|
$10M ARR
In Just 2 Months
|
1M+
Total Signups
|
~20,000
Paying Customers
|
10–20K
Daily Active Apps
|
How does a brand-new AI company go from basically zero to $10 million ARR in two months? Not "two months after the seed round," not "two months after the pivot" — just two months. That's the question hanging over Emergent, the vibe-coding platform co-founded by Mukund Jha and his twin brother Madhav. The speed is the headline, sure, but the more interesting story sits underneath: a relentless shipping culture, a product built for real production apps (not just demos), and a growth approach that looks more like an engineering sprint than a marketing campaign.
Emergent competes in one of the most crowded corners of the AI landscape — alongside Lovable, Replit, and a growing stack of code-generation tools. Yet in a space full of "artifact wrappers," this team made a different bet: build for production, not prototypes. The gap that created — between what dev shops charge and what Emergent costs — turned out to be the real engine behind the growth.
This is the full story — from a Visual Studio CD that sparked a coding obsession, to a bakery owner in the Philippines who unknowingly redefined the company's market, to 500 AI-generated videos a day and a Pixar CTO who became a power user.
- The $10M ARR Moment That Turned Heads
- What the Team Felt — and How They Kept Shipping
- The Surprise Pivot: When Non-Technical Users Showed Up
- Twin Founders, One Shared Obsession
- Vibe Coding, Explained Like a Normal Person
- What People Are Actually Building on Emergent
- Why Emergent Thought It Could Beat the Incumbents
- The Growth Playbook: From No Plan to 1M Signups
- Pricing, PMF, and the "Real Software" Trade-off
- Contrarian Decisions and What's Next
- What I Learned From This Startup Story
- Key Takeaways
- FAQ
The $10M ARR Moment That Turned Heads
The reaction to Emergent's run wasn't polite admiration — it was closer to disbelief. The milestone hit a nerve across the ecosystem, especially for people who've been asking the same question for years: can a company built by founders from India become a global AI player at the highest level? That's why names like Lovable and Replit come up in the same breath, and why comparisons to the fastest-growing AI labs hang in the background. The subtext is clear: this isn't supposed to be easy, and it definitely isn't supposed to happen this fast.
Mukund's response to the moment is refreshingly unromantic. The numbers were "still sinking in." The team celebrated briefly — clapping, cake — then went straight back to shipping a new release. That rhythm matters. The milestone wasn't treated as a finish line; it was treated as a checkpoint, then back to work.
Sitting underneath the excitement is a bigger belief: new technology waves reset the board. Globalization helped. Access helped. Talent helped. And now, with AI, the founders see a rare opening to — in their own words — "play the big game."
What the Team Felt — and How They Kept Shipping
Big revenue numbers can do weird things to teams. Some get nervous. Others get sloppy. A few get complacent. Here, the emotion sounded more like fuel than pressure. Mukund described the mood as "super pumped," but he framed it as the start of a marathon — speed is high, expectations go up, and now you find out whether your systems and your people can hold.
What stands out is how early they optimized for a core team that could move fast. The first hires included people Mukund had worked with before — including a "rockstar engineer" from Dunzo — plus a product leader named Sorab. The shared mission wasn't just to build a tool. It was to pick problems they could stay excited about for the next 20 years. That's a very different filter than "what's trending this quarter."
The work cadence is intense, even by startup standards. Early mornings. Long days. Six to seven days a week. It's not presented as a badge — more like the obvious byproduct of trying to out-ship a crowded market. And that detail about shipping on the very day they hit their big milestone matters more than it seems. When a team hits a big number and immediately returns to rollout work, it usually means one thing: they still think the product is the main story.
The Surprise Pivot: When Non-Technical Users Showed Up
Early on, the bet looked pretty standard for developer tooling. The assumption was that senior developers, product managers, and semi-technical builders would be the primary users. Then something happened that changed the internal story entirely.
A non-technical user — a bakery owner in the Philippines — built an ordering website on Emergent during beta. It spread fast. It was one of the first moments where the team saw "real usage" spike in a way they didn't expect, and it forced a total rethink of who this product was actually for.
Mukund's takeaway was blunt and kind of beautiful: there are a billion people with ideas stuck in their head, and tools like this can unlock them. Emergent also made an important architectural choice — not just "generate some code," but covering the full path to hosting, deployment, and getting an app genuinely live. They believe they're building real production software, which means best practices have to be baked in, not bolted on later.
Twin Founders, One Shared Obsession, and a Not-So-Glam Origin Story
Emergent has that rare complementary-founders setup — except here it's literally a twin pairing. Mukund and his brother Madhav (called "Maddie" in the conversation) started programming around age 12. The spark wasn't a fancy bootcamp. It was a Visual Studio CD their dad brought home when the boys wanted a gaming CD. Mukund was annoyed at first. Then it changed everything.
Their paths diverged in a useful way. Mukund leaned into engineering — experience at Google, then leading tech at Dunzo. His brother went research-oriented: a PhD, work in research labs, and an early role on a deep learning team at Amazon. That blend shows up in how they describe building: constant evals, benchmarking, rapid experiments, and big architectural bets.
