Written from 50 named sources · Nov 5 · first result Fact-Checking and Source-Verification Report: Global Voice-Over & AI Synthetic Speech Industry (2017–2026) Prepared for: Dan Price, Founder, BionicVO Domain: Media Economics, Market Intelligence, and Speech Technology Temporal Horizon: 2017–2026 (CY2026 preliminary through November 1, 2026) Methodological Standard: Open-Source Intelligence (OSINT) and Chain-of-Verification (CoV) Framework Scope, Boundary Definitions, and Anti-Double-Counting Protocols Core Human Voice-Over (TAM Boundary 1): Gross market spend on professional spoken-word voice performance, including talent session fees, residuals, usage/buyouts, casting commissions, and full-service creative localization where voice adaptation and direction are bundled into talent delivery. Factual Status: Modeled / Estimated — analytically coherent MECE construct, but no supplied aggregator respects it. Market.us explicitly includes AI-driven Voice-over and Dubbing inside its $4.2B [3], and Dataintelo's $5.8B explicitly spans Software, Hardware, and Services [33]. Published totals therefore cannot be decoupled from engineering/software charges without proprietary report access. Explicit Exclusions: Standalone audio post-production engineering, physical studio rental, mastering fees, Foley, non-performance translation, and general audiovisual localization pipeline charges not bundled into voice production fees. AI Speech Technology (TAM Boundary 2 — Carve-Out): Synthetic voice licensing, voice cloning APIs, enterprise Text-to-Speech (TTS) SaaS, and generative audio software treated as distinct, non-additive submarket. Factual Status: Modeled / Estimated protocol. The protocol is sound in principle but was violated in the prior draft by summing human + AI without deducting overlap. This report corrects that: combined figures are presented as gross non-additive upper bound with explicit overlap deduction. Factual Status Taxonomy: Verified: Reconciled against audited financials, statutory filings, registered collective bargaining agreements, or official rate sheets. Triangulated: Derived from multi-vendor consensus across independent market intelligence databases. Modeled / Estimated: Synthesized via top-down allocation or bottom-up capacity modeling; midpoint of conflicting aggregators. ND (Not Determinable): Data withheld or unavailable; proxy substituted with disclosure. Executive Summary: Decade-Long Trajectory (2017–2026) From Centralized Broadcast to Ubiquitous Synthetic Audio Between 2017 and 2026, the voice-over industry transitioned from an agency-brokered, studio-centric broadcast model to a distributed, software-mediated digital ecosystem. In the historical Tube Era, distribution was dictated by linear television networks, national radio syndication, physical theatrical releases, and centralized corporate training. Voice casting flowed through talent agencies in Los Angeles, New York, and London, anchoring high union day-rates and usage fees. The structural catalyst of this decade was Fragmentation of Digital Attention. The explosion of user-generated content platforms, direct-to-consumer OTT streaming, short-form vertical video, dynamic podcast advertising, and interactive gaming fractured media consumption. Vendor-blog proxy often cited: YouTube ingests ~360 hours of video every minute, with 75% of consumption on mobile screens [10]. Factual Status: Triangulated (weak) / quote-Verified, Low Confidence as 2025/2026 metric. The 360 hrs/min figure traces to ~2019-era YouTube disclosures and is repeated without disclosed methodology in a July 31, 2025 vendor blog [10]; it is not a verified 2025/2026 observation and must not be treated as primary telemetry. This fragmentation bifurcated buyer demand: The Volume Tier (Long-Tail Digital): Millions of continuous digital video units, corporate micro-learning modules, and programmatic ads requiring rapid turnaround and micro-budgets. This demand exceeded the financial and