# How Lovable Hit $100M ARR in 8 Months (Then $500M in 19)

> Launch to a $500M run rate in 19 months with 146 people. The engines behind it, and the cracks under it.

- This is the markdown version of: https://www.productgrowth.blog/p/how-lovable-dev-hacked-their-growth (cite that URL)
- Author: Rishikesh Ranjan · Published: May 5, 2025 · Updated: Jun 11, 2026
- Type: Teardown
- Tags: Case Study, AI, Product-Led Growth
- ~2201 words

---

In February 2026, Lovable [**added $100 million of ARR in a single month**](https://techcrunch.com/2026/03/11/lovable-says-it-added-100m-in-revenue-last-month-alone-with-just-146-employees/?utm_source=productgrowth.blog). Not over a year. One month.

The company that did it employs 146 people. The product was 15 months old at the time. There is no enterprise sales army, and for most of the run there was no paid marketing either.

I first wrote this teardown in May 2025, when the headline number was $17M ARR in 90 days. That version aged like milk. The story is now 30x bigger, so I tore it down and rebuilt the whole thing: the timeline, the growth engines, the pricing, and the cracks the LinkedIn victory laps skip.

## TL;DR

**Lovable went from a November 2024 launch to $100M ARR in eight months, then to a $500M run rate by June 2026, by converting an open-source audience into a pre-sold launch, turning all 146 employees into a distribution channel, and refusing to buy growth until $300M ARR.**

- **Trust is pre-sellable: **GPT Engineer had 50,000+ GitHub stars before Lovable existed. Launch day was not day zero.
- **Your team is a channel: **every employee builds on Lovable and posts about it, with the first two hours of engagement coordinated internally. They call it beeswarming.
- **Free weekends print demand: **one free day on March 8, 2026 drove 500,000 projects, about 2.5x a normal day.
- **Sequence your spend: **under 10% of first-year growth came from paid channels. Ads only started once organic retention was proven.
- **It is not all clean: **site traffic fell roughly 40% from its 2025 peak while ARR quadrupled, and the security record is bad. Both halves of that sentence matter.

## Lovable's ARR timeline: $1M to $500M in 19 months

The raw numbers first, because the compression is the story. Every figure here is company-reported; I've linked a primary source for each milestone in the sections below. Quick refresher on [what ARR does and doesn't tell you](https://www.productgrowth.blog/p/annual-recurring-revenue-arr) if you need it.

| When | Milestone | Context |
| --- | --- | --- |
| Jun 2023 | GPT Engineer open-sourced | Top trending repo worldwide, 50K+ stars |
| Oct 2024 | $7.5M pre-seed | Still called GPT Engineer |
| Nov 2024 | Lovable launches | #1 on Product Hunt and Hacker News |
| Jan 2025 | $10M ARR | Two months post-launch, 15 employees |
| Jul 2025 | $100M ARR | $200M Series A at $1.8B; 2.3M users |
| Nov 2025 | $200M ARR | About 8M users, enterprise push begins |
| Dec 2025 | $330M raised | $6.6B valuation |
| Feb 2026 | $400M ARR | $100M added in February alone, 146 employees |
| Jun 2026 | $500M run rate | 50M projects built, about 1M new per week |

![Lovable annual recurring revenue run rate by month: $0 at launch in November 2024, $10M in January 2025, $100M in July 2025, $200M in November 2025, $400M in February 2026, $500M in June 2026.](https://www.productgrowth.blog/media/posts/how-lovable-dev-hacked-their-growth/01-arr-growth-curve.webp)

Do the math on the team. $400M across 146 people works out to [**$2.77M of ARR per employee**](https://techcrunch.com/2026/03/11/lovable-says-it-added-100m-in-revenue-last-month-alone-with-just-146-employees/?utm_source=productgrowth.blog). A good SaaS company runs $200K to $400K. Lovable sits at roughly ten times that.

For slope context: Wiz took about 18 months to reach $100M ARR and held the record. [Cursor](https://www.productgrowth.blog/p/how-cursor-ai-hacked-growth) did it in roughly a year and looked untouchable. Lovable did it in eight months and [called itself the fastest-growing software company ever](https://sifted.eu/articles/lovable-hits-100m-arr?utm_source=productgrowth.blog). Nobody has produced a counterexample yet.