They also share a bigger, almost old-school ambition. Mukund mentions growing up looking at Bill Gates and Steve Jobs and carrying a question for years: why isn't there a Google or a Facebook from India? It's not said as a slogan. It's said like an itch that never went away. A clue to their competitive streak: they picked a very hard benchmark in coding and reached world number one in under a month. They don't ask "is it possible?" — they ask "why not us?"
For more background on the founders, this write-up adds useful context: profile on the Jha twins and Emergent Labs.
Vibe Coding, Explained Like a Normal Person
"Vibe coding" is a term that went popular fast, and like most fast memes, it gets misunderstood. In simple terms, it means you talk to AI like you'd talk to a developer. You give instructions in plain English. The AI builds. You don't live inside the code — you "vibe check" outputs, then give feedback like: change the color, adjust this flow, add authentication, connect a database.
Mukund sees it as the beginning of a bigger shift, not just in software but in knowledge work broadly. More tasks become "manage agents, verify outputs, refine instructions." Less about writing each line, more about guiding, checking, and iterating. This also explains why early vibe-coding tools felt rough — people complained about loops, broken states, and brittle outputs. Mukund's stance is bullish: the models improve, and teams should bet in the direction of AI from day zero.
What People Are Actually Building on Emergent (Not Toy Demos)
The examples Mukund gives aren't "hello world" apps. People have built portfolio tracking and management tools, an EV charging marketplace, full e-commerce apps, and a heavy share of AI-powered applications. Mukund estimates roughly 60 to 70 percent of apps built on the platform include an AI component — local models, RAG setups, and agent-style workflows.
One app went viral and brought about 100,000 visitors — enough to take the platform down for a while. The idea was simple to explain but hard to ship quickly without good tooling: upload a product photo, describe the product, and get a chatbot you can embed that sells it. That "took the platform down" moment is almost a rite of passage. It's painful, but it's also proof of real demand.
On the "production" question — the big critique of vibe coding — Emergent reports 10,000 to 20,000 apps used daily in production settings. They've also shipped features like authentication, the kind of "boring" requirement that separates a hobby project from a real business tool. For a broader take on what this category means for software roles, this is worth reading: when anyone can code, what becomes the differentiator?
Why Emergent Thought It Could Beat the Incumbents
When people hear "vibe coding," they often assume every product in the category is the same wrapper around a model. Mukund argues that's exactly where early competitors fell short. He describes some tools as "artifact++" — taking what a model like Claude outputs and presenting it cleanly in a web UI. That can work for prototyping. The gap shows up the moment you try to go from "looks good" to "runs in production."
Emergent's approach came down to two big bets. First, coding agent quality has to be top-tier — if the agent can't build well, nothing else matters. Second, infrastructure and feedback loops are everything. Agents are only as good as the signals they get back, and Emergent built the stack from scratch, end-to-end, so they could control the development process and tighten the loop.
Mukund also mentions landing on a multi-agent architecture after hundreds of failures — which is important, because it suggests they didn't get it right early. They just iterated harder than most teams can tolerate. A very practical internal detail shows up here too: they put one number on a big TV and align the whole company around it. The number changes over time (benchmarks, launch dates, revenue, app usage), but the focus stays singular. That's how you stop a fast team from thrashing.
The Growth Playbook: From No Plan to 1 Million Signups
Mukund says growth was almost an afterthought. They were heads down on product and didn't have a real launch plan even four weeks before shipping. Then reality hit: it's a crowded space, and even a strong product dies quietly if nobody tries it.
So they built a small growth team in about three weeks, then treated growth like engineering. They ran roughly 100 to 200 experiments before launch on small budgets, learning what content and what creators would actually move numbers. They studied platform algorithms on X and TikTok — reverse engineering the velocity and engagement thresholds needed for posts to travel. Influencer marketing played a role, but not in the lazy "pay a big creator and pray" way: they categorized creators (nano, micro, macro), studied which ones had worked for similar products, and planned content by geography and time of day.
At one point they were pushing around 500 AI-generated videos a day on TikTok and Instagram. That number sounds absurd until you remember the goal: impressions. Everything else flowed from that.
| Growth Lever | What They Did | Number / Detail |
|---|---|---|
| Launch Budget | Spent on initial launch push | $100,000 |
| Launch Target | Planned signups for launch day | 10,000 |
| Launch Result | Signups achieved (with invite codes) | 20,000+ |
| Experiment Volume | Tests run before launch | 100–200 |
| Content Volume | AI-generated videos per day | ~500/day |
| User Scale | Total platform signups | 1,000,000+ |
| Paying Users | Paying customers early on | ~20,000 |
The other underrated piece was re-engagement. A lot of users try once and drop off. Emergent built CRM loops to bring people back — including basic but critical hygiene like warming up emails so they don't land in spam. Product virality helped too: the "Made with Emergent" badge on sites built on the platform drives about 4 to 5 percent of incoming traffic on its own.
