and a list of “Cowboy Commandments” (“Honor yer Ma & Pa”, “Don’t be hankerin’ for yer buddy’s stuff”). On the lower level of the Ulysses Batcave is a swimming pool, next to which is a freestanding crane for lowering the ocean-robots into and out of the water. A twelve-foot American flag hangs from the far wall.

On my way over, Will texted me:

Ill need to get a cake in Whole Foods rn. So if you arrive early I’ll Be back in like 15!

A koan, of sorts.

By the time I arrived, the birthday cake had been devoured, save for a single, chocolatey slice. Nothing like a good day of robot-building to whet a man’s appetite.

When I met Will at the party last year, I saw right away that he was a sort of social magnet: drawing all the assorted iron filings around him into orbit. This is also the role he plays in his company. He’s the mascot, the spokesperson, the visionary. He’s the guy who gets everyone else onboard. Once a month, Ulysses hosts a pizza night for all of its engineers. You’d better believe that Will is there in the thick of it, apron at the ready, wrangling gallon-jugs of tomato sauce while laughing his uproarious laugh. At the end of last year, Ulysses hosted a rodeo, complete with bull-riding. Except the bull was — you guessed it — a mechanical shark.

Share

The new aquatic frontier

The American Frontier was officially declared to be closed in 1860. This event, according to various theorists, instigated a spiritual crisis from which we have yet to recover. America was founded upon the principle of expanse, of conquest, of the promise of the wide horizon. Since 1860, when this firehose of energy could no longer seek an outlet in the West, it has been forced to seek out other avenues: fighting the two European wars, policing the 20th century international political order, conquering the digital frontier.

New York Times columnist and podcast host, Ross Douthat, has written extensively on the phenomenon of American decadence. Douthat’s thesis is that, since the 1950s, the nation’s best and brightest minds have turned away from the material world and towards a dream-sequence sidequest of building out the infrastructure of the digital simulacrum. This simulacrum has become fantastically powerful and complex. Meanwhile, cities get uglier, highspeed rails go unbuilt, and even Elon dares not venture into the market for the flying car. “We wanted flying cars,” the proverbial Thiel quip goes, “instead we got 140 characters.”

Different tech oligarchs have different theories regarding what, exactly, the “next frontier” will be. Elon and Bezos think it’s space. The LLM-lords think it’s AGI. Brian Johnson thinks it’s the pursuit of immortality. Zuck thinks it’s the 3D rendering of the real world that we’ll all strap into via headset.

Most of these new frontiers exist beyond us; either above us in space, or locked in the recesses of an obscure new algorithm. But Ulysses thinks the new frontier is right beneath our feet, in the hundreds of millions of square miles of uncharted territory covering the surface of our planet. And now, their team is moving fast, building the technology that has the potential to unlock a whole new domain of human activity.

It’s funny. Ulysses, although founded in Dublin, with its frontier spirit, its Cowboy Commandments, its shark rodeo and gargantuan American flag emblazoned on the wall, might just be the most American company I know. And although he’s a consummate patriot, Will O’Brien, who arrived from Dublin just over a year ago, might be the most stereotypical instantiation of the American spirit I’ve ever encountered.

Which brings us to a broader question about culture. There’s a strange phenomenon where sometimes the most enduring cultural artefacts end up being created by people outside of the culture itself: the Scottish kilt invented by an Englishman, the modern pizza perfected in New York City, the White House designed by an Irishman, and the iconic horned viking helmets created by a German costume designer for a production of Wagner’s Der Ring des Nibelungen. There is a sort of cultural distillation, a consolidation, perhaps, that happens when the iconography of a culture is co-opted by someone from the outside: it can become, in some cases, more fully chosen, fully inhabited, fully adopted, than when one is born into it. To those of us inside a culture, it’s like water to a fish. Sometimes it takes someone on the outside to see it clearly.

But perhaps, even in the strictest sense, Ulysses is as American as you can get. Ulysses and its founders are immigrants, like most of us were at one point or another in this country of transplants. Like most of our grandparents did at some point, Ulysses and its founders have come to seek the next great frontier, and are attacking it with energy and enthusiasm.

When I met Will at the party last year, he was bursting with explosive energy. Today, he’s a bit quieter, more subdued. Coasting between the Whole Foods cake run and yet another late night of work. He’s easy to talk to, easy to be around. Will, unlike some young founders who have rapidly attained his degree of success, seems to feel no need to project an aura of righteous self-importance or overbearing potentiality. He’s just a nice guy. With a lot of sharks.

What is it to be American, anyway? Well, as anyone can attest, hardly one thing in particular. And yet, in this uprooted, cultureless, consumerist void of ours, which philosophers like Simone Weil have blamed for spreading the disease of hyperinstrumentalization throughout the entire globe, there is also a sort of bigness of spirit, an eyes-to-the-horizon sort of attitude that drew people here in the first place.

California is, at its worst, the vacuum of the SF cultural scene, the ugliness of the AI-billboard bonanza as you drive 101-South towards SFO, the Pandora’s box of attention-bending instruments our technologists-in-chief have unleashed upon the globe. But it is also the California of the bluffs and the bay, the sheer scale and expanse of the Pacific, a pool of relative anarchy and meritocracy in contrast to the fixed social orders of the Old Worlds back East. The poet Robinson Jeffers once wrote that California is the “end of civilization.” But it is also, at the same time, the beginning of a new history: the history of this strange new technological epoch whose features are only beginning to take shape.

Will O’Brien represents, I think, the legacy of this latter half of things: a sort of 21st century transplant from a bygone era, someone who is willing to take up the mantle of destiny and shape the world in the way he sees fit. California, for all its dysfunction, is a place where things still happen. And that, I think, is a pretty good thing.

It’s 11am in the West Village.

Kiki Couchman is walking down the street, hauling a bright-blue backpack full of yogurt.

She’ll stop at an office, then at a grocery store, then a cafe, to drop off anywhere from 25 to 100 cups of yogurt. On any given day, she’ll make a dozen of these stops, while capturing 0.5x zoomed microclips for the day-in-the-life vlogs that she films from the back of a CitiBike.

