Strengths
- Series D valuation/fundraise corroborated by company and CNBC sources.
- Product breadth is directly visible on public product, docs and pricing pages.
- Public sources consistently show rapid global headcount expansion.
ElevenLabs Diligence Research Report
Proceed only with data-room diligence focused on ARR quality, gross margin, consent/training-data legality, safety controls, customer contract durability, supplier economics and governance maturity.
ElevenLabs Diligence Research Report
ElevenLabs is a fast-growing private AI audio decacorn with public evidence of an $11B Series D valuation, >$500M company-reported ARR, broad AI voice products, major investors, and marquee customer logos. Investment attractiveness is offset by high diligence risk around unaudited ARR quality, margins, customer concentration, legal/biometric exposure, training-data provenance, safety controls, security evidence and hypergrowth governance.
Public lawsuits allege unauthorized voice use, cloning or BIPA voiceprint violations.
Diligence request: Counsel review dockets, complaints, reserves, insurance, training data licenses, consent logs and settlement terms.
Voice cloning and agents can enable impersonation, fraud or harmful synthetic media.
Diligence request: Request abuse metrics, red-team reports, classifier validation, enforcement logs and law-enforcement referral history.
Public $11B valuation relies on company-reported ARR without audited financials, gross margin, retention, CAC or concentration data.
Diligence request: Request audited financials, ARR bridge, cohorts, margins, CAC/payback, cash runway and top customer schedule.
Google, AWS, OpenAI, Deepgram and others can bundle or undercut voice capabilities.
Diligence request: Run benchmark, pricing and win/loss analysis.
Global expansion plus voice/biometric products create EU, U.S. state, privacy, AI and DSA exposure.
Diligence request: Request compliance memos, DPIAs, regulator correspondence, DSA/AI Act/GDPR/BIPA analysis and data-flow maps.
Safety page does not verify SOC 2, ISO 27001, pen tests, incident history, subprocessors or cyber controls.
Diligence request: Request SOC/ISO reports, pen tests, incident logs, subprocessors, data retention controls and cyber insurance.
NVIDIA/GPU relevance is public; allocation, cloud, data and telephony supplier terms are not.
Diligence request: Request supplier contracts, spend, SLAs, capacity, failover and unit-cost curves.
The broad portfolio can dilute roadmap focus and increase safety, support and compliance burden.
Diligence request: Request product-level ARR, roadmap, benchmarks, staffing and deprecation policy.
Public valuation and ARR signals are strong, but financial statements, ARR quality, cap table and contracts remain non-public.
partially verified confidence: medium
Public ARR milestones are visible; audited financials, margins, burn and concentration are not.
| topic | public evidence | diligence read |
|---|---|---|
| ARR growth | Company says >$200M ARR in Sep 2025, $350M at 2025 end, >$500M by May 2026. | High growth; unaudited and needs ARR bridge. |
| Revenue quality | No public audited financials, cohorts, margins, cash burn, or concentration. | Central valuation risk. |
| Enterprise transition | Tender blog says enterprise revenue grew >200% YoY and split was approaching 50/50 with self-serve. | Review CAC, sales efficiency, support costs. |
Company ARR is unaudited.
not publicly verifiable confidence: low
No public board-approved projections were found; growth targets require private plan and sensitivity review.
partially verified confidence: medium
Public sources verify major rounds and valuation, but cap table, preferences and rights are non-public.
| topic | public evidence | diligence read |
|---|---|---|
| Series B | $80M raised on 2024-01-22; later sources imply prior valuation near $1.1B. | Verify valuation, terms and dilution. |
| Series C | $180M at $3.3B on 2025-01-30; $281M total funding. | Primary financing; review rights and preferences. |
| Tender and Series D | $100M tender at $6.6B; $500M Series D at $11B; $781M total funding. | Separate secondary liquidity from primary cash. |
not publicly verifiable confidence: low
No public bank, tax, debt, receivable, payable, reserve or commitment schedules were located.
