The Fastest Wealth Creation in History: $0 to $2.7 Billion in 3 Years

0:00 Four college kids just became billionaires faster than Mark Zuckerberg. Their company went from zero to three billion dollars in revenue in three years. It was the fastest growing software company in history. So, why did they just sell it to Elon Musk for 60 billion dollars? This is Cursor and the timeline from founding it to exit is the fastest in software history. Let me show you what's going on and before we start, please don't forget to subscribe. It really helps the channel. Thank you so much.

0:38 In case you've been living under a rock, Cursor is an AI code editor. It is a fork of Microsoft Visual Studio Code with AI deeply integrated. You write code, Cursor suggests completions, rewrites functions, debugs errors. For technical users, it feels instantly familiar embedding powerful AI directly into your workflow rather than acting as a simple chatbot overlay. And it lets you easily toggle between different AI models. The company was founded in 2022 by four science students, Michael Truel, Suhas Asif, Aman Sanger, and Arvid Lunemark.

1:14 They spent nearly a year building AI tools for mechanical design before pivoting to software development tools mid-2022. They launched Cursor in March '23. As one of the pioneers of vibe coding, it went viral immediately. A massive word-of-mouth boom among developers, viral social media demonstrations, and endorsements from tech heavyweights propelled it to over a hundred million in annualized revenue with almost zero marketing spend. At the time, everyone at YC was using it and the trajectory is absurd. One million dollar AR in 2023, 100 million by January 2025. That is 21 month and at the time it was the fastest any SaaS company has ever reached a hundred million in revenue. Then, it kept doubling. 500 million by June 2025, 1 billion by November, 2 billion by February 2026, 3 billion by late April.

2:13 Management is guiding to more than 6 billion by the end of the year. Revenue doubled roughly every two months for most of 2025. By early 2026, Cursor had 67% of the Fortune 500 using the product. Uber, Adobe, Google, Nvidia, Jensen Huang, Nvidia CEO, publicly called Cursor his favorite enterprise AI service. But, maybe even more importantly, Cursor hit 2 billion ARR with about 300 employees. That is 6 to 10 million revenue per employee. For context, the Slack IPO generated $300,000 per employee. Snowflake was about $400,000. The best public SaaS companies today operate maybe at a 5 to $600,000 per employee. Cursor is 20 times more efficient than the benchmark we've been using for the last decade.

3:08 And the implications for venture are huge. It means that the right size for an AI native company at 1 billion in revenue might be 100 to 300 people, not 1 to 3,000. That changes how you model reserves. It changes how you think about ownership dilution. It changes what a normal ESOP looks like. If you're still underwriting AI companies using burn multiple and head counts projections from the 2010s SaaS playbook, you're pricing the wrong risk. The rise of companies like Cursor have shifted VCs to a new question. Not how much are they burning relative to growth, but how much revenue are they generating per person and is that sustainable without the model providers? However, this brings us to a dependency problem. Cursor's product works by routing code to large language models. For most of 2024 and 2025, that meant paying Anthropic and OpenAI for every API call. As of January 2026, Cursor had -23% gross margin. For every $100 in revenue, they spent $123 on compute. They were subsidizing growth with venture capital. And this is the trap. The faster Cursor grew, the more it lost. Revenue doubled every 2 month in 2025, and so did the compute bill.

4:30 Last year, Cursor burned $900 million. Now, Cursor knew that this was a problem. In October 2025, they launched Composer, a proprietary model. The goal was to stop paying Anthropic and OpenAI and route inference in their own stack. It did work, and by April 2026, Cursor reached what TechCrunch called slight gross margin profitability. Enterprise contracts were now profitable, but individual developer subscriptions were still losing money. But even with this progress, margins were still thin. This was not an 80% gross margin SaaS business. Cursor was a company barely breaking even on unit economics while scaling at warp speed. And when you're burning so much cash, you need capital. So, the margins became a problem when they tried to raise. In early 2026, Cursor went out to raise $2 billion at a $50 billion valuation. With such good growth, the pitch was strong, $3 billion in ARR, and the narrative of the fastest growth in SaaS history. But Iconic, one of the leading growth stage investors, who was set to lead the round, decided to pass, and so did other late-stage funds. The reason, they just deployed into OpenAI and Anthropic.

