The Twenty Minute VCElevenLabs CEO/Co-Founder, Mati Staniszewski:The Untold Story of Europe’s Fastest Growing AI Startup
CHAPTERS
- 0:00 – 1:14
$200M ARR milestone and why ElevenLabs is a global outlier
The episode opens with a punchy reveal of ElevenLabs crossing $200M in revenue, setting the stage for a discussion on hypergrowth in voice AI. Harry frames ElevenLabs as one of the fastest-growing AI companies and tees up themes like fundraising difficulty, enterprise traction, and defensibility.
- •Crossing $200M revenue and the pace from $100M to $200M
- •Initial framing of ElevenLabs’ growth relative to other AI startups
- •Early hints at pre-seed difficulty and investor skepticism
- •Enterprise use cases and large contract sizes teased upfront
- 1:14 – 4:14
Growing up in Poland: ambition, worldview, and “talent density”
Mati describes how moving from Warsaw suburbs into more competitive academic environments expanded his sense of what’s possible. He emphasizes how proximity to exceptional peers shaped motivation—an idea he later tries to replicate inside ElevenLabs.
- •Smaller-world mindset vs. expanding horizons through education in Warsaw
- •Motivation from high-performing peers and competitive environments
- •“Talent density” as a driver of ambition and execution
- •Parallels between formative years and company-building philosophy
- 4:14 – 6:15
Founding spark: hack-weekends, Google/Palantir experience, and bad dubbing
Mati explains how he and co-founder Piotr iterated via weekend projects, eventually focusing on audio. The specific pain of Poland’s monotone movie dubbing helped crystalize a larger vision for expressive, emotionally rich voice AI.
- •Hack-weekend experimentation as a repeatable idea-generation engine
- •Early audio project: speech analysis and coaching concept
- •Polish dubbing problem as an initial wedge and inspiration
- •Vision expands from dubbing to a broader “everything voice” platform
- 6:15 – 8:45
From dubbing to creator needs: discovering the real early wedge
They run parallel tracks: Piotr tests stitching existing tech; Mati tests demand by cold-emailing creators. Dubbing interest is lukewarm, but creators strongly want simpler workflows—voiceovers, corrections, and script preview—leading ElevenLabs to prioritize core text-to-speech quality.
- •Parallel strategy: prototype with existing tools + market validation outreach
- •Cold email experiments and what low-intent responses revealed
- •Pivot from language dubbing to creator-centric voiceover/narration tooling
- •Decision: build proprietary models to escape uncanny valley
- 8:45 – 12:24
Model strategy and the “plateau” debate: research vs product advantage
Harry probes how investors should decide when a startup needs its own model and whether AI progress is plateauing. Mati argues some voice use cases (e.g., narration) may plateau, making product execution and platform depth essential, while voice overall still has room for rapid progress.
- •When owning the model is necessary vs leveraging existing foundations
- •2022 context: weak market options made custom modeling unavoidable
- •Multimodal direction (reasoning + speech) and Eleven V3 framing
- •Research advantage commoditizes; durable edge comes from product + platform
- 12:24 – 16:48
Defensibility vs giants: answering “Why won’t OpenAI just do this?”
Mati tackles the existential question of competing with platform giants. He argues focus on voice, a scarce pool of top voice researchers, and superior product-layer execution (tooling, integrations, monitoring, deployments) create a moat beyond raw model quality.
- •Focus as a moat in a world of many AI opportunities
- •Voice research talent scarcity and assembling a “top 100” caliber team
- •Product/platform layer: deployment, evaluation, integrations, monitoring
- •Speed of execution as a compounding advantage
- 16:48 – 20:35
Pre-seed reality: 30–50 investor meetings, $2M raised, $9M post
Mati recounts a difficult pre-seed amid muted public excitement for AI voice and concerns about market size and defensibility. They rejected an accelerator offer, burned savings on GPUs and early hires, and ultimately raised $2M at a $9M post-money valuation with a small set of believers.
- •Top investor objections: research risk, market size skepticism, defensibility
- •Turning down an accelerator and choosing independence
- •Bootstrapping with savings while ramping GPU spend and first hires
- •Pre-seed specifics: $2M raised, $9M post, key early backers
- 20:35 – 23:49
Beta launch and early PMF signals: waiting list, audiobook hack, momentum
They delayed announcing the round until it could pair with product news—beta release in Jan 2023. Mati shares the first strong demand signals: a blog post showcasing “AI that can laugh,” a surge in waitlist signups, and an author using the tool to produce an audiobook that passed as human narration.
- •Why they time fundraising announcements to product/customer milestones
- •“AI that can laugh” PR moment and waitlist spike
- •Early power users: audiobook author workflow and viral validation
- •PMF as a continuum—users loved it, but long-term durability was the real test
- 23:49 – 30:10
Launch distribution lessons: forums beat press; fundraising should not distract
Mati advises founders to optimize launches for user acquisition rather than traditional media hits. He argues grassroots channels—newsletters, Discord, Reddit, Hacker News—drove real adoption, and cautions that post-launch investor attention can become a costly distraction unless managed deliberately.