Kiki has self-described “oldest daughter syndrome,” which means she’s always done the “right thing.” Until Sourmilk. When she quit her job in private equity a year ago, Kiki went from being a respectable Stanford grad hustling her way up the corporate ladder to an early-stage probiotic guru and up-and-coming influencer, whose signature accessory is a giant blue backpack full of yogurt.

Kiki shares a name with the heroine of the iconic 1999 Miyazaki film, Kiki’s Delivery Service. In the Miyazaki film, Kiki is a young witch who starts a courier service, flying all over town on her broomstick with a black cat in tow. And her dress is the exact color blue as the package of a Sourmilk yogurt cup. The real-life Kiki is a startup founder, not a magical courier. But they both spend their days delivering things.

Trust your gut

The yogurt thing happened sort of by accident.

Kiki and her cofounder, Elan Halpern, were best friends at Stanford. Kiki studied human biology, and Elan was a D1 athlete, so both of them were serious about taking care of their bodies. In April of 2024, they both independently sensed an impending career transition, and started chatting about what their next steps might be. “And we both realized we wanted to be in the health space,” said Kiki.

Kiki and Elan were living together as roommates when Elan started her fateful sideproject: making yogurt to heal her gut. Homemade yogurt, made via an extended fermentation process, contains more probiotics than normal, store-bought yogurt. It also has less lactose, which makes it easier to digest. Plus, it won’t have any sugars, thickeners, or preservatives messing with those beneficial microbes. Elan was mixing up a veritable gut-boosting ambrosia, the kind that would make an Andrew Huberman or a Gwenyth Paltrow weak in the knees, right there in their kitchen. But at that point, it was just for herself.

One morning, Kiki was doing a puzzle alone on her coffee table. Elan came in and said, all of a sudden, “Kiki, I have this idea. I think we need to scale yogurt.”

What if you could put the thing that heals you in all of your food?

And that’s the Sourmilk thesis.

Day in the life

Kiki starts her day early.

At 7:30, she’s at the gym. By 9:30, she’s trekking over to Long Island City to pick up the day’s supply of yogurt from their warehouse. From 10:30 to 12:30 she’s delivering yogurt to offices, cafes, and workout studios. Then from 12:00 to 3:00, she’s delivering to grocery stores, zigzagging back and forth across Manhattan on her CitiBike.

At 3:00, she comes home, and spends two hours doing admin. Then from 6:00 to 8:00 she bops over to some kind of Sourmilk evening event, a trendy pilates popup or a bustling holiday market. Then at 8:00, she comes home and makes a video. Rinse and repeat.

In the early stages of Sourmilk, Kiki was constantly obsessing about scale. Before she would make a decision, she would ask herself — is this decision scaleable? Will this process still work if we have 10x more customers? 100x?

It was at this point that a founder friend staged an intervention. “Don’t even think about scale,” he said. For now, focus on what’s in front of you. “You will not have any customers tomorrow if you don’t make the decision that is best for those customers right now. So don’t even think about it.” “The scalable decision…that will come. You will face that mountain when it gets here. But the mountain in front of you is those ten customers right now.”

So Kiki and Elan threw scale to the wind, and adopted what they’ve called the “drug deal model” of yogurt distribution (which just as easily could have been called Kiki’s Delivery Service). No outsourcing, no modular logistics chain, no robot couriers: just Kiki on her CitiBike, Manhattan’s very finest probiotic plug.

Share

When the milk goes sour

I asked Kiki what the hardest day for the company was.

She told me about a time earlier this year when she and Elan traveled five hours upstate to collect a giant round of yogurt. Each round takes three days to complete and produces several thousand cups of yogurt.

When they’d finally arrived all the way upstate, and tasted the yogurt, something just wasn’t quite right. The formula was all set, but something had been tweaked ever-so-slightly in the production process. The result? Yogurt that tasted… just a little bit wonky. Or to use Kiki’s words: “it was awful.”

Kiki and Elan faced a tradeoff. Do they drive the thousands of cups of wonky yogurt back to the city, pay the money to store it, and watch the light temporarily depart their customers’ eyes as they taste the yogurt that has embraced, perhaps too literally, the name of the brand?

Or do they eat it?

In the end, they threw away the yogurt. All 3000 wonky cups of it. They hauled all that yogurt to the dumpster and they said sayonara. “It was ridiculous,” said Kiki. “But once we committed and decided, we could be funny about it. It could be a whole bit.”

It was a huge loss to throw away all that food. Kiki even had to strategize about how she would break the news to her followers; just in case the Food Waste people decided to get involved. But it would, in the long run, do more harm to the brand to distribute a product that wasn’t meeting their quality bar. So that’s why, when SourmilkPocalypse hit, Kiki and Elan thought about the customers, first. Goodbye, wonky yogurt.

Playing on hard mode

When Kiki was deciding whether to quit private equity and embark on her probiotic Odyssey, she consulted a council of trusted elders who knew a thing or two about starting a business.

Their advice pretty much boiled down to the same thing. Whatever you do, “don’t do this.”

The logic was clear. “Consumer in general is really, really hard. Specifically food.” “Now that I’m in the industry, I know exactly where the advice is coming from,” said Kiki.

In order to start making real money selling food to consumers, you need to be incredibly scaled. As a local operation, “you make no money. If a tech company sells a contract to another company, they could make $100k in a single transaction. That’s a lot of yogurt.

Wanna make it even harder? Make it an animal product, make it refrigerated, and give it a quick expiration date. “Cool, I chose all three,” said Kiki.

But there are a couple of subtle factors that make the Sourmilk strategy different.

First of all, not many other founders are trying to riff off of yogurt. This gives Sourmilk a bit of a moat. “You see people doing frozen yogurt, you see people doing coffee drinks,” said Kiki. But there are “very rarely any new yogurt brands.”

Second of all, yogurt is a product that’s consumed almost every day. “People buy yogurt at a much higher rate than they buy chips on the shelf,” said Kiki. Once you get people to convert, they’re gonna be buying your product just about every time they go to the grocery store.

And finally, Sourmilk is not just about flavor; it’s about health. The internet is awash with recommendations for improving your gut health. But many of these suggestions are difficult to put into practice. Like fasting for sixteen hours every day. Or consuming 30 different plants every week. Or never exposing yourself to stress. But yogurt? That’s pretty simple. Slap it in a smoothie, mix it into a dressing, or eat a big bowl of it for dinner.