5:45 According to Crunchbase Q1 report, OpenAI raised 122 billion, Anthropic 30 billion, xAI 20 billion, and Waymo 16 billion. Collectively, 65% of all global venture investments in that quarter. I did a specific video about this. If you are a late-stage fund and you just wrote a 500 million check to Anthropic, you're not writing another 500 million-dollar check to a company paying Anthropic for compute while competing with Anthropic's Claude Code product. Claude Code hit itself a 2.5 billion run rate by February 2026 with 300,000 businesses customer. GitHub Copilot has 4.7 million paid subscribers and 37% market share.

6:29 So, Cursor was stuck. They needed capital to keep scaling, but the late-stage market had concentrated all its capital into the foundation model layer. There was no room for a well-funded competitor in the application layer. What's more, investors feared that Cursor would struggle to maintain its moat against those frontier labs building their own agentic coding assistants. But just as Cursor lined up 2 billion from Nvidia, Andreessen Horowitz, and Thrive, SpaceX showed up with an option to acquire the company for 60 billion. That option looks better than the alternatives. But why on earth would SpaceX want Cursor?

7:10 Well, in February 2026, Elon Musk merged xAI into SpaceX. The combined entity was now valued 1.25 trillion dollars. But xAI had a problem. It was 7 month behind Anthropic, Google, and OpenAI. Buying Cursor gives SpaceX three things. One is distribution. Cursor has millions of developers and 67% of the Fortune 500 companies. That is instant enterprise access for xAI's models. Two is training data. Every interaction a developer has with cursor generates training data for coding models. That is 3 years of the highest quality coding data in the market. And three is compute monetization. Cursor's biggest constraint has always been compute.

7:55 SpaceX has Colossus, one of the largest AI clusters in the world. Some of it is sitting idle. Routing Cursor's inference to Colossus turns idle capacity into revenue and improves Cursor's margins. For SpaceX, this is a $60 billion bet that coding is the wedge to catch up in AI. For Cursor, this solves the dependency problem and gives them access to compute they could not afford to build themselves. The acquisition is paid in SpaceX class A stocks at the IPO price. The IPO is June 12 and the deal closes about 30 days later in July. If SpaceX lists at 1.75 trillion and the stock trades down to 1.5 trillion, then the 60 billion shrinks proportionately.

8:41 Cursor's investors are not getting 60 billion in cash. They are converting one stock into another. Of course, if the stock goes up, then so does the acquisition price. And SpaceX itself is burning cash. According to their S1, they had 19 billion revenue in 2025 but lost 5 billion. In Q1 26, the revenue was almost 5 billion but the loss was 4.3 billion. The only profitable segment is Starlink. The AI division is the loss engine. XAI spent 12.7 billion in capex last year while generating only 3.2 billion in revenue. Cursor lost 900 million last year. SpaceX is acquiring a cash burning business to bolt onto another cash burning business. The bull case, of course, is that margins improve and the combined entity scales profitably. The bear case is that Cursor's growth stalls, margins stays thin, and SpaceX shareholders end up diluted 3.4% for a business that never breaks even. It all sounds crazy, and maybe to some people a little bit like a house of cards. So, what does this all mean for venture? Well, firstly, Cursor is the fastest wealth creation event in software history. Four founders went from zero to 2.7 billion each in 3 years. That is faster than Zuckerberg, faster than the Google founders, faster than anyone, but it only happened because they sold. And they only sold because late-stage VCs stopped writing checks. If Iconic and the other mega funds had been willing to back Cursor at 50 billion, this would have been series E, not an acquisition. Cursor would still be independent, and the founders would still be in an uphill battle to fix margins and fight the gods like Anthropic and OpenAI.

10:30 >> >> Instead, well, they are billionaires. Congrats. This is clear evidence for venture that timeline compression has broken the old playbook. I've been saying it for years. Companies are going from zero to 60 billion in 3 years instead of 10. As I'm shooting this video, Corgi went from 1 billion to 2 billion valuation in only 3 weeks. That's a YC startup from 2024. But the window is narrow, and the only way to capture that value is to concentrate early and sell before >> >> the music stops.

11:02 For funds that build diversified AI portfolios, most of those companies will not be Cursor. They will either run out of capital before they fix the margins, or they will get crushed by foundation model providers building competing products. Winners will be the funds that concentrated at the top two or three outcomes per vintage and exited before the capital markets closed. Cursor found the exit. Most companies will not, but all it takes is one. So, who will be next? Let's wait and see. Thank you for watching. Don't forget to subscribe to the channel. Thank you and see you next week.