You know that moment on Dragon’s Den…

The founder walks in confident. The branding’s eye-catching. The pitch is smooth. There’s talk of scaling, growth, maybe even international expansion. 

It all sounds quite promising.

And then Deborah Meaden leans forward and asks something very simple.

“What’s your gross margin?”

Then, there’s a pause. A slight shift in posture. A bit of mental arithmetic happening out loud. Suddenly the energy drains from the room, and you can almost see investor confidence evaporating in real time.

The numbers don’t stack up, and that’s what tends to kill the deal.

When you’re raising capital, investors test survival first. Beneath every funding conversation sits the same question: will this business last?

People frame it as numbers versus story, but the reality is a sequence.

And at FundOnion, we follow that order. 

1/ Can you deliver what you’re promising? 

2/ Do the economics work? 

Only then does the narrative carry weight.

In other words, the numbers can’t lie, but they also can’t inspire. That’s why you need the story as well.

Just never - and I mean never - let a good story cover up bad maths.

Now, onto today’s topic.

1/ ElevenLabs raised $180M Series C at a $3.3B valuation (tripled in a year) led by a16z and ICONIQ Growth. Total funding now $281M since 2022.

2/ Two founders from Poland got frustrated with terrible movie dubbing and decided to fix AI voices - just voices, nothing else.

3/ They focused ruthlessly narrow: make AI voices that sound human. No chatbots, no AGI dreams, no "transforming industries".

4/ The proof showed up in usage: 60% of Fortune 500 companies now use ElevenLabs, and developers built 250,000 voice agents in just two months.

5/ Enterprise adoption created stickiness - once you're through procurement and compliance, you're embedded in the system and harder to replace.

6/ The lesson: focus beats breadth, numbers beat narrative and specialism creates a defensible moat that investors actually want to fund.

LinkedIn right now is basically a fever dream of self-proclaimed "AI Visionaries" announcing that their new model can write your code, manage your calendar and probably negotiate your divorce settlement while it's at it.

Then you have ElevenLabs.

Two guys from Poland, Mati Staniszewski and Piotr Dabkowski, got annoyed by the absolute state of dubbed movies back home. 

You know the ones… where a single overly-enunciated, Disney-esque voice narrates every character’s dialogue like it’s a toothpaste advert.

They didn't wake up one morning determined to build a digital deity or revolutionise human-computer interaction. They just wanted AI voices that didn't make you want to mute the entire film. 

That's it. Simple problem, specific solution.

That decision matters more than it sounds on the surface.

AI is absolutely packed right now. Everyone's shouting about building "transformational technology" that's going to reshape civilisation as we know it. 

A few years ago it was crypto that was going to fix everything. 

Before that, blockchain was the answer to every question nobody asked. 

Before that, the internet was going to democratise knowledge and make everyone enlightened.

Before that, the automobile was going to free humanity from geographical constraints.

There's always a new revolution. The hype cycle never stops.

In markets like that, big sweeping claims don't actually cut through the noise anymore. What cuts through is focus. Sharp, undeniable focus on one thing you can own completely.

When everyone's promising to solve everything, the real advantage goes to the company that picks one clear problem, plants a flag and solves it so well that no one else can compete. 

Investors don't want a bigger story at that point - they want proof. They want traction. They want a niche you can actually dominate and defend, not vague promises about changing the world.

So instead of trying to become the "brains" of AI like everyone else, ElevenLabs made a far more commercially intelligent move: they became the vocal cords.

And by doing exactly that - and only that - they became the category leader.

That tight focus turned into something investors could actually wrap their heads around.

ElevenLabs raised $180M in a Series C at a $3.3B valuation - tripling their value in just one year. Andreessen Horowitz (a16z) and ICONIQ Growth co-led the round. Total funding since 2022 now sits at $281M.

Here’s a crucial part though.

AI voice technology isn't some completely unique, never-been-done idea. 

Other companies are doing similar things. This wasn't a "no one's ever seen this before" moment that blew investors' minds with novelty. The raise happened because people were already using the product at scale.

Specifically, 60% of the Fortune 500 companies use ElevenLabs. If that’s not market validation, I’m not sure what is.

With industry heavyweights on your customer list, it tells investors a few things:

From a business perspective, landing one large contract worth $20M a year changes your entire risk profile. The revenue is bigger, more predictable and significantly more stable than chasing millions of tiny subscriptions that churn every few months.

This is where "stickiness" becomes everything.

When a big company buys your product, they put you through procurement bureaucracy, compliance checks, and vendor onboarding. This process takes months.

But once you're through? You're embedded. Your product is woven into their daily operations. Their team knows how to use it. Their workflows depend on it.

Replacing you means going through that entire process again with someone new - disrupting workflows, retraining staff, risking things breaking during the switch. Most companies would rather keep paying you than deal with that mess.

That's your moat: the operational pain of switching.