operational capacity of traditional human studios, establishing market pull for generative AI synthetic speech. The Premium / Resonance Tier: High-stakes theatrical animation, triple-A gaming, flagship brand identity campaigns, and culturally sensitive narrative dubbing. Here the premium on human prosody, emotional delivery, and authenticity intensified, protected by enforceable union floors. Commercial Sizing and Localization Drivers — Corrected for Scope Independent research values the global video localization services market at ~$3.5B currently, expanding toward ~$6.5B by 2033 [10]. Factual Status: Triangulated (weak) / quote-Verified — accurately quoted from Travod vendor blog, no disclosed methodology. The combined dubbing and voice-over market is reported at $6,182.8M ($6.18B) in 2026, expanding to $10,988.56M by 2035 at 7.4% CAGR [4]. Factual Status: Triangulated with provenance correction required. This figure belongs solely to MarketReportsWorld, Sept. 1, 2026 [4], not jointly to Market.us. Market.us separately valued the market at $4.2B in 2024, with North America 43.5% / $1.8B [3]. The two firms, years, and scopes differ and must not be conflated. The broad voice-assistant application space surged from $2.8B in 2021 to $11.2B in 2026 (32.4% CAGR) [7]. Factual Status: Verified as quoted; inadmissible as VO evidence. It measures enterprise software/IVR and assistants, not talent spend or synthetic-voice licensing; using it as any VO proxy is a category error. Strategic Addressable Opportunity for BionicVO (TAM / SAM / SOM) — Restated to Prevent Double-Counting For BionicVO, treating this sector through a legacy talent-only lens artificially compresses the serviceable market, but summing human + AI without adjustment overstates it: Total Addressable Market (TAM), gross non-additive upper bound: Combined gross expenditure on core human spoken-word production, full-service audio localization, and AI synthetic speech software platforms: Low $9.09B / Base $10.72B / High $12.09B in 2026. Factual Status: Modeled / Estimated, Medium Confidence. Base = $5.57B human (modeled midpoint) + $5.15B narrow TTS/cloning (triangulated consensus) [39][40][50]. TAM, net de-duplicated: After deducting ~$0.8B–$1.2B overlap of AI-driven dubbing already counted inside both Market.us [3] and Dataintelo [33] aggregators: Low $8.1B / Base $9.6B / High $10.9B in 2026. Factual Status: Modeled / Estimated, Medium Confidence. Prior $10.5B–$11.5B claim is retained only as gross upper bound, not net market size. Serviceable Addressable Market (SAM): Commercial advertising, narrative gaming, corporate enterprise voice, and high-tier media dubbing across US and Western European markets requiring hybrid human-AI workflows, compliance, and multi-language rights management: ~$4.2B in 2026. Factual Status: Modeled / Estimated, Low-Medium Confidence. Allocation percentages and capture-rate pipeline math undisclosed; investor-grade use requires bottom-up pipeline validation. Serviceable Obtainable Market (SOM): Immediate contract opportunity for production platform delivering enterprise-grade voice solutions, ethical synthetic licensing, and bundled multilingual adaptation: ~$120M–$180M over 36-month horizon. Factual Status: Modeled / Estimated, Low Confidence. Insufficiently documented without win-rate and sales-capacity model. Market Segmentation Matrix: Geography and Language Status To isolate regional concentration and cross-border expansion, the market is divided across a 2x2 matrix: United States vs. Rest of World (RoW) by English-Speaking vs. Non-English-Speaking. Critical audit disclosure: No primary audited source publishes mutually exclusive geography-by-language splits. The two supplied regional anchors actively contradict each other: Market.us reports North America 43.5% / $1.8B in 2024 [3], while Dataintelo reports North America only 28.7% with APAC leading at 36.2% / $2.1B in 2025 [33]. Cell shares, CAGRs, and dollar ranges below are therefore