Public materials verify a broad AI audio product suite; revenue mix, security and independent performance benchmarks remain gaps.
verified confidence: medium
ElevenLabs publicly offers agents, creative tooling, API, TTS, STT, cloning, dubbing, music and sound effects.
| topic | public evidence | diligence read |
|---|---|---|
| ElevenAgents | Enterprise voice agents with integrations, testing, monitoring, reliability. | Verify production deployments, SLAs and agent revenue. |
| ElevenCreative | Creator/marketer tools for speech, music, image and video across 70+ languages. | Review rights, moderation and product-level margins. |
| ElevenAPI and models | REST API, SDKs, TTS/STT, cloning, dubbing, sound effects and voice library. | Benchmark quality, latency, uptime and compliance. |
| topic | public evidence | diligence read |
|---|---|---|
| Pricing | Free, Starter, Creator, Pro, Scale, Business, and custom Enterprise tiers. | Analyze discounting, usage breakage and gross margin. |
| Low latency | Docs claim Flash around 75ms and Scribe realtime around 150ms. | Independently benchmark under customer loads. |
| Voice cloning/library | 10,000+ voices and cloning/generation capabilities. | Review consent, licenses, takedown and royalties. |
Not validated internal architecture.
Public customer logos and adoption metrics indicate market pull, but top-customer revenue, churn and contract terms are not public.
partially verified confidence: medium
Named customers and use cases are public but not ranked by revenue or active paid status.
| topic | public evidence | diligence read |
|---|---|---|
| Named enterprises | Deutsche Telekom, Square, Ukrainian Government, Revolut, Duolingo, NVIDIA, TIME, Meta, Epic Games, Salesforce, MasterClass, Harvey, Cisco, Adobe, Aston Martin F1, Synthesia. | Validate active paid contracts and references. |
| Fortune 500 usage | 41% in 2024 and >60% in 2025 employee-use claims. | Usage is not equivalent to enterprise ARR. |
| Platform reach | Company says ElevenAPI platforms reach over one billion users. | Need contract and traffic attribution. |
Illustrative only.
partially verified confidence: medium
Strategic customers/investors and impact relationships are public; commercial terms are not.
| topic | public evidence | diligence read |
|---|---|---|
| NVIDIA | Investor via NVentures; company says it leverages NVIDIA Blackwell GPUs. | Review GPU supply, pricing, capacity, data and exclusivity. |
| Strategic customers/investors | Deutsche Telekom, Salesforce, Santander, KPN, RingCentral and others named. | Separate customer revenue from investment signaling. |
| Impact program | Company cites 268 nonprofits and 3,472 voices for people with aphasia/ALS/MND. | Review cost, consent, privacy and liability. |
not publicly verifiable confidence: low
Revenue by customer is not publicly available.
not publicly verifiable confidence: low
No public severed-customer list or churn evidence was found.
partially verified confidence: medium
NVIDIA infrastructure relevance is public; cloud, GPU, data and telephony contracts are not.
ElevenLabs competes against hyperscalers, model platforms and realtime speech specialists; differentiation requires benchmark and win/loss proof.
partially verified confidence: medium
Google, AWS, OpenAI and Deepgram public pages show credible TTS/voice alternatives.
| topic | public evidence | diligence read |
|---|---|---|
| Google Cloud TTS | Google cites AI TTS, 380+ voices, 75+ languages and custom voices. | Hyperscaler distribution and procurement pressure. |
| Amazon Polly | AWS managed TTS with dozens of languages and 100+ voices. | AWS integration can win cloud-native buyers. |
| OpenAI and Deepgram | OpenAI audio docs and Deepgram sub-200ms TTS for voice agents. | Foundation-model bundling and specialist latency pressure. |
| topic | public evidence | diligence read |
|---|---|---|
| Latency and quality | ElevenLabs claims 75ms Flash; Deepgram claims sub-200ms TTS. | Run independent bakeoff. |
| Enterprise compliance | ElevenLabs has enterprise tier and safety page; hyperscalers have mature enterprise trust programs. | Request security evidence and buyer win/loss. |
| Rights and consent | Voice cloning lawsuits and safety controls both public. | Legal defensibility can become a competitive differentiator. |
No benchmark performed.
