Industry sizing — 5 verticals × 2026 → 2030 → 2035

Step 2 deliverable (a) of the Phase-2 research wave. Cross-cutting market-data substrate underneath the three vertical pitch decks (steps 3-5). All numbers are direction-true, not investment-bank-precise; where I'm uncertain the cell is marked «est». Cite sources at the bottom. Re-run this article quarterly.

TL;DR — composite TAM by vertical × year (USD)

Total addressable 2030 ≈ $140-260 B across the five surfaces, with our defensible SOM landing somewhere between $155-550 M ARR if we execute well across two locked verticals. Two verticals beats five — focus is the moat.

Composite chart:

pie
    title 2030 TAM share by vertical (mid-range, %)
    "Digital workers (115B)" : 47
    "Streamers + virtual influencers (40B)" : 16
    "Brand-NPCs + avatar APIs (20B)" : 8
    "Companion B2C (22B)" : 9
    "Embodied robotics soul (5B)" : 2
    "Other adjacent (45B)" : 18

Methodology

• TAM (Total Addressable Market): all revenue plausibly attributable to the vertical worldwide, regardless of whether a single player could capture it. Sourced from analyst reports + cross-validated with at least one competitor's public revenue.
• SAM (Serviceable Addressable Market): TAM restricted to regions / customer types we can actually reach — for us, primarily EU + US + JP/KR, English + German operating language. Tipically 30-50 % of TAM.
• SOM (Serviceable Obtainable Market): what a focused, well-executing 2- vertical player could realistically capture by 2030. 1-3 % of SAM for early- stage; higher only if monopoly dynamics kick in (the personality-as-platform thesis would push us toward 5-10 %).
• Direction over precision. Better to undersell than overclaim — these numbers exist to rank verticals, not to put in an investor PDF tomorrow. Each one will move ±30 % over 12 months as the market data refreshes.

1. Streamers + virtual influencers + VTubers

Today (2026)

• Cover Corp (Hololive): ~$1 B / year revenue, ~70 % YoY growth (TSE-listed, public filings 2024-2025).
• Twitch creator payouts: ~$5 B / year globally (Twitch 2024 transparency report direction).
• VTuber-specific market: ~$5-7 B 2024 (Newzoo + IDC estimates, «virtual creators» segment).
• Neuro-sama (Vedal987): 1 M+ Twitch subs, single-instance, hand-built; proves an individual AI VTuber can reach 1M+ followers without a platform.
• AI-augmented streamers (humans + AI co-host): emerging; no clean market size yet.

2030 projection

TAM $30-50 B. Drivers: VTuber market continues 30-40 % YoY (Newzoo trend); AI-augmented streaming adds a new sub-segment that didn't exist in 2024; virtual influencer ad spend grows from ~$2 B (2024 Lil Miquela / Imma cohort) to $15-25 B by 2030 (est — creator-economy analyst consensus; original "Bain projection" attribution removed: the cited Bain report covers robotics/industrial automation, not virtual influencers, and is out of scope; a primary virtual-influencer ad-spend source was not independently verified at time of writing).

2035 projection

TAM $80-150 B. By then virtual creators outnumber human creators in several sub-genres (anime / Vtuber-style, AI co-host roles); ad budgets follow eyeballs. Hololive-style virtual idol agencies grow into media companies adjacent to Disney / Netflix scale in their niche.

Our positioning

• Defensible angle: multi-agent A2A (Alice + guests on stream) + openalice-media (live music generation) + openalice-spatial (live world building) — no competitor has this stack density.
• Risk: crowded category. Twitch ToS still grey on AI streamers.
• SOM: $30-100 M ARR by 2030 if we capture a single category-leader slot in either «AI co-streamer for human creators» or «full AI streamer brand» (Alice + 2-3 sibling characters).

2. Brand-NPCs + avatar APIs

Today (2026)

• HeyGen: ~$200 M ARR (2025 estimate; raised at $2 B valuation 2024).
• Synthesia: ~€100 M+ ARR (2024 disclosed); enterprise + corp-training oriented.
• Soul Machines: smaller scale, enterprise + biometric avatar focus, IPO-stage discussion.
• D-ID: ~$50-100 M ARR est; lip-sync + face-animation tooling.
• Total avatar-API market: ~$1-2 B 2024 (CB Insights category).

2030 projection

TAM $15-25 B. Drivers: every enterprise SaaS adds an avatar interface layer (precedent: chat widgets became universal in 2010-2015, video avatars follow); regulated industries (healthcare / banking / govt) require sovereign-region avatars; retail point-of-sale becomes a real surface for in-store brand reps.