## From weekend repo to category king

In June 2023, Anton Osika pushed a small Python project to GitHub over a weekend. [GPT Engineer](https://github.com/AntonOsika/gpt-engineer?utm_source=productgrowth.blog): describe the software you want, the model writes the code. It became the top trending repo in the world within days and now sits above 50,000 stars.

Osika and CTO Fabian Hedin incorporated in Stockholm, spent a year turning the CLI experiment into a web product a non-developer could use, raised a [$7.5M pre-seed](https://tech.eu/2024/10/07/lovable-raises-7-5m-for-gpt-engineer/?utm_source=productgrowth.blog) in October 2024, and launched Lovable that November. The launch hit #1 on Product Hunt and #1 on Hacker News.

Two months later: [$10M ARR with 15 employees](https://lovable.dev/blog/2025-01-29-zero-to-10m-arr-in-2-months?utm_source=productgrowth.blog).

Read the sequence again. The audience existed before the product did. By launch day, tens of thousands of developers already knew Osika, had starred the repo, and had argued about it in comment threads. That's not a cold start. That's a pre-sold launch.

> **Steal this:** Open source worked here as a trust pre-sale. If you can give away a credible v0 of your idea and grow an audience on it, launch day stops being day zero. The repo is the waitlist.

## The growth engines behind Lovable's revenue

Five engines, all compounding. None of them is paid media.

### Engine 1: a demo you can screenshot

Lovable's aha moment is stupidly simple: type *build me a booking app for my salon*, watch a working full-stack app appear in minutes. The output is shareable by default, so users post before-and-after demos on X, TikTok and YouTube without being asked. The product creates its own ad inventory.

The team amplified this early with [co-marketing partnerships with Supabase, Replicate and Resend](https://lovable.dev/blog/2025-01-29-zero-to-10m-arr-in-2-months?utm_source=productgrowth.blog), which put Lovable in front of developer audiences it didn't have to rent.

> “When your product's output is shareable, your users are your media plan.”

### Engine 2: beeswarming, 146 people as one media company

Osika spent years posting raw numbers, failed experiments and hiring photos before any of this worked. So when Lovable launched, people trusted the person before they evaluated the product.

The company then systematised founder-led content into what it calls beeswarming: [every employee builds side projects on Lovable and posts them](https://www.the-ai-corner.com/p/lovable-growth-playbook-0-to-400m-arr-14-months?utm_source=productgrowth.blog), and an internal channel coordinates comments and reposts within the first two hours of anything going live. The first two hours are what the algorithms judge.

> **Steal this:** Founder-led content tops out at one founder's reach. Get every employee shipping on your own product and posting the results, then coordinate the first two hours of engagement. The feed does the rest.

### Engine 3: free weekends that manufacture FOMO

Lovable runs periodic free days: early 2025, summer 2025, International Women's Day 2026. On [March 8, 2026, users built or updated 500,000 projects](https://techcrunch.com/2026/03/11/lovable-says-it-added-100m-in-revenue-last-month-alone-with-just-146-employees/?utm_source=productgrowth.blog) against a typical day of around 200,000.

> All I see is people talking about the Lovable free day. That is something you cannot pay for.
> — Elena Verna, Head of Growth at Lovable

Verna joined in mid-2025 and her fingerprints are all over this motion. Her read on the channel shift: five years ago organic marketing meant SEO, [now it's all about social](https://www.the-ai-corner.com/p/lovable-growth-playbook-0-to-400m-arr-14-months?utm_source=productgrowth.blog). A free day costs compute. An equivalent burst of attention from ads would cost millions, and people would trust it less.

### Engine 4: zero paid spend until $300M ARR

Total marketing spend on the way to $30M ARR: [about $2M](https://www.the-ai-corner.com/p/lovable-growth-playbook-0-to-400m-arr-14-months?utm_source=productgrowth.blog). Under 10% of first-year growth came from paid channels. Lovable only turned on real paid acquisition in early 2026, past $300M ARR, and even then it went for niche creator sponsorships over platform ads.

The sequencing is the lesson. Paid spend poured on before word-of-mouth works just buys churn at a markup. Lovable bought reach only after organic had proven that people stick around and expand.