If you walk through the aisles of a trendy Manhattan grocery store to find the yogurt section, you’ll notice Sourmilk right away. The electric Sourmilk blue cuts straight through the columns of demure beige. Kiki and Elan designed it this way; they ran a bunch of different tests to figure out which color would stand out the most given the current status-quo color palette for upscale health-forward yogurt. Look at me, says Sourmilk, I’m just a little bit different.

Crossing the cringe threshold

A year ago, Kiki was just a normal person who posted on Instagram twice a year. She had two thousand followers.

Since starting Sourmilk, she’s amassed nearly ten times that. And her 20k followerssee posts from Kiki nearly every day.

To get there, Kiki had to cross the proverbial cringe threshold, that zone of being seen trying-and-not-yet-succeeding. For Kiki, this meant tolerating the discomfort of her Stanford peers, her corporate coworkers, and her friends from elementary school witnessing her arc from respectable private equity grindmaster to aspiring yogurt influencer girlie.

To be cringe is to be seen trying and failing. And if the goal of content creation is to build a following, and everyone starts out at zero, then that cringe threshold is one that every content creator needs to cross. Some might get lucky and be swept over quickly by the mercurial tide of virality. Others are forced to spend months in the mines of the algorithm before they get any traction.

I asked Kiki whether she ever felt stuck on that cringe threshold.

“Totally,” she said. “So many times.”

To get over it, Kiki adopted a couple of principles.

Principle one: commitment. You can’t actually know whether or not your social media presence is viable until you’ve actually spent enough time posting. Unfortunately, this involves exposing yourself to the icy gusts of the cringe threshold. When Kiki started the Sourmilk social media presence, she told herself she had to post every day for three months. With that commitment in mind, the cringe threshold becomes less daunting. It’s about the follow-through, not the response.

Principle two: authenticity. When she posts, Kiki asks herself, “How do I make sure this is my authentic voice?” Anything that she posts online needs to feel like it came out of her head. “I wouldn’t be able to do this if it wasn’t authentic,” she said.

Principle three: purpose. At the end of the day, Kiki asks herself, “What do I want? Why am I doing this?” And the answer is clear. “I’m doing this because I want Sourmilk to be successful. I’m not trying to be an influencer. I’m trying to build a platform and build an audience. I’m doing this so I don’t have to pay an influencer to do this. So I have full control over what our brand and our message is.”

“And I care way more about whether Sourmilk is going to be successful than about whether or not someone is cringing.”

Subscribe now

Brand strategy

When Kiki walked me through her daily schedule, my first question was how she managed to fit a mid-sized content creation empire into a GCal that looks like a pro game of Tetris.

“The strategy is that I have to plan it out,” she said. And, like most other things in her life, she gets it done on the go.

Kiki’s jam-packed early-stage-startup-founder daily agenda means that she doesn’t have a whole lot of time for dedicated content creation. A traditional influencer might block out hours at a time to film atmospheric matcha-soaked morning routine rituals. For Kiki, this is not an option. “If I have to sit down and talk to a camera, it takes a long time.” Anything she does needs to fit into whatever is already on the agenda. Day in the life stuff is good because “it fits into the daily routine. She’ll film about 40 clips in any given day, edit them down, and voiceover the entire thing.

Kiki may be part of the vanguard shaping the next generation of content trends. As we close in on 2026, the lifestyle influencer brand has become a bit passé: the market oversaturated, the daily routines so obviously choreographed and false. Consumers are ready for a new kind of influencer, someone whose content is based around doing instead of just pretending to do. Kiki’s content is authentic, funny, and raw. She doesn’t have time to spend hours and hours perfecting her lighting or choreographing her morning breakfast routine. She’s got yogurt to deliver.

Her crazy schedule might actually be part of what makes her content so good. In what is increasingly a mimetic simulacrum, Sourmilk is a refreshing taste of real life. Sure, it’s still pixelated shortform content coming at you through a screen. But it’s a girl on a mission, showing up every single day with a bigger goal in mind. It’s this relentless energy and momentum that makes Kiki so engaging as a founder and as a person. It’s the reason she’s amassed almost 20,000 followers posting almost exclusively about yogurt. She’s got her eyes on the prize. And she’s moving fast.

Still, despite the unfilteredness, there’s still something remarkably cohesive about the Sourmilk brand: the design, the content, the narrative strategy. I asked Kiki about this, about how she planned it all out. In my head, I saw Canva moodboards, Midjourney mockup concept art, maybe even a small stack of trendy handbooks on design principles. “How did you decide what the vibe was gonna be?” I asked.

Turns out there’s no Sourmilk MegaPlan. The aesthetic and narrative brand is just, “you know, out of my head every day.” There was no grand strategy, no big visionboard, no neatly itemized brand agenda. Kiki just posts stuff that feels right to her. The result: a visually cohesive, narratively compelling brand to leverage this hot new consumer product.

The one rule that Kiki does follow is this. She asks herself: “How do I make sure everything I put into the world is valuable?” “I don’t need another thing clogging up your feed,” she said.

Aspiring content creators stuck in the far recesses of the Alo-set-matcha-pilates matrix might do well to follow in Kiki’s footsteps. Find an awesome mission, and share — honestly — about the journey. Kiki doesn’t need a brand strategy because she is the brand, and her real voice and day-to-day challenges are way more compelling than anything you could map out on Pinterest.

I asked Kiki what she thinks are the biggest mistakes people make when they’re trying to start a business. “There are so many people who wait until something is perfect in order to do anything. The only difference between me and someone else is that I decided to take the leap and do it,” she said. “Just do it. Just post that thing. What are you waiting for?”

In Miyazaki’s film, the only way the magic-witch Kiki can get around town is with her broomstick. Which means that every time she needs to get somewhere, she leaps in the air. The entire Sourmilk journey has been a series of these (metaphorical) micro-leaps: defying business-school wisdom, the obsession with scaling, and the gravity of the cringe threshold in order to keep showing up at your doorstep each day, backpack full of yogurt.

Share

The driver hops out, checks his iPad, and lifts the truck door. The pallet he unloads isn’t a West Elm couch, it’s enterprise-grade networking hardware—designed by the man waiting by the door.