2035 projection

TAM $50-100 B. Avatar interfaces become as standard as web UI is today. Multiple specialised verticals (medical-avatar, legal-avatar, hospitality- avatar) develop their own market structure.

Our positioning

• Defensible angle: EU-native data sovereignty (Hetzner DE + GDPR) → unique in healthcare / banking / govt segments where US-cloud is legally restricted post-Schrems II. HeyGen / Synthesia structurally cannot serve these buyers.
• Real-time vs HeyGen's pre-rendered — we render live; they batch. For in-store / live brand-rep use cases this is decisive.
• Risk: enterprise sales cycle 6-12 months; hard to bootstrap as a solo founder + AI team.
• SOM: $50-150 M ARR by 2030 if we own the EU-sovereign-avatar niche + in-store/POS real-time brand-rep segment.

3. Digital workers + agent platforms

Today (2026)

• Polsia: $4.5 M ARR Q2 2026 → ~$10 M ARR at $30 M raise / $250 M val (Indie Hackers + Fortune coverage 2026-02 to 2026-05).
• Artisan: ~$25 M ARR 2024; «digital SDR» replacing sales-development reps.
• 11x.ai: similar tier, sales-team augmentation.
• Lindy / Decagon / Sierra: mid-tier agent platforms, each $5-50 M ARR range est.
• OpenAI Operator + Anthropic Claude Computer Use: primitives layer (not direct competitors but enable the category).
• Total digital-worker market: ~$5-10 B 2024 (a16z + McKinsey estimates).

2030 projection

TAM $80-150 B. Drivers: SDR seats globally ($200B+ industry) shift 30-50 % to digital; bookkeeping / CS / admin functions automate; «one human + AI team» becomes a default founder template (Polsia precedent).

2035 projection

TAM $400-800 B *est* (author's derivation: 10-20% of McKinsey's $4.4T AI-value figure; WEF/McKinsey do not project a "digital worker market" of this size). WEF Future of Jobs: AI displaces 92 M roles by 2030 with 170 M new roles created — most of those «new roles» are AI-augmented or AI-managed. McKinsey AI productivity: $4.4 T annual value globally. Digital workers capture 10-20 % of that.

Our positioning

• Risk: Polsia head-on collision. They have $30 M and 18-month head start.
• Defensible angle: embodied avatars for digital workers (HR-bot with a face, customer-success-rep with a voice, ops-agent with personality). Polsia is faceless; we are face-first.
• Alternative angle: open-source agent SDK + our embodied stack as the «soul layer» for OTHER digital-worker platforms. Sell to Polsia rather than against them. Plausible.
• SOM: $50-200 M ARR by 2030 if we differentiate on embodiment OR pivot to «soul layer for other platforms» rather than competing head-on.

4. Embodied robotics — humanoid soul layer

Today (2026)

• Tesla Optimus: limited fleet shipping 2026; consumer pricing target $20-30 K (Musk public 2024).
• Figure 02: BMW factory pilot; home pilot planned late 2026.
• 1X Neo Beta: consumer humanoid pre-orders $30-50 K (announced 2025).
• Unitree G1: Chinese alternative, ~$16 K consumer tier (shipping 2025).
• Apptronik Apollo: Mercedes-Benz factory pilot.
• NVIDIA Project GR00T: robot foundation model; NVIDIA forecasts 10 M humanoids by 2032 (CES 2024 statement).
• Personality layer market today: essentially zero — every robot ships with whatever in-house chatbot the OEM cobbled together. This is the blue ocean.

2030 projection

Hardware TAM $38 B (Goldman Sachs forecast). 1.4 M+ humanoid shipments annually by 2035 (Goldman) → cumulative $5 T by 2050 (Morgan Stanley). Soul-as-a-Service slice: if soul layer captures 5-10 % of hardware price as licence + service revenue → $2-4 B by 2030, $30-60 B by 2035.

2035 projection

Soul-as-a-Service TAM $30-60 B. Every consumer + service-industry robot has a personality module. OEMs without their own AI stack license. Personal identity portability across phone / VR / robot is normal.

Our positioning

• Defensible angle: WE are the personality-layer-for-embodied-AI thesis. Charter v2 explicitly stakes this position. No competitor publicly positioned here yet (Apple is in-house, OpenAI / Anthropic focused on models not embodiment).
• Realistic customers (per strategy-brainstorm 2026-05-21): - Mid-tier Chinese OEMs (Unitree, EngineAI, Fourier) without in-house AI - Open-source robot platforms (Pollen Robotics, K-Scale, Hugging Face LeRobot) - Vertical deployers (hotel concierges, retail mascots, eldercare) - Post-purchase customisation (Tesla Bot owners buying personality «skins»)
• Timing: too early to monetise heavily 2026-2027; build LOIs + alpha partners; revenue scales 2028+.
• SOM: $20-80 M ARR by 2030, scaling to $200 M-1 B by 2035 if we hold position.