Modeled / Estimated midpoints of conflicting aggregators, not triangulated observations. Point estimates require ranges with explicit midpoint labeling. Market Segment Cell Revenue Share (2026 Base) CAGR (2021–2026) 2026 Sizing Range (Low / Base / High) Factual Status Methodological / Source Basis US / English-Speaking 34.5% 4.2% $1.70B / $1.92B / $2.15B Modeled / Estimated, Low-Medium Confidence Derived from US advertising and domestic media spend; cross-referenced against Market.us North American 43.5% baseline [3] adjusted for domestic non-English carve-outs. Reconciliation with Dataintelo 28.7% NA share [33] unresolved. US / Non-English-Speaking 5.5% 9.8% $0.24B / $0.31B / $0.38B Modeled / Estimated, Low Confidence US Hispanic broadcasting, corporate dual-language compliance, Spanish/Asian-language streaming dubbing within US borders. No trade-association or broadcast-filing support; high CAGR reflects demographic modeling, not observed revenue. Rest of World / English-Speaking 18.0% 5.6% $0.85B / $1.00B / $1.18B Modeled / Estimated, Low-Medium Confidence Aggregates UK, Canada, Australia, NZ, Anglophone African/Indian hubs. Scaled via UK equity agreements and Commonwealth e-learning production. Rest of World / Non-English-Speaking 42.0% 8.9% $2.05B / $2.34B / $2.68B Modeled / Estimated, Medium Confidence Primary driver of L10n. Anchored by European dubbing mandates (France, Germany, Italy, Spain), Latin American hubs, APAC OTT consumption [3][10][33]. Directionally supported by Perso single-platform evidence that English is #1 source 32.3% and #1 target 28.0% in AI dubbing Q1 2026 [2], with caveat below. Total Global Core Human VO 100.0% 6.7% $4.84B / $5.57B / $6.39B Modeled / Estimated, Medium Confidence Midpoint of conflicting aggregators: Market.us $4.2B (2024) [3], Dataintelo $5.8B (2025) [33], MarketReportsWorld $6.18B (2026) [4]. Scopes differ materially (Dataintelo includes hardware/software; Market.us includes AI-driven VO). Perso caveat: English 32.3% source / 28.0% target, 26.9% flows →English, and education 10.3% / animation 8.4% leading categories are Verified as quoted from Q1 2026 platform sample of tens of thousands of completed projects across 33 source/target languages [2], but Perso itself flags the sample as single-platform, creator-economy-biased and may not reflect the entire AI dubbing market. Must carry that caveat in any investor use. Year-by-Year Annual Trend Table (2017–2026) The table reconstructs the ten-year evolution of the global core human voice-over industry alongside the distinct AI voice technology market. 2017–2025: Reconciled historical years based on retrospective multi-vendor assessments. Factual Status throughout: Modeled / Estimated. No supplied source provides a year-by-year human VO series; anchors conflict and scopes differ. 2026 Preliminary: Data through November 1, 2026 (10 months observed, ~83.3% of calendar year). Annualization formula disclosed: Full-year equivalent = Jan 1–Nov 1 observed run-rate × (12/10), with Q4 seasonal uplift for advertising/entertainment (+3–5%) and straight-line for corporate/e-learning, cross-checked against Q1–Q3 earnings, strike-settlement wage escalations, and year-to-date AI venture run-rates. Presenting full-year equivalents without this formula overstates certainty; 2026 must be read as partial-year observation grossed up, not audited full year. Values in USD Billions ($B). Triangular bounds Low / Base / High. Year Core Total Human VO Advertising & Commercial Audiobooks & Narration Gaming & Animation Corporate & E-Learning Localization & Dubbing (Bundled) AI Voice Tech (Distinct Carve-Out, narrow TTS/cloning) Confidence Source Basis & Methodology 2017 $3.25 / $3.60 / $3.95 $1.05 / $1.15 / $1.25 $0.25 / $0.30 / $0.35 $0.45 / $0.50 / $0.58 $0.65 / $0.72 / $0.80 $0.85 / $0.93 / $1.02 $0.10 / $0.15 / $0.22 Medium Modeled / Estimated. Backcast from historic studio billings. Concatenative/parametric TTS dominant. 