Pricing, PMF, and the "Real Software" Trade-off
Mukund's definition of product-market fit is blunt: it's when you can raise prices without losing users. Emergent started with a low-barrier plan around $10, moved to $20, and has since introduced a $200 plan that — by his account — "a lot of people are buying." The intent wasn't to squeeze early users. It was to maximize trial, then move price once value was proven.
This connects directly to the comparison set in users' heads. If someone was quoted $100,000 by a dev shop and built something on Emergent for a few thousand, they see the platform as a steal regardless of what price it's at. That perception creates real pricing power.
There's also a counterintuitive product bet here: they didn't optimize for the fastest "wow moment." Many people told them to chase speed, but they chose quality software over instant results — believing that long-term winners will be the platforms producing better output, even if it takes slightly longer. One anecdote stands out as a genuine PMF signal: a power user took the team out for drinks, and that user happened to be a CTO from Pixar. It wasn't just usage — it was affection. That's a different category of validation entirely.
For more ideas on what solo builders can do as AI lowers the cost to build, this is a solid read: best AI business ideas for 2026 solo founders.
Contrarian Decisions, Selective Advice, and What's on the Big TV Now
The founders don't present themselves as rule-followers. Mukund says they ignored a lot of standard startup advice — they didn't "launch fast" (they shipped nine months later than the typical push), and they didn't avoid building for scale early. They built for scale from day zero, because their view of the category demanded it.
The bigger lesson isn't "ignore advice" — it's "filter advice." Generic startup rules are averages. If your product, market, and timing are genuinely different, following checklists blindly can push you in the wrong direction entirely.
Looking forward, the north star has shifted again. Instead of revenue being the only scoreboard, they're now heavily focused on end-user usage of apps built on the platform. Revenue should follow value, and value shows up as real apps getting real usage. There's also a quiet but telling internal moment Mukund shares: a teammate asked whether they'd delivered as much value as the company had spent. On launch day, after they estimated developer hours saved, the answer seemed to be yes. That value lens — not just the ARR lens — is where they want to operate.
For a deeper look at Emergent's positioning, this long-form write-up adds useful context: Emergent vibe-coding platform deep dive.
What I Learned From This Startup Story
Having covered dozens of founder stories on this blog, I've noticed a pattern: the ones with the most impressive headline numbers are usually the ones where the interesting story is buried three layers deep. With Emergent, the $10M ARR in two months is the number everyone talks about. But the detail I kept going back to was the bakery owner. Not the Pixar CTO, not the 1 million signups — the bakery owner in the Philippines. Because that's the moment the market told them something they hadn't planned for, and they listened. Compare that to the last few founders I covered who built purely for developer users and hit a ceiling they didn't see coming. The ones who scaled did it because they noticed an unexpected user early enough to change course.
The $10M ARR number is worth pressure-testing. Mukund mentions a $200 plan "a lot of people are buying" and ~20,000 paying customers. Run that quickly: if even 5,000 customers are on the $200 tier and 15,000 on a ~$30 average, you get close to $6.4M ARR from those two groups alone — which means the higher tiers and usage-based components are doing real work. That 15–20% margin critique I'd usually apply to service businesses doesn't land here the same way; software margins on a platform like this can be 60–70%+ once infra costs are controlled. The real question nobody is asking is what their compute cost looks like per app build. That number will define whether the $10M ARR is durable or thin.
Here's the uncomfortable truth about vibe coding platforms that the conversation danced around: the quality of output is still highly dependent on how good the underlying model is — and model providers can change pricing, access, or capability at any time. Emergent built their own stack end-to-end, which is smart, but they're still sitting on top of model infrastructure they don't fully control. Nobody asked what happens if their primary model provider raises API costs 40% next year. That's not a hypothetical. It's happened before in this space. The moat in vibe coding right now isn't the interface — it's the feedback loops, the eval infrastructure, and the accumulated training data from real app builds. If Emergent has that, they're harder to replicate than they look from the outside.
Honestly? This platform is worth trying if you have a real product idea and have been sitting on it because you can't code or can't afford a dev team. That gap — between having an idea and being able to ship it — is exactly what Emergent was built to close. But if you need a production app with complex integrations, compliance requirements, or enterprise-level reliability on day one, I'd treat this as a powerful prototype tool first and evaluate seriously before going fully live. The authentication features are a good sign they're maturing fast, but "production-ready" still means different things to a bakery owner and a Series B SaaS team. Know which one you are before you commit.
⚡ Key Takeaways
- Production-first beats demo-first: Building for real production from day one was Emergent's core differentiator in a crowded space.
- Listen to unexpected users: The bakery owner moment redefined the total addressable market. Don't dismiss users who "shouldn't" be using your tool.
- Treat growth like engineering: 100–200 experiments before launch, creator categorization by tier, platform algorithm analysis — this is a system, not a campaign.
- One number on the wall: Keeping the whole company aligned around a single changing metric prevents a fast team from losing focus.
- Filter advice, don't follow it: Generic startup rules are averages — know when your situation is different enough to deviate.
- Value before price: Emergent maximized trial with low prices, only raising them once users felt the product was a steal compared to alternatives.
- The moat is feedback loops: Not the UI, not the prompt — the eval infrastructure and end-to-end stack control is what makes this defensible.