Public evidence supports hybrid PLG/self-serve and enterprise GTM; CAC, pipeline, sales productivity and concentration are private.
partially verified confidence: medium
Public pricing/docs support self-serve, and company claims support enterprise expansion.
| topic | public evidence | diligence read |
|---|---|---|
| Self-serve | Free/paid tiers, signup, docs, REST API and SDKs. | Need funnel conversion and support cost. |
| Enterprise | Custom enterprise tier; >200% YoY enterprise revenue claim; named customers. | Review CAC, sales cycles, NRR and discounting. |
| Strategic/investor-led | Strategic investors/customers include NVIDIA, Deutsche Telekom, Salesforce, Santander, KPN. | Review channel contracts and conflicts. |
partially verified confidence: medium
Major logos are public but not tied to revenue.
partially verified confidence: medium
Observed avenues include self-serve signup, API adoption, enterprise sales and strategic relationships.
Conversion metrics not public.
not publicly verifiable confidence: low
ARR per teammate screens well publicly, but sales productivity metrics are not public.
| topic | public evidence | diligence read |
|---|---|---|
| ARR per teammate | >$500M ARR and 530 teammates imply roughly $0.94M ARR per teammate. | Validate employee/contractor definitions and ARR quality. |
| Sales efficiency | No public CAC, quota, pipeline, bookings, payback or churn. | High-priority data-room item. |
| Customer concentration | Named logos but no revenue by customer. | Need top customer schedule and renewals. |
not publicly verifiable confidence: low
No public marketing budget, CAC or channel ROI was found.
R&D signals include technical founder background, rapid product velocity, safety team and NVIDIA infrastructure; model provenance and evals are private.
partially verified confidence: medium
Public sources show research roots, founder experience, safety team and GPU infrastructure signals.
| topic | public evidence | diligence read |
|---|---|---|
| Founder technical/business fit | Endeavor and author page describe cofounders, research and enterprise deployment backgrounds. | Request full leadership bios and references. |
| Safety engineering | Company says dedicated safety team and multilayer safeguards. | Need staff, budget, authority and incidents. |
| Infrastructure | NVIDIA Blackwell GPU usage public. | Review capacity and inference cost curves. |
partially verified confidence: medium
Public announcements show fast launches across agents, voice, STT, reader, dubbing, sound effects and platform taxonomy.
| topic | public evidence | diligence read |
|---|---|---|
| Research roots | Endeavor says ElevenLabs began as 2022 research company building TTS models. | Verify model provenance and IP assignments. |
| Product velocity | Series C/D cite Conversational AI, Voice Design, Sound Effects, Reader, Flash, Dubbing and platform taxonomy. | Review roadmap focus and revenue by product. |
| Agents | 250k agents by Series C and >2M conversational agents by tender blog. | Separate demos from production deployments. |
Public founder and headcount evidence supports hypergrowth, but org chart, compensation, equity, attrition and employment controls are private.
partially verified confidence: medium
No full org chart is public; founders, board signal, hubs and safety team are visible.
| topic | public evidence | diligence read |
|---|---|---|
| Founders | Mati Staniszewski and Piotr Dabkowski identified as cofounders; Mati former Palantir; Piotr technical AI background. | Verify titles, agreements, equity and succession. |
| Board | Series D says Sequoia led and Andrew Reed joins board. | Request full board/observer rights. |
| Headcount | 40 in 2024, 120 in early 2025, 330+ in Sep 2025, 530 in May 2026. | Hypergrowth control and attrition risk. |
partially verified confidence: medium
Headcount scaled from 40 to 530 publicly; function/location plan is private.
Employee/contractor mix not verified.
partially verified confidence: medium
Founder biographies are partly public; complete executive bios are not.
not publicly verifiable confidence: low
No executive compensation or agreements are public.
| topic | public evidence | diligence read |
|---|---|---|
| Org and HRIS | Remote-first team across 30+/50+ countries and multiple hubs. | Need org chart, function/location, contractor classification. |
| Compensation/equity | Employee tenders disclosed, but stock plan and compensation not public. | Review retention, dilution, tax and tender participation. |
| Turnover/employee issues | No public attrition or employee-relations issues found. | Absence of public data is not verification. |
not publicly verifiable confidence: low
Employee tenders are public but stock plan and option pool are not.
not publicly verifiable confidence: low
No public employee-relations problems were identified, but HR review is required.
not publicly verifiable confidence: low
No public attrition data is available.