2018 $3.45 / $3.82 / $4.20 $1.10 / $1.20 / $1.32 $0.30 / $0.36 / $0.42 $0.50 / $0.56 / $0.64 $0.68 / $0.76 / $0.85 $0.92 / $1.01 / $1.12 $0.18 / $0.25 / $0.35 Medium Modeled / Estimated. Backcast using audio publishing growth and broadcast commercial budgets. 2019 $3.68 / $4.08 / $4.50 $1.15 / $1.26 / $1.38 $0.38 / $0.44 / $0.52 $0.55 / $0.62 / $0.70 $0.72 / $0.81 / $0.90 $1.00 / $1.11 / $1.22 $0.30 / $0.42 / $0.58 Medium Modeled / Estimated. Streaming wars catalyst (Disney+, Apple TV+); multi-language commissioning. 2020 $3.50 / $3.90 / $4.35 $0.95 / $1.08 / $1.20 $0.48 / $0.55 / $0.64 $0.60 / $0.68 / $0.78 $0.80 / $0.90 / $1.02 $0.90 / $1.00 / $1.12 $0.50 / $0.70 / $0.95 Medium Modeled / Estimated. COVID-19: physical studio shutdown offset by home-studio, gaming, remote e-learning. Annualization of dip undisclosed in sources. 2021 $3.85 / $4.30 / $4.78 $1.12 / $1.25 / $1.40 $0.58 / $0.67 / $0.78 $0.70 / $0.78 / $0.88 $0.88 / $0.98 / $1.10 $1.02 / $1.14 / $1.28 $0.85 / $1.10 / $1.45 Medium Modeled / Estimated. Rebound. Voice-assistant $2.8B benchmark [7] is Verified but inadmissible as VO proxy (enterprise software/IVR). 2022 $4.15 / $4.65 / $5.20 $1.20 / $1.33 / $1.48 $0.68 / $0.78 / $0.90 $0.78 / $0.87 / $0.98 $0.94 / $1.05 / $1.18 $1.12 / $1.25 / $1.40 $1.30 / $1.75 / $2.30 Medium Modeled / Estimated. Neural audio diffusion emergence; ElevenLabs founded 2022 [36]; early synthetic disruption in indies. 2023 $4.40 / $4.95 / $5.55 $1.24 / $1.37 / $1.52 $0.76 / $0.88 / $1.02 $0.85 / $0.95 / $1.08 $1.00 / $1.12 / $1.26 $1.24 / $1.38 / $1.55 $2.10 / $2.70 / $3.50 Medium Modeled / Estimated. Voices.com acquired Voices.ai April 2023 Verified [13]; SAG-AFTRA theatrical strike establishes digital replica bargaining protocols. 2024 $4.60 / $5.20 / $5.85 $1.28 / $1.41 / $1.56 $0.84 / $0.98 / $1.14 $0.88 / $0.98 / $1.12 $1.04 / $1.16 / $1.32 $1.35 / $1.51 / $1.70 $2.90 / $3.60 / $4.40 Medium Modeled / Estimated. Anchors: Market.us $4.2B [3] (includes AI-driven VO); TTS benchmarks emerging; SAG-AFTRA interactive strike begins July 26, 2024 Verified [38]. Prior $4.00B AI base revised down to consensus. 2025 $4.75 / $5.40 / $6.15 $1.30 / $1.44 / $1.60 $0.90 / $1.05 / $1.22 $0.92 / $1.04 / $1.18 $1.06 / $1.19 / $1.36 $1.45 / $1.63 / $1.85 $3.65 / $4.40 / $5.30 Medium Modeled / Estimated. Anchors: Dataintelo $5.8B [33] (includes software/hardware); TTS $4.14B [40], $4.8B [39], $4.25B [50]; SAG-AFTRA video game strike settlement Verified [38]. Prior $5.20B AI base revised down. 2026\* $4.84 / $5.57 / $6.39 $1.32 / $1.46 / $1.63 $0.94 / $1.10 / $1.28 $0.96 / $1.08 / $1.22 $1.05 / $1.18 / $1.35 $1.55 / $1.75 / $2.00 $4.25 / $5.15 / $5.70 Medium Preliminary through Nov 1, 2026. Modeled / Estimated human + Triangulated AI. Combined dubbing/VO $6.18B [4]; TTS $5.7B 2026 [39], $4.67B 2026 [40]; ElevenLabs ARR ~$600M Verified company-reported run-rate [36][42][44]. See annualization formula above. \2026 figures are annualized from 10 months observable data through November 1, 2026. Factual status: Triangulated industry baseline with econometric modeling for submarket allocations.* Submarket allocation audit note: The five-way 2026 split above is an unaudited allocation model (Modeled / Estimated, Low-Medium Confidence). Supplied sources give materially different shares: Dataintelo reports Film/TV 34.1% (~$1.98B), Advertising 17.6%, E-learning 16.8%, Gaming 13.7%, Corporate 11.3%, Others 6.5% [33]; Market.us reports dubbing 64.4%, human-based 58.2%, entertainment/media 32.5% [3]. Any investor use must present both taxonomies and reconcile that localization/dubbing (bundled) in this report maps roughly to Dataintelo Film/TV + portion of Services, not one-to-one. AI Voice Technology Deep Dive (2017–2026 / 2023–2026 Focus) Structural Shift: 2023–2026 Generative Acceleration Prior to 2022, synthetic speech was dominated by concatenative and parametric TTS, largely confined to automated telephony (IVR) and accessibility screen readers. Commercialization of neural audio codecs, diffusion