How Emergent Got to $100 Million in ARR

You've probably seen the headline. Emergent, the AI "vibe coding" startup where you describe an app in plain English and it builds the whole thing, frontend, backend, all of it, hit $100 million in annual recurring revenue eight months after launch. The founder posted it on X and people immediately started arguing about whether the number means what it sounds like.

Fair point. It's a run rate built off usage, mostly token consumption, not booked annual revenue the way SaaS companies normally report it. One critic on X asked it straight up. If you charge someone a discounted $1 for their first month and they renew at $20, does that count as $240 of annual revenue on day one? Nobody outside the company has a clean answer, and Emergent hasn't published which plans or trial periods actually feed the number. A Lightspeed India partner defended the company publicly. By the time Emergent's own head of growth was doing interviews about all this, the number he quoted wasn't $100 million anymore, it was closer to $120 million.

That's worth a paragraph, not the whole article.

The interesting part is how they actually got people to show up in the first place, and two long interviews came out this year where the people who built that engine just talk through it, mistakes and all. It's a much better story than "they went viral."

Two founders who'd already been through it

Emergent's founders are twin brothers, Mukund and Madhav Jha. Mukund studied engineering at Columbia, worked at Google, then co-founded the quick-commerce startup Dunzo, where he was CTO until he stepped down in late 2023 during a rough restructuring. Madhav spent time as an AI scientist at Amazon and later as a machine learning engineer at Dropbox. Neither of them was doing this for the first time or doing it on beginner's luck.

The money followed the growth curve almost exactly. A $23 million Series A from Lightspeed in September 2025. A $70 million Series B three months later, led by Khosla Ventures and SoftBank, that tripled the company's valuation to around $300 million. By the summer of 2026, a Series C led by Creaegis put the valuation near $1.5 billion, and Freshworks founder Girish Mathrubootham joined the board on behalf of Together Fund, one of the company's early backers.

They started with nothing

Four weeks before Emergent's public launch, there was no marketing team, not even a small one. One of the founders admitted as much in an interview. They sat down and asked themselves what marketing even meant and where their users already hung out online. That was the whole strategy at first.

They landed on influencers, people on Twitter and Instagram who already had an audience that trusted them. Instead of paying anyone for a hard sell, they gave a handful of them free access first and waited to see if they actually liked the product. If someone didn't, better to find that out quietly than have it blow up in front of their whole audience.

The Netflix clone that kicked it off

The first real spark came from a creator based in Singapore. He built a Netflix clone with Emergent, nothing fancy, but it looked good, and it went properly viral on Twitter, something like 700,000 views, all organic. Signups jumped right after.

That's when the team saw what actually worked. Creators showing off something they'd genuinely built beat creators reading an ad script, every time. The best posts weren't generic app plugs. They read like someone losing their mind over what they'd just built in twenty minutes.

So they kept testing with more creators like that. Same pattern every time, post goes up, signups spike. It worked enough times in a row that they stopped treating it as an experiment and went all in, from five creators to three hundred.

Buying a domain on a car ride to Mysore

Before he even joined the company, one of Emergent's early growth people was driving to Mysore talking with friends about how India should host a big hackathon for this new wave of AI coding tools. Someone asked what to call it. He said VibeCon. They checked if the domain was free. It was, for $1,500, and they bought it right there in the car.

He tweeted about the idea. Emergent's co-founder saw the tweet and messaged him to sponsor it. A tweet started the whole thing, not a marketing plan.

Once he joined the team, the first version of the plan was fairly ordinary, a hackathon at Y Combinator's office with AWS and Anthropic as partners, one more event in a busy week. What made it different came from one addition. He pushed to get YC to offer the winner a direct interview, skipping their usual application process entirely. Emergent had to actually ask YC for that, and YC said yes. The San Francisco edition pulled in four to five thousand applications. The India edition, a few months later, pulled in thirty thousand.

Hackathons aren't new. What's rare is a hackathon winner skipping YC's entire application process.