> **Steal this:** Treat paid as an amplifier, not an engine. If switching ads off would stop your growth, you don't have product-market fit yet. You have a leased audience.

### Engine 5: a north star that counts both sides

Lovable's north-star metric is Daily Active Apps: apps being actively built plus apps receiving real traffic. One number that captures both the builder and the audience of what they built. Vanity signup counts can't hide in it.

Verna's growth team also runs on a [95% innovation, 5% optimization split](https://www.the-ai-corner.com/p/lovable-growth-playbook-0-to-400m-arr-14-months?utm_source=productgrowth.blog), the inverse of most growth orgs. The team ships actual product (Shopify integrations, voice mode, agent workflows) instead of A/B testing button colours.

## Lovable pricing: credits, top-ups and the enterprise climb

The model is freemium with usage-based credits. Free gets you about 30 credits a month, Pro starts around $25 a month, Business around $50, Enterprise is custom. Since the [Agent launch in July 2025](https://lovable.dev/blog/agent?utm_source=productgrowth.blog), credits are complexity-weighted, so the price of an action tracks the compute it burns.

Top-ups absorb bursty builders without forcing a plan upgrade, and that usage layer is why Lovable reports [net dollar retention above 100% and 85% day-30 paid retention](https://www.the-ai-corner.com/p/lovable-growth-playbook-0-to-400m-arr-14-months?utm_source=productgrowth.blog). The customers who stay spend more over time, which is how ARR quadrupled while traffic fell (more on that below).

Up the stack, Klarna, HubSpot and Photoroom are named customers, and Osika claims [more than half the Fortune 500](https://techcrunch.com/2026/03/11/lovable-says-it-added-100m-in-revenue-last-month-alone-with-just-146-employees/?utm_source=productgrowth.blog) touch Lovable in some capacity. Take that line as marketing arithmetic; one team on a free plan counts.

This is the credits-eat-seats story I keep coming back to. If you're pricing an AI product right now, [my breakdown of credits vs seats vs outcomes](https://www.productgrowth.blog/p/ai-pricing-credits-vs-seats-vs-outcomes) pairs well with this one.

## The cracks in the $500M story

Now the part the victory laps leave out. You'd be right to distrust 2,500 words of pure admiration, so here are the four problems I'd be staring at if I ran growth there.

### Run rate is not ARR

The $500M figure is an [annualised run rate](https://www.techtimes.com/articles/318072/20260609/lovable-says-it-hit-500-million-run-rate-vibe-codings-maintenance-test-still-looms.htm?utm_source=productgrowth.blog): take a hot recent period, multiply it forward. It's self-reported and unaudited, and in a company moving this fast it flatters the trend. None of that makes the growth fake. It makes the precision fake.

> **Before you quote $500M:** Run rate annualises a recent period. ARR proper is the contracted recurring base. For a business with monthly plans and promo spikes, the two can diverge a lot. Quote the number, but say which one it is.

### Traffic fell 40% while revenue quadrupled

In September 2025, [Barclays analysts flagged](https://dnyuz.com/2025/09/26/ai-vibe-coding-tools-may-be-going-from-boom-to-bust-new-data-shows-heres-why/?utm_source=productgrowth.blog) that Lovable's web traffic was down about 40% from its summer peak. Vercel's v0 was down 64%, and Bolt (I've done [that teardown](https://www.productgrowth.blog/p/how-bolt-new-hacked-its-growth) too) was down 27%. Their question: had vibe coding already peaked?

Then ARR went from roughly $130M at the time of that report to $500M. Both facts hold. The tourists left, and the people who stayed are paying much more. That's what NDR above 100% looks like from the outside.

The honest read: durability is still unproven. Barclays' other point was that month-to-month subscribers can leave as fast as they arrived, and Lovable has never disclosed [churn](https://www.productgrowth.blog/p/churn-rate). When a company shares its NDR but not its logo churn, the silence is data too.