architectures, and zero-shot voice cloning between 2023 and 2026 transformed synthetic speech into enterprise-grade media technology. By Q3 2026, pure-play market leader ElevenLabs achieved an annualized recurring revenue (ARR) run-rate of ~$600M (up from $330M–$350M in late 2025 and $100M in 2024 [36][42][43]), culminating in a September 2026 employee tender offer valuing the company at $22B [36][44][45][49]. Factual Status: Verified as company-reported (CEO to TechCrunch Sept 2026 [36][44]; Sacra independently estimates $600M by June 2026 [42]); annualized run-rate, not audited revenue — High Confidence for existence, Medium for precision. Enterprise clients now represent over 55% of its revenue mix [36][44][45]. Factual Status: Verified as stated on ElevenLabs blog [45] and corroborated [49]. ElevenLabs agents handle 15M+ conversations/week, 3× February level [45][49]. Concurrently, traditional voice platforms integrated synthetic infrastructure: Voices.com acquired Voices.ai in April 2023 Verified [13]; transaction value undisclosed; and deployed proprietary AI Studio [22], enabling actors to license cloned replicas alongside human auditioning [21][26]. Voices.com revenue scale remains Modeled / Estimated, Low Confidence / unresolvable without Canadian statutory filings: $123M [13] vs. $170.2M [22] for a private company; both secondary estimates, ~$47.2M / ~38% gap; no midpoint should be presented as fact. ENTERPRISE WORKFLOW ALLOCATION MATRIX ┌────────────────────────────────────────────────────────┐ │ CREATIVE COMPLEXITY │ │ LOW HIGH │ ├──────────────────────────┬─────────────────────────────┤ │ • Internal Corporate VO │ • Brand Identity Campaigns │ │ • Standard E-Learning │ • Triple-A Game Leads │ │ • IVR / Telephony Bots │ • Theatrical Animation │ │ • Localized News Clips │ • Emotional Narrative Reads │ │ │ │ │ [DISPLACEMENT VECTOR] │ [HUMAN RATE RESILIENCE] │ │ 100% Synthetic / TTS │ SAG-AFTRA Union Minimums │ │ Marginal Cost -> $0 │ + Overscale Protections │ └──────────────────────────┴─────────────────────────────┘ Economic Displacement vs. Supplementation Vectors Direct Displacement Vector (Low-Tier Commoditization) — Modeled Proxy, Directionally Plausible AI voice generation has severely disrupted entry-level, non-union voice work [12][30]. Factual Status: Modeled / Estimated, Low Confidence for quantification. Affected submarkets: Corporate explainers, internal compliance modules, lower-tier e-learning, regional digital promos, scratch tracks [12][30]. Dubbing Cost Equation: Academic and trade assertions claim generative AI reduces dubbing costs by 60% to 70% [20]. Factual Status: Modeled / Estimated, Low Confidence — single academic assertion [20] with no disclosed sample or time-and-motion methodology. Dataintelo cites up to 40% per-minute reduction on selected projects [33], and MRFR cites ~70% pipeline cost cuts [40]; range plausible only as modeled. Nimdzi supports only directional hybrid-workflow margin compression, not a fixed percentage [12]. The mathematical reduction in raw studio hours is genuine (replacing multi-day bookings with algorithmic lip-sync [20]), but net realized margin savings for Tier-1 entertainment are ~25% to 35% Modeled / Estimated with no audited studio budget behind it — constrained by human-in-the-loop adaptation, phonetic QA, cultural nuance tuning, and legal indemnity review [12]. Audition Notice Proxy Trends: Direct platform disclosures of audition counts remain proprietary [21]. OSINT indicators from casting hubs (Voices.com, Voice123, Backstage) suggest fewer public postings for corporate narration under $250, with clients migrating to enterprise TTS SaaS [21][22][25]. Factual Status: Modeled / Estimated — unsupported as quantitative claim. Sources [21][25][30] offer only anecdotal creator/forum commentary; no scraped time-series exists. Claim of marked drop and wholesale migration over-indexes on synthetic-hype narrative; requires Wayback/Common Crawl audit before any quantitative