Sunil Varanasi signs for the delivery. He thanks the driver but politely declines the offer to help open the crate. That’s part of the challenge.

Inside the garage, he starts the timer on his Apple Watch. He slices the straps, peels back the wrap, and slides open the casing. Forty-three seconds. He notes this in a small Leuchtturm notebook. Another twenty-four seconds to extract the first unit. Another note. The page fills with timestamps. If you found the notebook on the street, you’d think it belonged to an F1 pit crew.

Sunil works methodically, mounting one of the units onto a rack. One clean turn. Then another. The soft click of perfect torque.

“That’s why we changed the screws,” he tells me later. “The old ones needed both hands. Hold here, screw there. Took over two minutes. Now they twist and lock. No tools. Forty seconds.”

Anil, his brother and co-founder, looks up from his phone and grins. “Forty seconds? You’re getting slow, old man.”

“I’d like to get it to thirty-five,” Sunil replies.

When you’re dealing in bits and atoms, you want an atomic attention to detail.

Subscribe now

The bet

Networking is the forgotten substrate of modern computing. It is also one of the most fragile. The big vendors are stitched together from decades of acquisitions. The hardware is bolted on and designed to be torn out. The software feels unloved. And the people who understand how the whole thing fits together are retiring with no real pipeline behind them.

Inside the industry, the reaction is a shrug. This is just how networking is. This is how it has always been. Meter is betting that this assumption is wrong.

The scale of the category is almost absurd once you look directly at it. Broadcom sits near two trillion. Nvidia earns a massive share of its revenue from networking. Fifteen to twenty percent of a data center’s capex and opex flows into networking and the labor around it. It is one of the largest markets in technology, hiding in plain sight because no one thinks to examine the plumbing.

When I asked Lachy Groom about it, he called Meter the clearest infrastructure bet of the cycle. Models and chips get the attention, but the deepest leverage sits in the layers that outlast hype waves, the layers that get harder, not easier, as demand grows. Networking is one of those layers.

Anil puts it more simply. You cannot build real AI without caring about the plumbing.

And that idea is about to be tested. Every headline points at the frontier labs. Every chart points at GPUs. But all of that activity depends on one thing: packets. A packet is the smallest unit of data on the internet. Every message, stream, query, training run, and inference step becomes thousands of packets racing across the world.

One of Anil’s core beliefs is that we will use the internet far more in the future than we do today. The internet runs on packets, and if you are the company that can best move those packets, the world begins to lean on you.

And the world is beginning to notice. Earlier this year, Microsoft formally partnered with Meter: investing in the company, supporting the training of Meter’s models on Microsoft infrastructure, and adding Meter as the only networking provider on the Microsoft Marketplace.

For decades, networking has been treated as a legacy category rather than a frontier one. The large vendors accepted complexity as destiny. The industry resembled the pre-Tesla car market. Sprawling supply chains. Heavy intermediaries. A belief that vertical integration was impossible.

Meter rejects that premise. You do not inherit the constraints of an industry. You redesign the whole thing. One stack. One interface. One control plane. If the system breaks, you fix the system. If it is too slow, you redesign the loop.

While the spotlight sits on OpenAI, Nvidia, and the model and chips race, the bulb itself depends on the wiring underneath. Meter is just outside the frame, powering the thing everyone is looking at.

This is the bet. That infrastructure, done properly, compounds harder than anything built on top of it. That the AI moment will force the world to see the plumbing it has ignored for twenty years. And that the company capable of owning the full loop will define the next era.

Meter believes networking should run like electricity. It just has not been built that way yet.

The brothers

If you were guessing the path two ambitious brothers would take toward building an internet infrastructure company, you would probably picture San Francisco. But Anil and Sunil Varanasi have never followed the obvious route. Their work has always come from a different instinct. When something is hard or opaque or fundamental, they don’t take it for granted, they take it apart and rebuild it better.

Shenzhen was the clearest expression of that instinct. They went there not for adventure but because they wanted to learn how hardware really gets made. Not from textbooks. From factories. They lived in a small apartment beside a school. Sunil slept in his bed. Anil often slept on the floor, his bed taken up by circuitry. Days were spent inside workshops. Evenings were split between basketball courts and technical manuals. Sunil, trained in biology and genetics, taught himself PCB design from library books. Four months later he was producing functional circuit boards from scratch.

They learned to solder. They learned how heat moves through components. They learned how hard it is to convince a factory to change a production line for a fifty unit order. When they first tried to get production contracts in Taiwan, almost every factory turned them away. Too small. Too early. Years later, the same factories were competing to win Meter’s business.

But the deepest reveal about the brothers came later, when San Francisco finally beckoned. One morning Sunil walked into the office carrying a document he had written overnight. More than a hundred pages. A complete rethinking of how the entire networking stack could work if you threw away the industry’s assumptions and started from physics. Anil compared it to Moses coming down the mountain. Except instead of commandments, Sunil had diagrams, topologies, and rules for a world that did not yet exist.

He placed the document on the table. Anil read it quietly, nodding along as he went through. Turning the final page, he said “Let’s do it.” And the company shifted around the document.

Alan Kay once said that people who are serious about software should make their own hardware. For the brothers, seriousness meant something even stricter. You go to the factories. You understand the physics. You master the whole stack. And you do not stop until the system does exactly what you intend.

Share

The stack

The document Sunil placed on the table was a new physics for how a network could work. In the old world, hardware and software lived on different timelines. Devices arrived with their own quirks, constraints, and silent assumptions. Software bent around those quirks. Operators bent around both.

Meter started somewhere else entirely. If you were going to rebuild the system, you would vertically integrate it. Make the physical and the digital match perfectly. Reject the precedent to commoditize networking hardware and instead build beautiful hardware that fits with its precise software twin. Every configuration would live in the cloud, not on the box. A switch would reach the customer fully configured, waking up and behaving as an engineer intended. Logic would dictate hardware.

Once you build everything as a single unit, the whole system shifts. You can simulate a rollout before anyone unspools a cable. You can redesign a topology with the freedom of refactoring code. You can validate thousands of networks on your phone. You can build powerful models that automate network design, deployment, and management. A network stops looking like patchwork equipment and starts looking like a coherent fleet.

