The $500M Lesson Nobody's Board Wants to Hear
What ElevenLabs' revenue engine reveals about the real cost of "we'll figure out sales later"
A friend sent me a thread breaking down ElevenLabs’ go-to-market org last week. I read it twice, then sat with it for a day, because most “here’s how [hot AI company] scaled” threads are recycled wisdom with a new logo stapled on top.
This one wasn’t.
It was a teardown from Carles Reina, the guy who joined ElevenLabs as its first investor and fourth employee and built the revenue function from a spreadsheet into a $500M ARR machine in about three years. Before that, he’d built go-to-market at Sonantic (acquired by Spotify), cut his teeth at Tractable, and worked international expansion at Uber. This isn’t a marketer’s theory of sales. It’s an operator’s record of what actually happened.
So let’s establish why this company is worth studying before we get to how.
The company: proof that the model works
ElevenLabs started in 2022 as a text-to-speech company — two founders, Piotr Dąbkowski and Mati Staniszewski, trying to build voices that didn’t sound like a GPS unit reading a eulogy (my wife makes fun of my British accent Siri).
By 2026, it isn’t a voice company anymore. It’s audio infrastructure. ElevenAgents handles conversational AI for customer support and sales. ElevenCreative handles content generation and localization across more than seventy languages. ElevenAPI is the raw developer layer underneath both.
The customer list is the part that should make every B2B founder sit up. Deutsche Telekom. Square. Revolut, running voice agents across four million customers in the UK and Europe. Klarna, using ElevenLabs as first-line phone support for thirty-five million U.S. customers. The Ukrainian government. HarperCollins. Roughly 41% of the Fortune 500 use the platform in some form.
The capital tells the same story from a different angle. ElevenLabs closed 2025 at $330M in ARR, then added another $100M in net-new ARR in a single quarter, crossing $500M ARR by April 2026. The Series D — $500M, led by Sequoia — priced the company at $11B, more than tripling its valuation from thirteen months earlier. By the time you read this, reports are already circulating about a raise at double that number.
That’s not hype-cycle noise. That’s a company that figured out how to convert product into recurring revenue at a pace almost nobody manages. Which is exactly why the “how” matters more than the “what.”
The playbook: revenue as a designed system, not a headcount problem
Here’s the pattern I see with almost every founder I advise: they treat sales like a staffing decision. Hire a couple of AEs, throw together a comp plan that seems fair, hope pipeline materializes, and six months later they’ve got a bloated team that still can’t forecast, close, or expand with any consistency.
Reina’s approach at ElevenLabs inverts that completely. He designed the system first, then hired into the system. Three pieces of that design are worth stealing outright.
First: quota as a culture filter, not a spreadsheet input.
The standard SaaS quota multiple is somewhere around 6–8x base salary. ElevenLabs runs at 20x. A rep with a $100K base carries a $2M quota. That number sounds unreasonable until you learn that over 80% of the team hits it.
The multiple isn’t really about revenue math. It’s a screening mechanism. It filters for a specific kind of operator before you ever get to the comp conversation — and it sets the bar high enough that mediocrity self-selects out during the interview process, not eighteen months into a bad hire.
Second: pessimistic forecasting as an operating discipline.
Most sales orgs forecast optimistically because optimism is what gets funding rounds closed and boards off your back for a quarter. Reina builds forecasts on the opposite assumption — brutal, deliberately conservative numbers that force reps to justify every dollar of pipeline with evidence, not vibes.
The output isn’t a lower number. It’s a trustworthy number. And a trustworthy forecast is worth more to an investor than an aggressive one, because it means the next number you give them will also be true.
Third: expansion revenue is a shared incentive, not a handoff.
The most common failure mode I see in growth-stage companies is the AE closes the deal, tosses the account over the wall to a CSM, and everyone’s incentives point in different directions from that point forward. ElevenLabs pays both the AE and the CSM on upsell revenue. Growth on an existing account becomes something both people are financially rowing toward, instead of a line item nobody owns.
That one change — compensating two roles on the same expansion dollar — does more to prevent the “we signed them, then they churned quietly for a year” problem than any amount of customer success tooling.
Reina’s GTM lookback
Carles’s X post is an inside look at the GTM moves he would make sooner knowing what he knows today.
Hire earlier, even if your GTM model is PLG.
Sales enablement is key. I love this and its why I built the first Product GTM team at my 50+ year organization.
Open more markets. This works IF you have scalable process in place.
Be more vocal about ROI.
Hire senior sellers. I agree with Carles. They cost more, but their network should be worth it.
Build a fun brand. Brand is one of my Six Pillars of Modern GTM. People like to do business with fun people.
Ditch bad deals. We talk about this a lot lately in my company. If a prospect isn’t your ICP, you are likely fighting for a bad deal.
Why this translates beyond a $500M AI company
I want to be direct about something here, because it’s the part founders skip: you don’t need ElevenLabs’ valuation to run ElevenLabs’ system.
The 20x quota multiple is calibrated to their business, not a universal law — swap in whatever multiple reflects your margins and deal size. But the principle — using comp design as a hiring filter instead of a retroactive fix — works at $0 ARR the same way it works at $500M.
Pessimistic forecasting costs you nothing to implement tomorrow. It’s a discipline, not a headcount. The only thing standing between your current sales meetings and this practice is whether you’re willing to hear a smaller, truer number instead of a bigger, hopeful one.
And the shared expansion incentive is just math you can rewrite in your next comp cycle. It doesn’t require a Series D.
I’ve raised close to $800M in debt and equity across my career, and the pattern I’ve watched sink more founders than bad product is this: they build a revenue org reactively, chasing whatever the current fire is, instead of designing the system the fire is supposed to run through. ElevenLabs didn’t get to $500M ARR in three years because their product was better than everyone else’s — plenty of good audio AI exists. They got there because revenue became a system before it became a headcount problem.
Here’s the question worth sitting with: if you audited your own sales org the way Reina audited ElevenLabs’ — quota logic, forecast honesty, expansion incentives — would you find a designed system, or three years of decisions made under pressure that nobody’s gone back to fix?