Daily Grind July 24, 2025: The new fastest growing startup in history

Welcome to Tech's morning newsletter, where you get one headline, one page of a great book, and one question to start your day.

Good morning, and welcome to The Daily Grind for Thursday, July 24.

Today, we have a NEW fastest growing company in history, and it signals an inflection point in the way companies are built.

Let’s get right into it!

📰 One Headline: We have a new fastest growing startup in history

In February 2025, Cursor, the leading AI code-editor for engineers, became the fastest-growing startup in history, reaching $100 million in ARR in just 12 months.

Today, just 5 months later, we have a new record-holder: Cursor’s consumer-friendly rival, Lovable.

Lovable is a Sweden-based vibe coding tool that allows non-developers to spin up websites and apps in a fraction of the time.

Unlike Cursor, which caters to professional developers, Lovable has given the power of programming to millions of non-programmers. It is the equivalent of Canva in the design world: beginner-friendly yet surprisingly powerful.

From the article:

Lovable doesn’t pretend to cater to the pros. Its fans are tinkerers, designers and entrepreneurs. “Developers are really important, but they’re only 1% of the market,” [Accel partner Ben] Fletcher says.

Lovable has been on a run. Last week, the startup announced a new $200 million fundraising round that made it Europe’s latest Unicorn.

They achieved that milestone, plus the $100 million ARR mark today, with just 45 full-time employees.

It’s not a coincidence that a vibe-coding startup has unlocked new levels of explosive growth with such a small team—that’s exactly what Lovable and its competitors promise to do for their customers.

But the ability to grow 10x faster also comes with the risk of 10x the competition and, potentially, a 10x faster downfall.

Lovable is now booking around $1 million a day in subscriptions, but competition is fierce. Its focus on portfolio sites and simple prototypes puts it on a collision course with an earlier generation of unicorns like Figma and website builders such as Wix and Squarespace. All these companies are also building their own AI tools. Figma launched a code generator earlier this year, and in June, Nasdaq-listed Wix spent $80 million to buy a six-month-old AI coding startup.

Another challenge is Lovable’s reliance on the large AI models, which are increasingly competing against its customers. Cursor and Windsurf have recently grappled with this reality. Lovable’s hightened profile is likely to cause riffs as well:

Lovable relies on the same set of underlying AI models, notably Anthropic’s Claude. Lovable is spending millions of dollars a month on these models to power its coding. Some other startups are even heavier spenders, which has put Anthropic on track to make a reported $4 billion in revenue this year. Now Anthropic, last valued at $60 billion, is selling its own code tool directly.

Then there is the question of building in Europe. With the EU’s tight restrictions on AI, can an AI startup truly thrive?

Despite the headwinds, the tailwinds for vibe-coding tools like Lovable are too strong. This milestone signals a massive change in the way startups are built. You no longer need hundreds of employees to work for years to reach Unicorn status.

I expect Lovable’s record to be broken again in just a few months. Any bets on who it will do it?

Follow up from Friday:

📚 One Page: The Experimentation Machine by Jeffrey J. Bussgang

Rapid revenue growth—without scaling headcount—is one of the main theses of Jeffrey J. Bussgang’s newest book, The Experimentation Machine: Finding Product-Market Fit in the Age of AI.

Bussgang, a Harvard Business School professor and General Partner at Flybridge Capital, has invested in AI startups for over 15 years. In that time, he has seen the Revenue Per Employee metrics grow exponentially.

In this page, Bussgang shares some examples of this exponential growth and teases the secret to achieving it:

Imagine if every person you hired was worth $10 million to your company. In other words, if you hired the right person, you’d earn an additional $10 million. If you hired the wrong person, the opportunity cost would be $10 million.

Would this change your hiring process? Would you think more carefully about who you hired and how you set them up for success?

This is a hypothetical question, but not for long. Revenue-per-employee metrics are skyrocketing across the tech industry as companies become leaner and more efficient with AI. Let’s do a quick comparison:

Ovia Health, at the time of their acquisition in 2021, reached $20 million ARR with 100 employees. That’s $200,000 in revenue per employee—a good number for a growing company in pre-gen AI days. Today, Flybridge has a portfolio company that just reached $40 million ARR, in less than a year, with just twenty employees. That’s $2 million in revenue per employee—a 10x increase.

And we’re still in the early innings of this super-scaling revolution. One founder friend, who recently raised his seed round from Sequoia, promised the firm that he would never grow to over 100 employees. He’s committed to reaching a billion dollars in revenue with no more than 99 employees by incorporating AI into every part of the company—$10 million in revenue per employee.

The question for founders today is this: If each new hire is worth 10–50x more than they were previously, how do you ensure you hire the very best? What does a $10 million employee look like?

There used to be a rule at Flybridge that any AI startup we invested in needed to have a co-founder with a PhD in machine learning. In 2012, I wrote a cheeky blog post with a nod to the classic movie The Graduate titled, “Hey Graduates: Forget Plastics—It’s All About Machine Learning,” encouraging young people to study this important field in the era of Big Data.

We’ve scrapped that requirement today because a PhD is no longer needed to build an AI-forward company. The technical playing field is being leveled for all but the most bleeding-edge companies.

❓ One Question: If you could build anything in the world, what would it be? 

This whimsical question is becoming more realistic every day.

For non-engineers like me, there is a learned helplessness when it comes to software. I don’t have a habit of thinking about tools and apps I could build, because I was never capable of doing it in the past.

Today, anyone can build anything—or at least an impressive prototype that accelerates a project by weeks or months.

I have been keeping a list of tools and apps to build, and I suggest you do the same. Everyone needs to unlearn the “not a coder” mindset.

🗳️ Wrap Up and Feedback:

That’s it for today! We are back on schedule. I would love your thoughts on this newsletter, but let me know what you tools, apps, or companies you are building!

See you tomorrow!

Cheers,

Ben