Meter is creating the Tesla moment for networking. The difference between coordinating a constellation of inherited parts and owning the entire loop. Tesla controls battery, motor, firmware, data, fleet operations. Similar for Meter: hardware, firmware, cloud, installation, operations, and now the model layer. One stack. One interface. One control plane. A universe they designed end-to-end. Intelligence is no longer an add-on, it is the next step.

The model is not just trained on dashboards or logs. It learns from the structure itself. Switch behavior. Wireless patterns. Routing logic. Power constraints. Physical layouts. Failure modes. The physics of how a real network breathes. Because the physical and digital match perfectly, the model can reason with the same fidelity as a human engineer, only without the limits of place or time.

Give it a floor plan and a few requirements and it produces the full design. Placement. Cabling. Switching. Wireless. Redundancy. All validated across thousands of simulations. Weeks fall into minutes. And as the available talent pool of network engineers declines, Meter’s uniquely positioned to step in and support them. The old world loses expertise while the new world gains leverage.

This is the moment where Meter’s strategy begins to feel inevitable. When the stack is unified and the network becomes knowable, it becomes something that can think.

You can also communicate with it. With Meter, you can DM your network. You type a message as you would to a colleague. The system understands. You tell it what you want. The network decides how to make it happen.

This is the model. The intelligence that follows naturally from rebuilding the world underneath it. This is why Meter can build what no incumbent can attempt.

The distribution

If you want to change an industry, you need a distribution engine that compounds faster than anyone else’s. For Meter, that engine came from observing how companies grow.

Startups sign years-long office leases that they outgrow in months. They expand, overflow, and move again. Each move forces them to rebuild their network from nothing. Hardware orders. Contractors. Cabling. For companies used to shipping daily, waiting weeks for internet plumbing feels obscene. And every time they leave, they have to rip out the network they just paid for.

Meter realized that if the hardware stays in the walls, the next tenant can use it on day one. No waiting. No rebuild. The departing customer takes Meter with them to their new space, and the incoming tenant inherits a working network, already shaped for the physics of that space. One installation now serves several users across time. Once Meter is in a building, demand starts to arrive on its own.

To make this work at scale, Meter needed buy-in from the landlords. The building itself had to become the channel. And once Meter is in the building, it’s there for good.

Girish Gopalan on Meter’s real estate team spent months visiting landlords, often with boxes of donuts, showing them what it means for a tenant to open the door to a new office and finding the network already alive. The pitch was simple. You have water. You have power. Now you have Meter. A third utility.

With Meter, the space rents faster. Tenants settle sooner. Brokers close deals more easily. Even if the tenant chooses another space, they return months later and ask for Meter in the one they finally select. Win a tenant and you meet their landlord. Many landlords own dozens of buildings. Some own entire portfolios. When a building changes hands, the new owner often reaches out first. They have heard that Meter removes uncertainty. They want it from day zero.

Clusters of installations also drive down labor costs, turning every new building into momentum for the next. As the footprint grows, so does the credibility. It becomes harder for anyone else to even enter the conversation. There are no real brands in networking. Ask someone who provides their connectivity and they will almost always shrug. The category is invisible.

And the part no incumbent can imitate is the part that makes the flywheel turn. None of them control hardware, firmware, installation, support, and operations in one continuous loop. Without vertical integration, you cannot ask landlords to standardize on you. Meter can.

Which is why the strategy begins to feel inevitable. Offices were the beginning, Meter now operates across industrial, retail, healthcare, education, and data centers. Meter wants to be the connectivity layer waiting inside every space before anyone signs a lease. The default because it is already there.

Subscribe now

The flywheel

Meter discovered early that networking could be financed the same way you would finance a long-lived utility. A physical footprint that pays for itself many times over.

When Meter wins a customer, the customer isn’t just buying a network. They are helping Meter acquire the location itself. The hardware stays in the space long after the tenant leaves. The next tenant inherits it. One installation begins to serve two users, then three, then four. The cost of the original deployment spreads across every future occupant.

This is where traditional networking breaks. Most vendors treat hardware as an expense that has to be clawed back immediately. They don’t control enough of the value chain to earn from it twice. Meter does. Vertical integration makes the investment self-amortizing.

That self-amortization is what sets the flywheel in motion. Better integrated hardware enables better software. Better software creates more customer value. That value supports higher lifetime revenue per location, which is reinvested into tighter integration, faster installs, and the next generation of hardware. Each turn of the loop makes the system cheaper to operate, harder to replicate, and more valuable to every customer that follows.

The economics follow naturally. The recurring price covers hardware, firmware, software, installation, and support. Hardware is the minority of the cost. Labor is the majority. But labor collapses when you install multiple suites in the same building. A cluster of ten deployments can cut labor costs by nearly half. Meter only needs a small percentage of those suites to become paying customers for the numbers to work. The rest becomes upside.

This is why Meter treats buildings as part of the network. Once a location enters the footprint, it behaves like infrastructure. The cost to support it falls. The revenue compounds. A building becomes a long-lived asset in the same way a stretch of fiber or a piece of grid does. Something you invest in once and earn from repeatedly.

It also explains why Meter is comfortable with capital intensity. Most hardware companies avoid capex because they don’t own the loop. Meter owns every part of it. Hardware, firmware, cloud, installation, support. The capital they deploy today returns through tenants they haven’t even met yet.

This unlocks financing mechanisms most software companies never see. Equity for speed, venture debt for footprint, revenue-linked financing tied to buildings already under contract. Growth funded by the predictability of infrastructure, with less volatility than quarterly sales.

Most people still see boxes of hardware and a subscription price. Meter sees a footprint that compounds.

The team

Meter the company is built from two tribes. On one side are the networking veterans: people whose careers have been shaped by the physics of real systems. They speak in failure modes, tolerances, and thermal budgets. They know what happens to a switch when a room gets a few degrees too warm. They know the ghosts that live inside legacy networks and what’s required to keep them running.

On the other side are the software-natives who grew up in a world of SaaS. Designers, operators, and engineers who move quickly, improvize well, and have a natural instinct for turning something complex into something a customer can actually use.

Most companies would separate these groups, or let them operate on different tempos. At Meter, they sit a few feet apart. They work with the same temperament: a calm intensity, a grounded focus, a sense that the work matters but panic does not. The networking folks bring the why. The builders bring the why not. Together, they give the company shape.