assertion. Directionally plausible (both platforms added AI voice options [21][26]) but not verified displacement rate. Union Rate Defense and Structural Stabilization — Verified Floors In contrast to erosion in unrepresented low-tier narration, top-tier talent established enforceable floors: SAG-AFTRA Commercials Agreement (2025–2026): Verified, High Confidence. Effective through March 31, 2026, traditional digital 4-week options stand at $469.92, 1-year digital buyouts at $1,315.76, digital-plus 4-week at $501.85 [5]; day-performer session fees at $551.00 [5]. Current Year 2 sheets (Apr 1, 2026–Mar 31, 2027) step to $488.72 / $1,368.39 / $521.92 and $643.00 TV session [9]. SAG-AFTRA Interactive Media Agreement & Strike (2024–2025): Verified, High Confidence. Performers struck July 26, 2024 [38]; on July 9, 2025 members ratified 2025 Interactive Media Agreement at 95.04% [38]. Contract locked in compounded 3% increases in November 2025, 2026, and 2027, doubled overscale caps, and raised retirement contributions to 17.5% by October 2026 (Oct 31, 2026) [9][14][38]. Precise 15.17% immediate step is Triangulated, Medium Confidence — plausible as compounded scale delta but not in supplied sources verbatim. Crucially, agreement mandates explicit informed consent and disclosure before any AI digital replica creation/utilization, establishing union-sanctioned pricing floor [9][38][47]. Video game voice rates Verified: $567.62 (1 hr) / $1,134.95 (4-hr day), stepping to $584.65 / $1,169.00 on Nov 1, 2026 (~3% step) [9][14]; P&H 17% rising to 17.5% Oct 31, 2026 [9]. Supplementation and New Revenue Streams — Verified Existence, Modeled Scale Digital Voice Licensing & Marketplaces: Through SAG-AFTRA/Narrativ pact (Aug 2024) and Replica Studios agreement [38], performers can license verified digital replicas at custom price points, earning passive royalties without physical sessions [21][26][47]. 2025 IMA requires separate written, reasonably specific consent for digital replicas, with defined Digital Replica / ICDR / Real-Time Generation categories and compensation triggers Verified [47]; strike-suspension-of-consent right noted [38]. Enterprise Conversational AI Agents: ElevenLabs and competitors report 10,400+ voice contributors earning $22M+ in marketplace payouts by mid-2026 Verified as company-reported [36][49]; TTS benchmarks triangulated: GMI $4.8B 2025 / 22.4% CAGR to $35.3B 2035 [39]; MRFR $4.14B 2025 / $4.67B 2026 [40]; AutoFaceless $4.25B 2025 with cloning $3.29B [50] (aggregator, modeled). Growth driven not merely by capturing human budgets, but by creating net-new operational demand (automated support, hyper-personalized marketing, real-time NPC dialogue, statutory accessibility under European Accessibility Act and US ADA Title II [39][40]) that human production could never economically service. Diverging Revenue Trajectories: Traditional Human VO vs. AI Synthetic Speech The visual model presents the structural divergence between traditional core human voice-over revenue and the rapid ascent of AI/clone-driven speech platforms from 2017 through preliminary CY2026. Human VO sustains steady, inflation-aligned growth anchored by high-end union rates, theatrical dubbing, and narrative media; AI voice software expands along a tech S-curve on a narrower, de-duplicated TTS/cloning definition to avoid double-counting AI-driven dubbing already inside human aggregators. Analytical Commentary on Divergence Traditional Human VO Curve ($3.60B in 2017 → $5.57B in 2026\*): Modeled / Estimated, Medium Confidence. Exhibits ~4.9% annualized growth. Sustained by non-English localization volume and union rate hardening in US/UK insulating top-tier performers [3][5][9][14][38]. Anchors sit between conflicting $4.2B (2024) [3], $5.8B (2025) [33], $6.18B (2026) [4] with irreconcilable scopes; series is midpoint, not observation. AI Synthetic Speech Curve ($0.15B in 2017 → $5.15B in 2026\* narrow definition): Triangulated (GMI/MRFR) for 2025–2026,