Michael Truell: How Cursor Grew to $1B ARR in 24 Months

Cursor was a slow burn for 22 months, then went vertical. Four MIT founders, zero marketing spend to $100M ARR, and a $400M-to-$29B valuation run. The teardown — and the parts you can't replicate.

TL;DR

• Cursor reached $100M ARR in 22 months and $1B ARR in 24 months from $1M — the fastest SaaS company to either milestone on record.
• It was not an overnight hit. The first 22 months (March 2023 – July 2024) were a near-flat slow burn to ~$48M ARR. The vertical takeoff was compressed into the 13 months after.
• Co-founder Aman Sanger has stated publicly that Cursor spent zero on marketing through $100M ARR. The growth came from a structurally better product and a founding team that treated itself as a media property.

The Numbers

Supporting metrics: 1M+ daily active users by April 2025; enterprise revenue share climbed 25% → 45% → 60% across 2024–2025; valuation ran from $400M to $29.3B in 15 months. Anysphere — the company behind Cursor — was founded in January 2022 by four MIT undergrads: Michael Truell, Sualeh Asif, Arvid Lunnemark, and Aman Sanger.

What they did differently

Move 1: Take the structurally harder path

In mid-2022, Anysphere made the decision the entire company rests on. GitHub Copilot was a plugin living inside VS Code. Cursor could have been a plugin too — faster to ship, easier to distribute. Instead the team forked the entire VS Code editor, betting that owning the whole product surface would matter more long-term than easy distribution.

That choice is why Cursor could ship multi-file editing, then an agent layer, then its own in-house model — capabilities a plugin architecture structurally cannot reach. It also meant the product had to be genuinely better to overcome the harder install. It was. Engineers at OpenAI, Midjourney, and Shopify started posting Cursor screenshots unprompted, and the product spread on that signal alone. In August 2024 a Cloudflare VP posted a video of his daughter building a working chatbot:

My 8-year-old daughter built a chatbot using Cursor and Cloudflare. It took her about 45 minutes... she didn't write a single line of code herself.

— Ricky Robinett, X, August 2024 (~2.7M views, reshared by Tobi Lütke and Andrej Karpathy)

This is what "zero marketing spend through $100M ARR" actually means: the product was good enough that credible people did the marketing for free.

Move 2: Treat the founding team as a media property

Through 2023–2024, all four founders posted product updates and demos to developer Twitter on a weekly cadence — in personal voice, not corporate tone — for 18 months before any of it mattered.

The payoff event was Lex Fridman episode #447 on October 6, 2024: all four founders, two hours and 38 minutes. A long-form technical conversation does something short posts can't — it converts developer credibility into investor and executive credibility, reaching the exact audiences who price Series B rounds. In the four months after the episode, Cursor's valuation went up 6.5×.

The deeper return came in February 2025, when Andrej Karpathy coined "vibe coding" and reached for an example:

There's a new kind of coding I call "vibe coding"... I'm building a project... it's not really coding — I just see stuff, say stuff, run stuff, and copy-paste stuff, and it mostly works.

— Andrej Karpathy, X, February 2025, naming Cursor Composer + Sonnet

You don't get a KOL to name you by asking. You make sure that when the KOL needs an example, you're the obvious one. Eighteen months of mind-share investment is what made Cursor that obvious example.

Move 3: Never let a milestone fire alone

Cursor's funding announcements are a study in compounding coverage. A solo "$X raised" headline buys 3–5 days in the capital press. A "$X raised + $Y ARR + new product" headline buys the same window across capital press, developer press, SaaS press, and business press at once.

So Cursor never let a milestone fire alone. The Series C in June 2025 was announced with the disclosure that ARR had crossed $500M. The Series D in November 2025 — a $2.3B raise at a $29.3B valuation — was announced with the $1B ARR milestone, and arrived two weeks after Cursor 2.0 shipped a multi-agent architecture. Same announcement budget, three to four times the coverage surface, every single round.

What you can copy

1. Take the structurally harder path when it buys long-term control. Being a plugin is faster; owning the whole product surface is what lets you ship the things competitors structurally can't. Choose the path that compounds.
2. Post in founder voice, on your buyers' platform, before anyone is watching. Cursor's founders did 18 months of weekly developer-Twitter posts with no audience. The cadence is the moat — it can't be bought later.
3. Do one long-form podcast that forces real depth. A 2.5-hour technical conversation converts niche credibility into investor credibility. It only works if the founders can survive the questioning — so build the depth first.
4. Earn the default, don't ask for the mention. You can't make a KOL cite you. You can make sure that when they need an example in your category, you're the most obvious one on the board.
5. Bundle every funding announcement with an ARR milestone, a product launch, or both. Never spend an announcement on a single headline.

What probably won't work for you

The slow burn requires investors who'll tolerate it. Cursor's backers — the OpenAI Startup Fund, a16z, Thrive — let the company stay quiet through 22 months of near-flat growth without external pressure. Most investors would have forced a premature push. If your cap table can't absorb a long latent phase, this exact curve isn't available to you.

"Zero marketing" is a product bar, not a growth tactic. Cursor spent nothing on marketing because the product was good enough that senior engineers evangelized it for free. If your product doesn't earn unprompted word-of-mouth from credible users, the answer is not "copy Cursor's marketing" — it's "fix the product." Zero-spend growth is a symptom of product quality, not a substitute for it.

The category moved at 5× normal speed. AI coding tools compressed five years of category formation into 18 months. The Lex episode, the Karpathy post, the valuation re-pricings — all of it rode a wave most categories never get. The moves transfer; the tailwind doesn't.

The founders could survive 2.5 hours of expert questioning. All four are technically deep. A weaker team puts the Lex episode — the catalyst — at risk. Don't book the long-form slot until the depth is real.

Sources & references

• Anysphere raises $60M Series A at $400M valuation — TechCrunch
• "My 8-year-old daughter built a chatbot" — Ricky Robinett, X, August 2024
• Lex Fridman Podcast #447 — Cursor team — the founder-as-IP catalyst
• Andrej Karpathy's "vibe coding" post — X, February 2025
• Anysphere acquires Supermaven — TechCrunch
• Cursor nabs $9.9B valuation, soars past $500M ARR — TechCrunch
• Cursor's $29.3B Series D — CNBC
• Full timeline: The Cursor Growth Story — every round, launch, and inflection charted

This case study is part of GrowthHunt's growth teardown series. Compare it with the Lovable teardown — the same category, the opposite curve shape — or watch the fastest-growing AI repos update weekly on GrowthHunt Velocity.

More from the growth playbook.

See it run in the wild.