Lachy Groom calls Anil a “generational recruiter.” It shows. The team Meter has assembled shouldn’t exist at this scale, in this industry, at this moment. A group that moves like a single unit, despite coming from opposite ends of the universe.

Directly across the street from the office sits Meter’s distribution center. Not an outsourced warehouse an hour outside the city. Not a black box. Across the road. You can watch hardware move from design to assembly to deployment in one continuous loop. Few companies get to see their product cross the street in real time.

The clock

But none of this guarantees anything. Meter thinks in decades, but the opening to define the next era of networking is much shorter. As Anil puts it, the next thousand days will decide who becomes the default provider. The window is wide today, but it will close.

Meeting that moment means stepping out of the quiet that protected them for so long. Almost no one knows the Meter name. At first that was intentional, then a byproduct of focus. Now it is a liability. The company rebuilding the networking layer of the internet finally has to tell the world what it has built.

But telling the world is not foreign territory. Before Meter, the brothers ran a small film studio as a side project. Anil in particular has a habit of spotting creative talent early: he backed Dwarkesh Patel when his podcast had only a few hundred listeners. These are small signals, but they point to something important. Meter understands taste. They understand narrative. They understand the outside world as well as the inside one.

The next thousand days are the test. A unified architecture. A unified model. A unified team. And now, finally, a unified story.

The company is built like its product. And the clock is running.

The end

It’s Friday late-afternoon, and I’m in a conference room with Anil and Sunil. Sunil is at the whiteboard in his usual stance, sketching packet paths at speed. At first the lines look chaotic. Then, with a few final strokes, everything snaps into place. You get the sense the whole system is sitting behind his eyes, waiting to be drawn.

Across the table, Anil watches in his familiar way. Quiet. Focused. Letting the explanation run until one sentence from him lands in the exact spot that makes the whole diagram click. Their rhythm is easy to miss if you’re not looking for it. One builds the world. The other orients you inside it.

We’ve gone past our allotted time, but Sunil is still drawing. He’s onto installation now, talking through how it used to require two hands, then one, and how every small improvement frees a technician somewhere in the real world. Now that I’ve seen how Meter’s model works in the macro, this part makes perfect sense: the big shifts come from a thousand tiny adjustments.

I mention something Anil told me earlier, about wanting Meter to feel like it was designed by one person. I say that it already does, and that person is Sunil.

Anil smiles. “That’s right.”

Sunil caps the marker and apologizes for having to run. They’re launching nine new hardware models, and one of them has just arrived across the road.

As he ducks out, Anil calls after him, “Two hands last year, one hand this year, it better be hands-free next year!”

The thousand-day clock is already running. And, somewhere on a tree-lined street just south of San Francisco, another pallet is being opened — a little faster than the one before.

Share

Thanks to Anil, Sunil, Sarah, Liam, Lachy, Nitan, Girish, Pike, James, and everyone else who took the time to help with this piece.

So, Mati heads out alone towards Half Dome, and Piotr stays in the passenger seat, editing. Head down, headphones on. The car, already their bedroom, now doubling as a study hall.

Out in the valley, the sun falls faster than expected. On the descent Mati’s phone dies. No signal. No light. No map. By the time he reaches the base, the trailhead is empty. No Piotr.

He boards the last shuttle of the night, a slow loop back to the gravel patch where they slept the night before. A forgettable corner of the valley. A place that feels like a misprint against all the beauty around it. Mati stands there for a moment, doubt creeping in. He weighs his options. Walk back. Wait. Find a motel. Or trust the logic that has carried them through fifteen years of shared problems.

He chooses trust.

Five minutes later headlights cut through the dark. Piotr pulls in beside him. Calm. Unhurried. A drift of printed pages and protein bar wrappers on the passenger seat that Mati brushes aside as he climbs in.

“How’d you know where to go?” he asks.

“You weren’t at the trailhead,” Piotr says. “So I figured you’d double back.”

Not instinct. Not luck. Just reasoning. A Bayes-flavored friendship honed across years of math olympiads, side projects, and the learned practice of thinking in tandem with someone.

That was May 2017. Today, they run ElevenLabs, one of the fastest growing and most technically ambitious companies in AI. But the old pattern remains intact. One moves outward. One goes inward. One tests the edges. One traces the logic. Both arrive at the same point.

They were lucky to have found each other early. And the world, as it happens, was lucky too.

A perfect marriage

Mati and Piotr have this quiet, practiced way with each other, a mutual understanding at a molecular level. “They’re like an old married couple,” a16z’s General Partner Jennifer Li tells me.

Sitting in Eleven’s London office, Mati pulls up an old photo from high school: the two of them at Poland’s national math olympiad. “We had all the same classes,” he said. “We were competing on everything.”

But the competition was surface. Underneath was recognition. Piotr could sit with a problem until the structure appeared. Mati would push outward, try things, break things, bring back a piece that made the whole thing clearer. A natural split of focus, but with reflexive collaboration.

It carried through everything they did. Officially, they first worked together at Opera in 2015, but unofficially they’d been working together for years. They spent one summer road tripping across the Balkans, once again deciding a car was accommodation enough. “Not optimal,” Mati said, looking back, “but it was us.”

And they built things. A recommendation engine. An early voice analyzer. A crypto analytics engine during the 2021 frenzy. The substance of the projects mattered less than the pattern: they kept choosing each other.

“We used to speak every day. Now it’s not exactly daily, but I usually predict what he’d say,” Mati told me. “Piotr and I know each other so well we even know the weaknesses the other person is trying to hide.”

It’s part married couple, part platonic ideal: two people shaped by the same culture, trained on the same problems, building on decades of shared instinct before a single line of ElevenLabs code existed.

Subscribe now

A gap in the market

A movie in a foreign country usually comes in two forms. A dubbed version, where actors are fully replaced by local voice talent. Or subtitles, where you keep the original performance and read your way through it.

Poland took a third path.

Mati and Piotr grew up on lektor films, the uniquely Polish format where every foreign movie is voiced over by one man reading every line in the same flat tone. You hear the original actors murmuring underneath, like ghosts trapped inside the film. Jarring when you first encounter it and invisible if you grow up with it.