Public legal risk is material due to voice/biometric lawsuits, safety/misuse exposure and gaps in IP, contracts, insurance and regulatory evidence.
partially verified confidence: medium
Vacker was dismissed with prejudice, while 2026 Illinois BIPA voiceprint suits require current docket review.
| topic | public evidence | diligence read |
|---|---|---|
| Vacker litigation | Delaware docket filed 2024 and dismissed with prejudice in 2025; Bloomberg summarized voice-cloning allegations. | Historical but material IP/likeness risk signal. |
| 2026 BIPA suits | Loevy, CBS and Sifted report Illinois suits involving ElevenLabs and voiceprint/consent allegations. | Current high-priority legal risk; verify docket status. |
| DSA/safety reporting | Safety page lists EU DSA and law-enforcement request channels. | Review regulatory correspondence and response SLAs. |
not publicly verifiable confidence: low
No affirmative lawsuits by ElevenLabs were found in public sources reviewed.
not publicly verifiable confidence: low
No public EHS liabilities were found; global offices/remote work require standard compliance review.
partially verified confidence: medium
Voice cloning, lawsuits and product breadth make training data, consent and IP diligence critical.
| topic | public evidence | diligence read |
|---|---|---|
| Training data and voice consent | Cloning products and lawsuits make provenance material. | Request data inventory, licenses, consents, retention and takedown logs. |
| Security and insurance | Safety page public; SOC/ISO, pen tests, cyber/E&O/media policies not public. | Request security reports, insurance, exclusions and claims history. |
| Material contracts | Customer, strategic, GPU/cloud and supplier terms not public. | Request MSAs, DPAs, SLAs, side letters, supplier commitments and termination rights. |
not publicly verifiable confidence: low
Insurance policies and exclusions are not public despite litigation and biometrics risk.
not publicly verifiable confidence: low
Material customer, strategic, supplier and side-letter terms are not public.
partially verified confidence: medium
No agency action was found, but DSA reporting, BIPA, global AI/privacy and misuse risks require counsel review.
| ID | Claim | Status | Sources |
|---|---|---|---|
| EC-001 | ElevenLabs has an active website positioning the company as an AI voice generator and voice agents platform. | verified high | SRC-001 |
| EC-002 | The company describes three platforms: ElevenAgents, ElevenCreative, and ElevenAPI, plus safety and impact programs. | verified high | SRC-002 |
| EC-003 | Series B public source says ElevenLabs raised $80M, had millions of users, generated 100+ years of audio, grew from 5 to 40 employees, and was used by employees at 41% of Fortune 500 companies. | partially verified medium | SRC-003 |
| EC-004 | Series C public source says ElevenLabs raised $180M at a $3.3B valuation with $281M total funding, 120 employees, >60% Fortune 500 employee usage, and major product milestones. | partially verified medium | SRC-004 |
| EC-005 | The September 2025 tender source says $100M of employee liquidity at a $6.6B valuation, >$200M ARR, expected >$300M year-end ARR, enterprise revenue >200% YoY, and 330+ team members. | partially verified medium | SRC-005 |
| EC-006 | The February 2026 Series D source says ElevenLabs raised $500M at an $11B valuation, reached $781M total funding, closed 2025 with >$330M ARR, and named major customers and expansion plans. | verified high | SRC-006 |
| EC-007 | CNBC independently corroborates the $500M raise at $11B valuation and notes ElevenLabs is London-based and eyeing a potential IPO. | verified high | SRC-007 |
| EC-008 | The May 2026 company update says ElevenLabs ended 2025 with $350M ARR, surpassed $500M ARR in the first four months of 2026, closed another tender, and had 530 teammates across 50+ countries. | partially verified medium | SRC-008 |
| EC-009 | ElevenLabs publicly lists Free, Starter, Creator, Pro, Scale, Business, and custom Enterprise pricing tiers using credits. | verified high | SRC-009 |
| EC-010 | Documentation describes TTS, STT, voice cloning, conversational agents, generative audio, SDKs, 10,000+ voices, eleven_v3 70+ languages, Flash ~75ms, and Scribe 90+ languages. | verified high | SRC-010 |
