At 25x ARR, why is Cursor so cheap?
And: Bad times in Chinese real estate
Welcome to Cautious Optimism, a newsletter on tech, business, and power.
📈 Trending Up: Bluesky? … Bluesky? … government subsidy … fire safety startups … Anduril in Ohio … Mistral-AFP, Axios-OpenAI …
Data centers in space: I sat down with the co-founder and CEO of Lumen Orbit yesterday (interview coming over on TWiST) for a chat. Want to feel bullish? The company, which wants to build incredibly huge solar arrays in space to power huge orbiting data centers, had such a high-demand $11 million Seed round that it added another $10 million afterward in a SAFE. That’s a lot of cheddar for a very small, very new space startup that has just about the biggest goals I have ever encountered at a new tech company.
Nice. (I believe Phillip Johnston shared the SAFE cap, but I need to confirm against the transcript before I shoot my mount off.)
📉 Trending Down: The shadow fleet … my genius, after buying a Switch 1 in December … free speech online … fintech exit valuations? …
Why is Cursor so cheap?
The venture market for AI fundraising is hot — and especially so here in the United States. Thanks to the investor demand, there have already been at least ten nine-figure investments into global AI companies this year.
One of those deals was AI coding tool Cursor raising $105 million, a huge chunk of change. And it did so at a very, very interesting revenue multiple. Here’s the Times with the critical riff:
But [Michael Truell, co-founder and CEO] said that Anysphere [the startup behind Cursor] had sought to keep a lid on costs, including spending on computing resources. He also argued that his company’s valuation was within historical norms. (It’s about 25 times Anysphere’s $100 million in annual recurring revenue, in line with other A.I. start-ups.)
$100 million ARR for a company founded in 2022? That’s insanely impressive, full stop. But 25x ARR? In 2025? For a super-hot AI startup with bonkers momentum? That seems — don’t yell at me — a little cheap?
As CO wrote while discussing reports that Anthropic was hunting $2 billion more at a $60 billion valuation earlier this year:
Anthropic wants to raise $2 billion at a $58 billion pre, $60 billion post-money valuation. It’s annualized revenue “recently hit about $875 million” per the WSJ. That means Anthropic is looking for a roughly 68.5x run rate multiple to start the yera.
OpenAI raised last year at a $157 billion valuation and revenue expectations of $3.7 billion for the 2024 calendar year. It likely ended on a faster run rate cadence, but even using a more conservative price/sales denominator, OpenAI is worth around 42.4x its revenues. That’s much cheaper!
A fun question to ask yourself: Would you rather invest half your 401k in OpenAI at 42.4x or Cursor at 25x? The latter, right?
So, what’s going on? A few things could be true: