Jeff Bezos - one of the wealthiest people on earth - once asked Warren Buffett - who sits comfortably in that same category - what his secret was. 

Buffett's answer was a self-confessed, simple strategy that nobody ever likes to hear:

In other words: time, patience and the discipline to keep going when nothing exciting is happening. That’s the secret.

We're surrounded by stories that tell us the opposite - the guy who built a one-man AI business from his bedroom seemingly overnight, the founder who raised £10M on a napkin and a vibe. We see those stories and absorb them as the norm.

But those stories represent a tiny fragment of a much larger pool of people who tried to grow at exactly that speed. For every bedroom founder who made it, there are a thousand who just learned an expensive lesson and went back to a day job.

I think people are massively underestimating the power of going slowly right now. I keep asking myself: what's the rush?

That's a strange note to open on when I'm about to walk you through a $30B capital raise. But the number is almost beside the point - the way they built towards it is what's worth paying attention to.

1/ Anthropic closed a $30B funding round - the second-largest private raise in tech history - at a $380B valuation. They were already generating $14B in annual revenue and already profitable when they did it.

2/ At a 27x revenue multiple, investors are betting on Anthropic becoming one of the largest software businesses ever built - bigger than Salesforce, bigger than Oracle.

3/ Anthropic had something most businesses raising capital don't - genuine scarcity. There aren't many of them on the market, and that changes who has the leverage.

4/ The lesson is about the position they were in when they raised. Profitable, growing at 10x, raising from strength - not desperation.

In 2021, Dario Amodei - who was VP of Research at OpenAI - and his sister Daniela - who was VP of Safety and Policy at the same company - handed in their notices. Around nine colleagues went with them. 

Together they founded Anthropic in San Francisco with a specific idea: that the people most concerned about the risks of powerful AI should be the ones building it, not watching from the sidelines while someone else does.

They structured it as a public benefit corporation - a type of company that is legally required to pursue a social purpose alongside profit. For Anthropic, that purpose was building AI that could be understood, controlled and corrected by humans. At a time when most of the industry was focused purely on making models more powerful, that was a fairly unusual thing to be putting at the centre of your business.

It took two years before they released anything to the public. Two years of research, testing and deliberate patience before Claude launched in 2023. 

For context, that's an eternity in an industry where the standard move is to ship fast and fix the problems once real people are using it. Anthropic made a different call - they wanted to understand what they were releasing before they released it. That decision cost them time and first-mover attention. It also built the foundation of trust that made Claude so attractive to developers and enterprises when it did land.

And if you're reading this, you've almost certainly used it. To summarise a document. To clean up an email before it went to a client. To write a LinkedIn post and then paste it straight back in with the prompt "make this sound more human." Most people use it every day without knowing Anthropic built it.

Developers took to it particularly quickly. It became the tool of choice for anyone building something that needed to actually work reliably, not just impressively. Amazon noticed. So did Google. Both made multi-billion dollar commitments, becoming cloud partners and strategic investors in the process. 

By 2025, Anthropic had over 3,000 employees and Claude was embedded in the workflows of businesses across the world.

$30B raised. $380B valuation. 27x revenue multiple. 20 investors. The second-largest private funding round in the history of technology.

The number worth understanding here is 27 - the revenue multiple. It means investors paid 27 times what Anthropic currently earns in a year. A well-run, growing business typically raises at 5 to 10 times revenue. Exceptional companies in rare circumstances might hit 15 to 20x. At 27x, investors are paying for the business Anthropic is going to become, not the one it is today.

To give those investors the return they're expecting, Anthropic probably needs to hit $100B in annual revenue. For context - Salesforce does $34B. Oracle does $50B. Anthropic needs to become larger than both.

The round was led by Coatue and Singapore's GIC, with Microsoft, Nvidia and a long list of other investors involved. The term sheet on this was negotiated to within an inch of everyone's life. At a $380B valuation, every investor came in with enormous growth expectations and significant legal protections built around them. These things don't get agreed over a handshake and a good feeling.

And yet Anthropic almost certainly had more leverage in that room than you'd expect. There simply aren't many companies doing what they do. You've got OpenAI, a handful of others - Cohere, Perplexity - and that's about where the list ends. 

When something is genuinely rare and genuinely in demand, the people selling it have more power than the people buying it. If talks got difficult, Anthropic could turn around and say they'd find another investor. Capital is everywhere. A business with Anthropic's technology and position is considerably harder to come by.

Anthropic were profitable and growing at 10x before they raised. They went to market because the timing and position were right - and the terms reflected that. A lot of founders do the opposite. They wait until the money is nearly gone, and that pressure is written all over them by the time they sit down with an investor.

The best time to raise - equity or debt - is when you have options. That's when the terms are better, the interest is higher and you can walk away from a deal that doesn't work. Get your position right before you go looking. You can start at FundOnion:

*It’s free and it only takes 90 seconds.

A $380B valuation sounds like a win. It's also a commitment to every investor, board member and observer that Anthropic will grow into that number. Every decision they make from here will be shaped by it.

The same applies at any scale. Raise at a number your business cannot support in the long run and you will feel it - in how you spend, how you hire and how fast you feel you need to move. Be clear on what you're committing to before you agree to the number.

With lots of investors and $30B on the table, you might assume Anthropic were the grateful ones. They weren't - because there simply aren't many companies like them. 

When something is rare and in demand, the people selling it have more power than the people buying it. And when I say Anthropic is in demand, they are so much so that they can’t build the infrastructure quick enough to serve their customer base:

So, before your next funding conversation, ask yourself: what do we have that is hard to find elsewhere? What would an investor actually lose by walking away from us? Those answers matter more in a negotiation than anything in your pitch deck.

When Nvidia - the company that makes the chips powering AI - put in up to $10B, it was buying considerably more than equity. Anthropic takes that capital and spends a significant chunk of it on Nvidia's own compute and infrastructure. 

Anthropic gets the funding. Nvidia gets the revenue, a front-row seat to where AI is going and early access to whatever gets built next. Both sides benefit beyond the initial cheque.

When thinking about who to bring onto your cap table, look beyond the money. The right strategic investor brings relationships, credibility and sometimes a commercial arrangement that directly benefits the business. The wrong one just brings opinions.

Most people look at a 27x valuation and assume the only way this works is if Anthropic dominates AI. But that's not how the people writing the cheques are thinking about it.

Bing has maybe 4% of the search market. Google has the rest. And yet Bing is a profitable business that Microsoft is perfectly happy to own. Second place in an enormous market is still an enormous business.

AI is a bigger, more fragmented opportunity than search ever was. There won't be one winner. Which means Anthropic doesn't need to beat OpenAI to justify its existence - it just needs to be the right choice for enough people. At the scale AI is operating, "enough people" is a very large number.

Anthropic raised $30 billion not because they were running out of money, but because in AI right now, infrastructure is the battlefield. GPUs are constrained. Data centres take years to build. And the partnerships that come with capital - Microsoft, Nvidia, Amazon - lock in compute supply before competitors can get to it. Whoever secures capacity early shapes what the market looks like for everyone else.

Founders tend to think about a raise in terms of what it buys internally - runway, headcount, new hires. It's worth thinking about what it does to the landscape around you too.

Raising from strength, at the right moment, with the right partners, can shift your competitive position in ways that go well beyond what you spend the money on. The capital is the signal. The infrastructure it buys is the moat.

Anthropic got here carefully. In an industry moving faster than almost any other in history, they made a deliberate choice to understand what they were building before they scaled it - and that discipline put them in the position to raise $30B from a place of strength.

Buffett said get rich slowly. Anthropic's version was get rich responsibly. The $30B came after. That order matters.

If you're thinking about a capital raise and want to work out what the right position looks like for your business - drop me an email: [email protected]

Till next time,

James