| EC-011 | Safety page describes safeguards to prevent, detect, enforce, and inform misuse, including C2PA-related work, classifiers, vetting, verification, reporting, and law-enforcement channels. | partially verified medium | SRC-011 |
| EC-012 | Careers page says ElevenLabs is remote-first with a global team spanning 30+ countries and hubs in New York, London, and Warsaw. | partially verified medium | SRC-012 |
| EC-013 | Expansion blog reports UK, US, and European headcount growth, multiple offices, NVIDIA Blackwell GPU usage, and impact-program metrics. | partially verified medium | SRC-013 |
| EC-014 | CourtListener records Vacker v. ElevenLabs as filed in Delaware in 2024 and dismissed with prejudice by stipulation in 2025. | verified high | SRC-014 |
| EC-015 | Bloomberg Law summarized Vacker allegations that ElevenLabs allegedly cloned or misappropriated voice actors likenesses. | partially verified medium | SRC-015 |
| EC-016 | Loevy + Loevy announced 2026 class actions against AI voice companies including ElevenLabs alleging BIPA voiceprint violations without consent. | partially verified medium | SRC-016 |
| EC-017 | CBS/Capitol News Illinois reported nine Chicago federal class actions against companies including ElevenLabs over alleged voice theft to train AI. | partially verified medium | SRC-017 |
| EC-018 | Sifted excerpt reports seven journalists and voice actors sued ElevenLabs in Illinois alleging voice models used recordings without consent. | partially verified medium | SRC-018 |
| EC-019 | The unicorn list source identifies ElevenLabs as AI Audio technologies with $11B valuation dated February 2026 and country Poland. | verified medium | SRC-019 |
| EC-020 | Competitor pages show Google Cloud, Amazon Polly, OpenAI and Deepgram offer public TTS or voice/audio capabilities. | verified medium | SRC-021SRC-022SRC-020SRC-023 |
| EC-021 | Endeavor profile identifies Mati Staniszewski and Piotr Dabkowski as Polish cofounders and describes 2022 research-company origins building TTS models. | partially verified medium | SRC-024 |
| EC-022 | ElevenLabs author page says Mati is cofounder, leads AI teams, previously worked at Palantir, grew up in Poland, and studied Mathematics at Imperial College London. | verified medium | SRC-025 |
| EC-023 | No audited financials, cap table, customer concentration, contracts, security audit, insurance schedule, or full legal docket were public in reviewed sources. | not publicly verifiable high | SRC-003SRC-006 |
| EC-024 | Current website, May 2026 update, and CNBC/Series D IPO language support active private/not-yet-IPO status; no public IPO, acquisition, or shutdown was identified. | verified high | SRC-001SRC-007SRC-008 |
| EC-025 | Customer logos and Fortune 500 usage are mostly company-published and do not reveal revenue by customer, contract term, renewal status, or concentration. | not publicly verifiable high | SRC-006SRC-008SRC-013 |
| EC-026 | Public sources indicate NVIDIA investor/infrastructure relevance, but cloud/GPU/data supplier contracts and unit costs are not public. | partially verified medium | SRC-008SRC-013 |
| EC-027 | Product breadth across agents, creative tools, API, STT/TTS, cloning, dubbing, music, image, and video creates execution and control complexity. | verified high | SRC-001SRC-002SRC-010 |
| EC-028 | Misuse/deepfake risk remains material because the company operates voice cloning and generative audio despite public safeguards. | partially verified medium | SRC-010SRC-011 |
| EC-029 | Global expansion and EU DSA reporting create multi-jurisdiction regulatory, privacy, AI, employment, and biometric compliance complexity. | partially verified medium | SRC-006SRC-011SRC-013 |
| EC-030 | At $11B valuation and >$500M ARR, the public screen is roughly 22x ARR before verifying margins, retention, customer concentration, and legal risk. | partially verified medium | SRC-006SRC-008 |
This report is a public-evidence diligence snapshot, not investment advice. Important financial, legal, technical, and contractual facts remain non-public and should be verified directly with management and primary documents before any investment decision.