How did ElevenLabs raise $781M and reach an $11B valuation?

The rounds at a glance

Mati Staniszewski and Piotr Dabkowski founded the company in April 2022. They grew up together in Poland and bonded over badly dubbed films where the voices never matched the actors, which sounds like a small irritation and became the founding insight for an $11 billion company. Mati came from Palantir. Piotr was a machine learning engineer at Google, working in the exact technical domain they went on to build in.

The pre-seed that nobody remembers being hard

Those credentials are about as strong a signal as you can send in AI, and the $2M still took a grind. Think about when they were pitching. Founded April 2022, closed January 2023, which means most of those conversations happened before ChatGPT launched and before the market accepted AI as a platform shift. Two Polish founders, no US network, selling voice synthesis into a category most investors filed under niche.

The round closed with Credo Ventures and Concept Ventures, two European funds. Not a16z. Not Sequoia. Neither of the names that would compete to lead later rounds.

This is the part founders should sit with. A company that reached $11B had trouble raising $2M, and that is the rule rather than the exception. Investment rates across the industry sit around 0.1 to 0.3% of companies seen, so a pre-seed founder should expect 30 or more clear passes before finding the few who lean in. Rejection at that stage carries almost no information about the company. What separates the founders who get through is volume and process, and I've written up how that works in why investors reject startups.

The year they didn't fundraise

After the pre-seed closed, Mati and Piotr did something most first-time founders get wrong. They didn't start working the Series A. For roughly twelve months the whole team, about eight people, pushed on one thing: making the voice models as good as they could possibly be. No press tour, no conference circuit, no premature investor coffees.

Then the beta launched in January 2023 and the internet found them. 4chan users cloned celebrity voices. Harry Potter by way of Balenciaga went everywhere. They hit a million users inside five months.

There's a framing I use with founders that explains why this worked: investors buy lines, not dots. A single meeting or deck is a dot, a snapshot with no direction attached. A line is what an investor sees when they watch you ship over months. ElevenLabs drew a very steep line in public, in the form of a product improving faster than anything else in the category, and by mid-2023 that line was visible to every AI investor alive. They never had to argue for it in a meeting.

What the viral moment did to the Series A

By June 2023 they weren't looking for investors. a16z led $19M at a $100M valuation five months after the beta launched, and the angel list around it reads like a roll call of the people who built the last platform generation: Nat Friedman, Daniel Gross, Mike Krieger, Brendan Iribe, Mustafa Suleyman.

$19M is an ordinary Series A. The signal was not the money. When those specific names put personal capital into a company, every fund in the market reprices its attention, and the round stops being a fundraise and becomes a competitive process. That dynamic is worth understanding even at a fraction of the scale, because it's the same loss aversion that makes a first term sheet pull three more behind it.

What the later rounds tell you, which is less than you'd think

The Series B in January 2024 valued the company at $1.1B, roughly 11x the A, seven months later. Revenue went from close to nothing in early 2023 to an estimated $80 to $90M ARR, and the team grew from 8 to around 40. The Series C added $180M at $3.3B. Sequoia led $500M at $11B in February 2026, with an IPO reportedly on the table.

Impressive, and mostly uninstructive. Once you're raising a B at a billion-dollar valuation, the round is driven by numbers, and the numbers were determined years earlier. Every decision that produced this outcome was made between April 2022 and June 2023, while nobody was watching.

What a founder can take from this

• Your credentials buy meetings, and nothing else. Palantir and Google got them in rooms. It still took a grind to close $2M, because at pre-seed the credential is the start of the assessment.
• Splitting your attention costs you both things. They put twelve months into the product and let the product create the round. Half-building while half-raising tends to produce a weak version of each.
• Virality was downstream of quality. The product spread because it was better than the alternatives by a wide margin, which is a consequence of the stealth year and not a marketing tactic anyone could copy.
• Being early to a category is a fundraising cost before it's an advantage. They pitched voice AI to people who didn't yet believe in it. That's a narrative problem, and it's solvable, but it explains the pre-seed difficulty better than anything about the team.

Gian Seehra Ex-Octopus Ventures, part of deploying $200M as a VC. 3x founder, raised $16M. Has coached 120+ founders who have raised $250M+ collectively.