ElevenLabs had something else working for them too. Developers built 250,000 voice agents in just two months using their platform. That's builders actively choosing the product when they could've used anything else.

So investors saw enterprise adoption proving the business worked at scale, and developer traction proving the product was genuinely good.

The story made it interesting. The usage made it investable. 

I know which reputation I'd rather have...

This newsletter is sponsored by Longview Tax.

Manual tax processes drain your team's time and energy, leaving no time for analysis.

When you automate and centralize your tax data your team can access what they need, when they need it and confidently make decisions.

The best part? Your team gets time back to focus on strategic initiatives.

Download the paper

If you’re raising money - whether you’re a SaaS founder in Manchester or a manufacturing business in Leeds - the mechanics are the same.

We live in a rules-based system that tends to reward specialism over generalism.

Adam Smith - the grandfather of modern economics - illustrated this perfectly with his pin factory example. Instead of one person making an entire pin from scratch, productivity exploded when each worker specialised in just one small part of the process. That principle hasn't changed. It still applies in modern markets, maybe even more so now.

When you specialise, two things happen: you get dramatically better through repetition, and the market starts perceiving you as the expert in that specific thing.

I saw this firsthand when I worked at Lehman Brothers. There was a lawyer whose entire job was paragraph nine of the ISDA documentation - that's the International Swaps and Derivatives Association contract used in derivatives trading. Not the whole document. Not general advisory work. Just paragraph nine.

Sounds absurdly narrow, right? Frankly, quite boring too. But from a positioning perspective, it was absolutely genius.

When Citibank or Goldman needed clarity on that clause, they just called him. Because when someone only does one thing, we instinctively assume they must be exceptional at it. And usually they are, because that's literally all they do.

That's how specialisation becomes its own marketing strategy.

I've been thinking about this a lot lately, especially as FundOnion grows. There's this constant, nagging temptation to broaden out - to offer every financial service under the sun because we technically could. The ElevenLabs story is honestly a cold shower for that impulse. It's a reminder that owning one thing completely is almost always better than being okay at ten things.

Once you've actually mastered something and built that dominant position, however, then you can think about expansion. But you expand from strength, not ambition. You've already proven you can execute at the highest level in one domain. The market trusts you. Your operations are solid. That's when adjacent moves make sense.

An important point here is: don’t dilute your proposition just to look bigger. Own something tightly first. Let mastery and traction build from there.

AI is apparently going to solve... everything. Investors have heard this promise before. Many times. And after a while, the claims start to blur together.

You don't cut through that exhaustion by announcing you're revolutionary. You cut through by showing actual adoption - real customers, measurable usage, revenue on a spreadsheet. Enterprise adoption especially, because that's what shifts you from "trendy concept" to "fundable business".

ElevenLabs didn’t raise because AI is hot. They raised because large multinationals were already using the product at scale.

When industry heavyweights are on your customer list, it proves you can survive procurement hell, pass compliance audits, and navigate buying cycles that would destroy most startups.

Once you're through those gates, you're embedded.

Bureaucratic stickiness - replacing you is too painful. Development stickiness - your API is woven into their infrastructure. Operational stickiness - their teams are trained on your platform and don't want to learn something new.

One large, recurring contract changes your entire risk profile. The revenue is bigger, steadier, and way more durable than chasing thousands of flaky $10/month subscriptions.

For SMEs, this doesn't mean you need Apple as a client tomorrow. You need customers who stick around and pay consistently. The real moat is operational dependency (rather than branding, for example). Once your product is part of their daily workflow, their team knows how to use it and their processes are built around it, switching becomes a genuine headache they'd rather avoid.

ElevenLabs operated like a completely normal company. The business model was standard. The operations were unremarkable. A bit dull, really.

Thank god.

Raising capital doesn't require reinventing how businesses work. Sometimes - most times, actually - the advantage is just doing one specific thing absurdly well instead of building some sprawling "super app" that does everything poorly.

Operational discipline gets zero hype in a market that's obsessed with vision statements and disruption theatre. But when investors are actually deploying capital, execution beats inspiration every single time.

Back to Dragon’s Den.

When investors assess a business, they test survival first. 

Can you deliver? 

Do the economics actually work? 

Is there real proof people care, or just your conviction that they should?

The narrative only matters after those boxes are ticked. Numbers can't inspire on their own, but they have to stack up before anyone takes your story seriously.

ElevenLabs got the order right: ruthless focus, proven traction, enterprise validation, operational stickiness, sensible economics. That's why investors paid attention.

In capital markets, sequence is everything. Lead with vision before proof and you're dead in the water.

ElevenLabs made the decision simple for investors. The usage was visible. The enterprise contracts were meaningful. The growth was measurable. 

That clarity compresses doubt.

If you’re building, your job isn’t to sound impressive. It’s to remove friction from the decision. 

Raising soon? Reply to this email and I'll help you bulletproof your numbers before you're in the room - so when they ask the hard questions, you've got answers ready.

Till next time,

James