### The security record is bad, and the response was worse

In February 2026, a researcher showed a single [Lovable-built app exposing data on 18,000 users](https://www.theregister.com/2026/02/27/lovable_app_vulnerabilities/?utm_source=productgrowth.blog). March was worse: a [BOLA vulnerability in Lovable's own API](https://thenextweb.com/news/lovable-vibe-coding-security-crisis-exposed?utm_source=productgrowth.blog) let anyone with a free account read other users' source code, database credentials and AI chat histories.

The response did the real damage. The report was closed without escalation, the hole stayed open for 48 days, and the company's first public move was denial before the eventual apology. When your customers can't read the code they ship, your security bar has to be higher than a dev tool's, not lower.

### The maintenance test nobody can answer

50 million projects have been built on Lovable. How many still work, get updates, and have an owner who can debug them a year from now? [Nobody outside the company has that data](https://www.techtimes.com/articles/318072/20260609/lovable-says-it-hit-500-million-run-rate-vibe-codings-maintenance-test-still-looms.htm?utm_source=productgrowth.blog), and Lovable isn't volunteering it.

Building software was always the easy half. The category's future rests on whether non-technical builders can maintain what they shipped, or whether Lovable can make maintenance as promptable as creation. My money says maintenance becomes the actual product by 2027.

## What to steal from Lovable's playbook

Strip out the venture budget and the once-in-a-cycle category timing, and five moves transfer to almost any product:

- **Pre-sell trust with a giveaway: **open-source a credible slice of your idea, publish the research, or build in public. Make launch day a harvest instead of a cold start.
- **Make the output the ad: **if users can screenshot the value, design for that screenshot. Gate nothing that fuels sharing.
- **Turn the team into distribution: **every employee using the product in public, with the first two hours of engagement coordinated. Costs nothing but culture.
- **Sequence paid after retention: **Lovable spent $2M getting to $30M ARR. Prove people stay before you pay for more people.
- **Count what compounds: **a north star like Daily Active Apps measures builders and their audiences at once. Pick the metric that captures your loop, not your funnel.

None of these need Lovable's funding. All of them need a product people would talk about anyway. That part you can't growth hack.

![Four lessons to steal from Lovable's playbook: open source as a trust pre-sale so launch day isn't day zero; beeswarming where every employee ships and posts instead of broadcast founder content; paid ads as amplifier not engine; and credit pricing that meters the moment of value.](https://www.productgrowth.blog/media/posts/how-lovable-dev-hacked-their-growth/02-steal-this-card.webp)

## FAQ: Lovable's ARR, valuation and funding

#### What is Lovable's ARR in 2026?

Lovable reported a $500 million annualised revenue run rate in June 2026, up from $400 million in February 2026 and $100 million in July 2025. The figure is company-reported and annualises recent revenue rather than reflecting an audited yearly total.

#### How fast did Lovable reach $100M ARR?

About eight months. Lovable launched in November 2024 and announced $100 million ARR in July 2025, claiming the fastest software ramp on record. Cursor took roughly 12 months to the same mark and Wiz about 18.

#### What is Lovable's valuation?

Lovable was valued at $6.6 billion in December 2025 after a $330 million round, up from $1.8 billion at its $200 million Series A in July 2025.

#### How much funding has Lovable raised?

Disclosed rounds total over $530 million: a $7.5 million pre-seed in October 2024, a $200 million Series A led by Accel in July 2025, and a $330 million round in December 2025.

#### Who founded Lovable?

Anton Osika (CEO) and Fabian Hedin (CTO) founded Lovable in Stockholm in 2023, building on Osika's open-source project GPT Engineer, which drew more than 50,000 GitHub stars.

My prediction: Lovable crosses a $1B run rate before the end of 2027, and the thing that gets it there won't be vibe coding demos. It will be boring enterprise control: SSO, audit logs, security reviews that actually pass, and an agent that maintains apps instead of just birthing them. The maintenance bill always arrives. The winner is whoever makes paying it feel like typing a prompt.

Which of the five engines would move the needle for your product? Hit reply and tell me. And if ramps like this are your thing, the [Gamma](https://www.productgrowth.blog/p/gamma-growth-teardown-100m-arr) and [ElevenLabs](https://www.productgrowth.blog/p/how-elevenlabs-hit-330m-arr) teardowns are the natural next reads.

---

All posts: https://www.productgrowth.blog/archive · Site: https://www.productgrowth.blog