“It shapes how you listen,” Mati told me. “You start paying attention to what’s missing.”

So they learned early on that tone isn’t purely decoration, it’s meaning. That delivery shapes understanding. Which is why the rest of the world’s indifference to voice always felt a little strange to them. For most of the 2010s, voice tech plateaued. Siri and Alexa handled weather and kitchen timers, but nothing that required nuance.

Then the generative wave hit. Text exploded. Images followed. Video became the next gravitational center. Voice stayed peripheral. For Mati and Piotr, voice hadn’t been solved. It was the one category that still felt open.

By 2022, when Mati began thinking seriously about building something together, the pieces had aligned: “Piotr came to me and simply said ‘the models are ready.’” The tools had finally caught up to the problem they’d been circling since they were teenagers.

Meanwhile the industry still hadn’t noticed. Voice looked irrelevant. And that disconnect showed up immediately in fundraising.

They pitched. Most VCs passed. But Credo Ventures said yes. They led the pre-seed with support from London-based Concept Ventures. At this stage, Mati and Piotr didn’t even have an office, so the first few months of ElevenLabs were built from Concept’s boardroom.

As Concept’s investment memo put it at the time, audio had been neglected by recent advances in research, and ElevenLabs looked like the team to bring the medium back to life. “It felt like a golden window,” Oliver Kicks told me. “A moment where the technology and the team lined up.”

It wasn’t obvious. Voice wasn’t dead. It had simply been abandoned. And the only people who noticed were two founders who had grown up attuned to absence.

The birth of the company

If you ask Mati about the origins of ElevenLabs, there’s no eureka moment. What he described instead was a narrowing. The sense that the problem they’d been circling for years had finally come into focus, sharp enough that they could step straight into it.

So they fell back into the rhythm they always had. Piotr went inward. He dug into papers, built prototypes, and pushed at the edges of the new architecture. Mati went outward. He talked to early users. He sketched use cases.

Their first product was a wedge: a dubbing tool for creators. The product was small, but the ambition wasn’t: they figured creators alone could reach one hundred million in ARR over time. But they never saw that as the destination. They saw voice as infrastructure. As a new interface waiting for its moment.

The first version of ElevenLabs was a simple text-to-speech tool with a tweet-length character limit. They launched it quietly, but users loved it: one author copied and pasted an entire book into it, 240 characters at a time.

From the outside, the rise looked sudden. From the inside, it felt like a long, submerged line finally breaking the surface. Even the structure of the company reflected the balance Mati and Piotr spent years protecting. They split equity evenly. They split their focus the same way: Piotr on research, Mati on go-to-market. They moved differently but always in parallel, each layer reinforcing the other.

This wasn’t two founders discovering a market. It was a market finally catching up to two founders who had been preparing for it their whole lives.

Share

The two-body solution

I didn’t get to meet Piotr. Almost nobody does. The distance can make Piotr sound mythic, but he just loves the work, and it’s hard to pull him away from it. Inside the company, he is deeply present in the work itself, protecting the space where the real thinking happens.

Mati moves differently. He is the surface area of the company, the part that meets the world. He gathers information, translates ambiguity into direction, and pulls the horizon closer for everyone else. Even investors who passed on the early rounds told me they walked out of their first meeting with him convinced he would eventually run something enormous.

Jennifer Li told me about a fundraising meeting with a16z. Both sides had been going back and forth for hours. Then Piotr gestured for Mati. They stepped aside, spoke to each other in Polish for five minutes, came back, and closed the deal.

Piotr’s research team is tiny, but unusually effective. They often hire by artifact, searching GitHub repos. One early researcher came straight from a call center. Another was building voice models in his spare time.

Mati is the counterweight. He built the commercial engine, the partnerships, the teams, the surface of the product. People kept telling me the same thing: he learns at a frightening pace. Introduce him to someone sharp and he absorbs their best instincts within a week. The company compounds because he does.

On paper, ElevenLabs is a high-growth LLC. In practice, it still runs like a partnership. The culture is what forms in the magnetic field between the two.

Early-middle game

Mati is a chess player, so I asked him how he sees the board. “We’re in the early middle game,” he said. Enough clarity to see the lines, enough uncertainty that every move matters.

The first phase of AI voice is over. The primitives are stable. The world has accepted that expressive audio is solvable. Now the real contest begins: latency, reliability, multimodal agents, distribution, and all the invisible seams that determine whether a company becomes infrastructure or an afterthought.

But now is where most companies lose their way. This is where patterns calcify. It’s where teams start optimizing for the wrong thing, or get distracted by adjacent categories, or confuse motion with progress. “It’s so easy to copy the primitives now,” Mati told me, “that people assume that’s the work.”

So I asked him what the endgame looks like.

There are real risks for ElevenLabs. Their research advantage is powerful but fragile. Fifteen researchers competing with labs that have hundreds. As Mati noted: fall behind on research, and nothing else matters. Product momentum can’t compensate for losing the frontier. Culture is another pressure point. They’re over 400 people and growing. The original rhythm—Piotr deep, Mati wide—gets harder as the company expands. Growth rewards alignment. It punishes noise.

The market is changing too. Voice is no longer an afterthought. The first wave of competitors is arriving. The big labs are now paying attention. Customers want more reliability, more control, more integration. Winning the opening doesn’t guarantee winning the game.

Mati knows all of this. That’s why he says early middle game. The opening gave them shape; the middle game decides whether that shape endures.

Ghosts in the machine

All the chess talk has lit Mati up now, and we end up talking about Richard Feynman. He’s one of Mati’s heroes. Feynman believed that if you truly understood something, you could explain it simply. Strip the idea down until the shape shows itself.

When I asked who else he admired, he mentioned von Neumann. A different creature entirely. The kind of mind that sees the full architecture of a problem all at once. Hearing Feynman and von Neumann side by side, it seemed obvious he was sketching the archetypes of him and Piotr.

So I asked him.

Mati gave a small smile. “Yes,” he said. “That makes sense.” In recent weeks, ElevenLabs announced they now offer a licensed Richard Feynman voice—one can only hope for a future von Neumann.