- •Tie announcements to substantive product/customer progress, not the round itself
- •Traditional press often has minimal impact on actual user growth
- •Community channels (Discord/Reddit/HN) as primary distribution levers
- •Fundraising: line up investors, re-engage only when needed; avoid “always raising”
- 30:10 – 41:14
Series A mechanics: Nat Friedman’s API testing, a16z help, and speed to win
The Series A forms as momentum builds in early 2023, with Nat Friedman standing out by personally testing the APIs and offering detailed product feedback. a16z demonstrates partnership before investing via introductions and support, and Mati highlights speed (in diligence and decision-making) as a key differentiator.
- •Nat Friedman’s hands-on diligence and rapid understanding of the vision
- •a16z’s pre-investment support (introductions, voice licensing connections)
- •Why US brands helped signal credibility globally (especially SF)
- •Speed of investing and execution as decisive in competitive rounds
- 41:14 – 48:44
Operating model: small teams, no titles, culture tensions, and learning loops
Despite rapid scaling, Mati prefers small, high-ownership teams organized by product areas and execution pods. He explains dropping titles to reduce politics and keep impact-centered decision-making, then candidly shares a low point: a partner shipping dubbing before ElevenLabs, hurting morale and exposing partnership risk.
- •“Small and mighty” teams: 5–10 person pods with high ownership
- •Product-area team structure (studio, core app, voice agents; enterprise vs self-serve)
- •Why removing titles reduces distraction and enables rapid leadership mobility
- •Culture low point: partner launched dubbing first; advice—be authentic, then execute relentlessly
- 48:44 – 1:01:34
Compute strategy and unit economics: building data centers and balancing cost vs speed
Mati discusses why ElevenLabs built its own training infrastructure—control and a projected two-year breakeven—while using partners for inference. He addresses broader AI margin concerns, arguing many apps have poor unit economics today but can win through cost improvements, brand, and data feedback loops; ElevenLabs aims to stay healthier by controlling research, product, and distribution.
- •Rationale for owning training data centers: breakeven, control, faster experimentation
- •Tradeoff: optimizing for shipping new models quickly vs minimizing cost
- •View on application-AI margins: risky, but winners can emerge with brand + iteration
- •ElevenLabs’ claimed advantage: integrated stack (research + product + distribution)
- 1:01:34 – 1:04:56
Hypergrowth metrics and business focus: $35M to $200M, enterprise agents, global expansion
Mati shares the growth curve: roughly $35M by end of 2023 and a rapid climb to $200M, alongside plans to scale headcount from ~250 toward ~400. He explains enterprise traction (largest contract around $2M) and positions conversational voice agents for customer support as the biggest future business line, with international outposts to accelerate distribution.
- •Revenue trajectory: end-2023 ~ $35M; 20 months to $100M; ~10–15 months to $200M
- •Enterprise + self-serve mix; biggest contracts centered on call centers/support/assistants
- •Strategic focus: conversational/voice agent platform with deep enterprise integrations
- •Scaling approach: global outposts (Brazil, Japan, India, Mexico) while monitoring efficiency
- 1:04:56 – 1:11:22
$3.3B valuation raise, brand-value of top funds, and staying focused while doing more
Harry and Mati discuss the pricing and timing of the latest round (announced Jan 2025; done late 2024) at a $3.3B valuation, and the rationale for raising: bringing forward bets on multimodal models, international expansion, and enterprise-grade agent capabilities. Mati explains how top-tier investors (a16z, Sequoia, others) materially impact credibility with customers and talent, while cautioning that new initiatives must not distract the core.
- •Latest valuation: $3.3B; round timing and revenue multiple context
- •Use of proceeds: multimodal investment, international expansion, enterprise agent features
- •Fund brand as credibility amplifier (customer trust and hiring signal)
- •Decision filter: can new efforts be parallelized without harming the core product?
- 1:11:22 – 1:22:03
M&A interest, employee liquidity via tenders, and the quick-fire worldview
Mati confirms acquisition offers but emphasizes building independently, with secondary liquidity and tender offers each round to reduce employee risk and sustain ambition. The episode ends with quick-fire topics: Europe can build global winners, voice as a primary interface, regulation concerns, founder brand ambivalence, and risk-taking as a guiding principle.
- •Handling acquisition interest: diligence, but strong bias to continue building
- •Secondary liquidity strategy: tenders for vested employees to reduce pressure to sell
- •Quick-fire takes: Europe can build global-scale companies; voice as key interface
- •Policy and leadership: align EU AI regulation closer to US approaches; risk-taking mindset