They picked one channel and went all in

Emergent had every channel available to them and could have chased all of them at once. They went deep on one instead, and that choice is worth copying even without $100 million behind you.

Their competitor Lovable built its growth mostly on product-led growth and SEO, template pages in the style Zapier and Canva made famous. Emergent tried a bit of everything early on too, but once influencer marketing started working, they poured effort into it instead of splitting attention across five other channels. Lovable eventually built its own creator program, roughly six months after Emergent had already gone all in. By then Emergent had hit its billion-dollar valuation.

It wasn't just Lovable playing a different game. Higgsfield, an AI video startup in a similar space, went all in on a different version of the same bet. They turn every single feature release into a full press cycle across short-form, long-form, Twitter, and LinkedIn, so that even a minor update feels like the company just launched again. It's the same logic, just applied to a different channel, betting everything on one thing done enormously well instead of five things done adequately. And the field Emergent was fighting through wasn't small either. Replit, Rocket.new, Wabi, and Anything were all chasing the same non-technical builder audience at the same time.

If a channel gives a real signal, and that signal repeats two or three times, that's reason enough to stop testing everything else and go win it hard. Being decent at five channels rarely beats being genuinely dominant at one.

How they organized the team once it started working

Once influencer marketing was clearly the engine, Emergent split the growth work into three separate teams instead of one group doing everything.

One team owned acquisition, channel by channel. Whoever ran influencer marketing lived and died by that one channel, same for whoever ran organic or affiliate. Each person got a clear brief, which geographies, which persona, and just as importantly, which users not to bother chasing. An acquisition team's only job was getting people in the door, nothing about what happened after.

A second, smaller team owned what happens after someone signs up, conversion, activation, retention, turning a signup into someone who sticks around. Emergent's growth lead described these as the people who get energized by cutting the same data fifteen different ways to find out where a campaign actually worked. Mixing that job with acquisition, in his telling, is how you end up hurting both.

The third team was creative, and it sat across the other two instead of living inside either one, writers, editors, and video editors, the people actually making the content a creator posts or a performance ad runs. There was a simple reason for splitting it out this way. Performance marketing runs on creative, organic runs on content, and influencer marketing runs on content too, so whoever makes that content should serve all three channels instead of being boxed into one.

It wasn't smooth

For the public launch, they lined up several creators to post at the same time as the announcement. One of them posted four hours early anyway. You can brief a creator, but you can't fully control one. Signups still spiked, and a "coordinated campaign" that looks clean on paper turned out messier once it was actually happening.

The first piece of short-form content that really took off came from a creator whose whole channel was about sketchy corners of the internet, the kind of "sites you didn't know existed" content that no brand safety review would sign off on. It worked anyway.

What scales, and what doesn't

Twitter and LinkedIn influencer posts tap out fast. You can only run that play so many times before people stop reacting. Short-form video scales better. And the most scalable thing of all, by their own account, is just the founder showing up and talking, consistently, as himself.

Attribution was a mess too. On Twitter or LinkedIn you can put a tracked link at the end of a post and know exactly what it drove. On short-form video, you mostly can't. They leaned on timing instead, if signups spiked right when a video dropped and nothing else was happening that day, that was answer enough, even without a clean link to prove it.

Once revenue passed $50 million, they layered paid ads on top of what was already working, instead of starting there. Influencer marketing and what they called "moment marketing," basically treating every milestone like a small product launch with its own press push, stayed the backbone the whole way through.

What's worth stealing here

You don't need $100 million to copy the shape of this. Give a handful of real users or small creators free access before you ask them to post anything. Watch for a spike that repeats more than once, not a one-time fluke. When you find it, go deeper on that one thing instead of spreading your time evenly across ten channels. And if there's a chance to attach your launch to something bigger than your own name, take it, the way they did with YC.

Someone still has to find the creators, brief them, chase the ones who go quiet, and figure out what actually moved signups. That workload is what grew Emergent's team as they scaled. It's also the kind of busywork an AI CMO like Okara is built to take off someone's plate, so a two-person team can run something closer to what took Emergent a whole department.