Another voice on offer is Alan Turing. The man who gave us the cultural shorthand for machine intelligence is now a machine himself. Playing around with the product, I used ElevenLabs to hear myself speak Spanish. Properly speak it. Not the strained, halting version I manage in real life, but a version with flow and timing. The voice sounded like me. I passed my own Turing test.

I had given the model a clip of me doing standup. Partly because that’s the most clips I have of myself, but partly because so much of standup is dependent on voice. Standup is jokes, sure, but it’s also breath and timing and intention. Delivery is the difference between a joke landing and a joke dying.

This is the frontier Mati and Piotr are pushing. A synthetic voice that holds timing and tone with enough fidelity that you stop thinking about the mechanism and start reacting the way you react to people. You hear something that feels real, and your brain follows.

Maybe that is the real threshold. Not the moment a machine fools you, but the moment you stop thinking in terms of fooling at all.

The business behind it

The company closed 2025 with three hundred and thirty million ARR, with close to a fifty-fifty split between self-serve and enterprise. Last week, the company raised a $500m Series D at $11 billion, a relatively modest multiple in a frothy AI market.

The business today runs on three lines of revenue. There is the Creative platform used for expressive translations, dubbing, and text-to-speech. There is the developer API that powers thousands of voice, sound, and music applications. And then there is the Agents platform, ElevenLabs’ enterprise layer, growing more than three hundred percent in the past twelve months.

That shift in scale came with a shift in footprint. ElevenLabs now works with companies across many sectors, from customer support to entertainment to education, including names like Cisco, Epic Games, Adobe, and Nvidia. A large part of this momentum comes from their work on conversational agents. Developers and enterprises have already built more than two million of these agents on the platform, deployed across web, phone, and apps. Businesses can bring their own logic and knowledge base, connect to existing systems, and rely on low latency under the hood.

What makes all of this possible is the company’s decision to train and run its own models from day one. Most generative AI companies today pay large fees to upstream providers for every request. ElevenLabs does not. That means its margins look more like traditional software than most AI companies. It also means they can expand into new modalities without breaking the economics beneath them. Voice led naturally to sound effects, then music, then multi-speaker scenes, and those will eventually lead into richer studio and enterprise pipelines.

The creator wedge was never just about distribution. It stress tested the product at the highest possible resolution: accents, pacing, emotion, and edge cases. When the Fortune-level customers arrived, the foundation was already set. In a similar expansion to their original creator focus, ElevenLabs has recently added generative image and video models to the platform.

All of this has been built by a team of roughly four hundred people. Many other frontier AI labs number in the thousands, but ElevenLabs feels very different. Smaller, faster, more opinionated about taste and quality. It is a generative AI company, but also an audio company. A consumer product that grew into enterprise infrastructure. It now feels less like a single model and more like the early architecture of a full audio platform, the layer other companies build on top of.

From the outside, ElevenLabs looks like it clicked into place overnight. From the inside, it was a precise alignment of timing, taste, and two people who’d been circling the same problem for half their lives. Once the models matured, everything else snapped into place.

Subscribe now

Voice as interface

Voice technology is a good business, but there’s a bigger question underneath. Because voice isn’t just a feature. It’s the oldest interface we have. Long before writing, before screens, people taught one another, signaled trust, and revealed intention through tone and cadence. Meaning lives in how something was said as much as in the words themselves.

For years, voice technology gave us a flattened version of that. A thin synthetic voice on one end and you yelling ‘operator’ on the other. Expectations fell because imagining anything better felt abstract. The best we could hope for was eventually reaching a human agent. ElevenLabs shifts that baseline to the point where many people may prefer to speak to an agent. And not just for navigating refunds, but for the real stuff.

Mati’s been speaking with his hands, and I notice a bracelet on his wrist: black and white beads spelling out the number eleven. “My niece made it,” he tells me, and I think of the world she’ll grow up in. While Mati and Piotr had math textbooks and lektor films, she’ll grow up in a world of voice.

If she wants to learn chess, she won’t be staring at some dead static board on an app. She’ll be talking to an instructor that can match her pacing, explain a tactic three different ways, slow down when she hesitates, and speed up when she grasps it. And when she gets curious about the people she reads about, she’ll be able to talk to them too. Imagine learning about ancient Egypt with Cleopatra, or early aviation with Amelia Earhart.

Educators have dreamed of this for decades. Having a tutor can raise student achievement by two full standard deviations: going from the middle of the pack to the top of the class. It’s how we educated royalty, and now it can be for everyone.

Once models can express tone, pacing, hesitation, and emotional contour with fidelity, whole categories move. Education becomes more personal. Accessibility becomes more humane. Entertainment becomes easier to translate across borders. Agents feel present rather than scripted. Translation keeps the performer’s voice instead of replacing it.

None of this replaces human voice. It expands what’s possible. Voice is a medium we left underdeveloped, and what ElevenLabs is building is a world beyond flat assistants and canned prompts. It’s a set of primitives for expressive audio at internet scale.

And in a neat loop, it returns to where Mati and Piotr began. Two kids in Poland listening to flat lektor films, learning to notice what was missing long before they could name it. Now they’re building the infrastructure that restores the range they never heard growing up.

Coda

Later that night, they were back in the same gravel patch. Yosemite was a shadow outside the windows, and they’d split the car in two: front seat for work, back seat for sleep. Mati lay in the back, head against a rolled jacket. Up front, Piotr was still awake, his face lit by the glare of his MacBook. His wallpaper was a picture of the valley outside, the closest he’d come to hiking it that weekend.

Piotr’s paper would be finished in time and accepted into the conference. A clean result for a weekend spent sleeping in a rental car. A few hundred miles away, a group of Google researchers were working on a paper for the same conference. “Attention Is All You Need”, a similar last-minute sprint, would lay the foundations for the field Mati and Piotr would enter a few years later.

None of that felt real in the car. It was just two friends at the end of a long day. One reading. One sinking toward sleep. The valley was silent, save for the clicking of keys and the occasional rustling of pages. Mati closed his eyes, dreams unfolding like a lektor film, voices murmuring underneath.

Share

Thanks to Mati, Luke, Olly, Jennifer, Mark, Nev, Victoria, Maciek, Franek, Hannah, and everyone else who took the time to help with this piece.

Enjoyed this deep dive? Subscribe for future ones.

Subscribe now