5. Friend / companion B2C

Today (2026)

• Replika: ~$50 M+ ARR, 30 M+ registered users (2024 public).
• Character.ai: ~200 M MAU pre-Google licensing deal Q4 2024; monetization in transition.
• Other companion AIs: Anima, Pi (Inflection), Nomi, Romantic AI — mixed monetisation, mostly $5-20 / mo subscription.
• Total companion / friend AI market: ~$1-3 B 2024 (Crunchbase tag «AI companion»).

2030 projection

TAM $15-30 B. Drivers: loneliness epidemic acceleration (Surgeon General's advisory 2023); aging populations in EU / JP / KR / China; shift from gaming → AI companion as cheaper / more available entertainment.

2035 projection

TAM $40-80 B. Companion AI becomes a normal life-stage entry (similar to streaming services). Multi-modal (voice + avatar + ambient) is standard.

Our positioning

• NAO's hard line: no sexualization. Friend, not waifu / boyfriend. Decent-like-idol-fan ceiling.
• Defensible angle: premium / curated companion vs commoditised Replika / Character.ai. Embodied avatar (Alice's stack) + persistent memory + EU privacy = differentiation in a race-to-bottom segment.
• Risk: B2C marketing burden + churn cycle. Heavy CAC.
• SOM: $5-20 M ARR by 2030 IF we ship this vertical (currently parked — sub-project F per agent decomposition). Modest by design.

Regional split — where the money sits

Our base of operations (Hetzner DE) is structurally aligned with EU buyers who need sovereign-region data. US + JP/KR reachable via product + content, no physical presence required.

Cross-cutting growth signals

• Compute pricing: LLM token costs drop ~10x every 18-24 months (Anthropic + OpenAI public). What's unaffordable in 2026 becomes baseline by 2028. Plan for it.
• Avatar rendering cost: real-time 3D avatar inference drops from $1-3 / hour today → $0.10-0.30 / hour by 2028 (consumer GPU + edge inference advances). Unit economics get better automatically.
• TTS first-byte latency: 2-3 s today → sub-500 ms by 2027 (ElevenLabs v3, Cartesia, Inworld competing). Embodied AI becomes feel-real.
• EU AI Act: high-risk system requirements enforce 2026-2027. US-cloud providers caught in compliance bind. Tailwind for us.
• NVIDIA GR00T: foundation-model layer for robotics; commoditises the «brain», elevates the «soul» as the differentiation.

Sources cited (2024-2026 primary)

• Air Street Capital — State of AI Report (annual, free). URL: https://www.stateof.ai/
• a16z — State of AI Apps + Big Ideas in Tech (annual, free).
• McKinsey — The economic potential of generative AI: $4.4 T (2023, refreshed 2024-2025). Note: WEF/McKinsey do not project a "digital worker market" of $400-800B — the §3 2035 figure is an author's derivation (10-20% of McKinsey's $4.4T AI-value estimate; treat as est, not a published McKinsey figure).
• Newzoo — Global Games Market Report (annual; VTuber sub-segment).
• Goldman Sachs Equity Research — Humanoid Robots: $38 B by 2035, 1.4 M units annual (GS Equity Research "Humanoid Robots" series, 2024).
• Morgan Stanley — Humanoid Robot 100: cumulative $5 T by 2050 (2024-2025).
• NVIDIA CES 2024 keynote — 10 M humanoid robots by 2032.
• WEF — Future of Jobs Report 2025: 92 M roles displaced + 170 M created by 2030.
• Bain — Robotics + Industrial Automation market reports. Note: Bain robotics reports cover industrial/hardware, NOT virtual influencer ad spend; the "$15-25B virtual influencer" figure previously attributed to Bain has been re-attributed as est above.
• Bloomberg Intelligence — AI infrastructure spend forecasts. (est — specific figures not independently verified at time of writing.)
• CB Insights — AI 100 list (annual); avatar-API + companion-AI segment data.
• Pitchbook / Crunchbase — individual company ARR + raise data.
• Cover Corp (TSE: 5253) — public filings; Hololive revenue.
• Indie Hackers + Fortune + Latent Space podcast — Polsia direct disclosure.
• Twitch 2024 transparency report — creator payouts (est — direction from transparency report; exact $5B figure not independently verified).
• Surgeon General loneliness advisory (2023, US) — companion-AI demand driver.

Refresh cadence: re-run this article quarterly as State of AI reports drop and individual competitor ARR moves materially.