The EU leaves tens of billions in legitimate tax refunds unclaimed every year, and a fragmented family of recovery operators is forming to capture it — each rebuilding the same infrastructure. This section sizes the recovery pools, the operator universe Tax Labs sells into, and the enabling-technology markets the platform rests on — then models TAM / SAM / SOM bottom-up and anchors it to real EU-infrastructure exits.
Tax Labs does not chase the fraud gap — it targets legitimate but operationally unclaimed refunds. The €89.3B VAT gap is context; the operative pool sits below it.
The European Commission's DG TAXUD puts the 2022 VAT compliance gap at €89.3B — roughly 7% of expected VAT revenue1. That figure includes fraud and error and is not our target. Stripping it back to the legitimate, operationally-recoverable refunds — cross-border B2B VAT, excise rebates, T&E VAT, customs overpayments, per-diem relief — yields a working pool of ~€38–45B/yr, of which roughly €17–23B goes unclaimed each year2.
| Recovery pool (EU) | Recoverable / yr | Claimed today | Unclaimed opportunity | Conf. |
|---|---|---|---|---|
| Cross-border B2B VAT — Directive 2008/9/EC | €15–20B | €8–10B | €8–12B | md |
| Corporate travel & T&E VAT | ~€15B | ~€9B | €5–6B | lo |
| Per-diem / income-tax — cross-border workers | €3–5B | €1–2B | €2–3B | lo |
| Excise / diesel rebates — incl. €10.8B trucking pool | ~€2.5B | ~€1.5B | ~€1B | lo |
| E-commerce VAT — IOSS / OSS | ~€2B | ~€1B | ~€1B | lo |
| Import duties & tariff recovery | ~€1.25B | ~€0.5B | €0.5–0.75B | lo |
| Total recovery pool | €38–45B | €21–24B | €17–23B | lo-md |
Reading the table: the recoverable column is the total legitimate entitlement; the unclaimed column is the friction Tax Labs' operators monetise. Category estimates carry honest confidence flags — cross-border B2B VAT is the most defensible; the smaller pools are directional.
The pools above are the revenue substrate of Tax Labs' customers — the vertical operators. The platform's value is letting each operator reach its pool at lower cost and higher accuracy than a bespoke build. A ~€40B recoverable base with ~€20B chronically unclaimed is a deep, structural, friction-driven opportunity — not a fad.
Six primary verticals, each a business someone is building. All need the same backbone; incumbents serve the fat head and leave the mid-market and long tail exposed.
| Vertical | Recoverable pool | Who serves it today | The gap Tax Labs closes |
|---|---|---|---|
| Trucking / road-carrier fuel VAT · diesel excise · toll · per-diem reference |
€10.8B pool ~€1.6B fee TAM |
Eurowag, DKV / UTA, Telepass, VAT IT, Taxback, national gestores | OCR of paper diesel invoices in 24+ languages; per-country filing & per-diem rules; open banking to pre-fund refunds |
| SME / entrepreneur VAT cross-border & domestic input VAT |
€5–8B unclaimed |
Accountants (manual), VAT IT, Vatax, Taxeo, Mooncard, Yokoy, Pleo | Invoice OCR from email / PDF inboxes; RAG over per-country eligibility rules; 8th-Directive e-filing across 5–10 countries per SME |
| Cross-border & posted workers income-tax · per-diem · social contributions |
€2–4B ~2M+ posted workers |
Vialto, EY / PwC mobility, Blue Umbrella, Taxback | Payslip / HRIS OCR; dual home + host tax-authority connectivity; RAG over 75+ bilateral treaties; confidence scoring before filing |
| Import / customs & tariff duty overpayment · FTA preference · import VAT |
€0.5–1.25B refund opportunity |
Ryan, KPMG Trade & Customs, Descartes, Avalara, customs brokers | OCR of customs declarations (SAD); HS-code classification via LLM + RAG over Combined Nomenclature; CDS / TARIC connectivity |
| Corporate travel / T&E VAT foreign hotel · transport · conference VAT |
€5–6B unclaimed |
VAT IT (largest agent), Taxback, Amex GBT, Yokoy, Spendesk | High-volume receipt / folio OCR; per-category eligibility rules; batched filing to 20+ authorities; open banking card-to-receipt match |
| E-commerce VAT IOSS / OSS · marketplace VAT |
€1–2B ~15% CAGR |
Taxdoo (Pagero), Avalara, TaxJar (Stripe), hellotax, SimplyVAT | Marketplace API connectivity (Amazon, Shopify); OSS / IOSS filing across 27 portals; AI product-category VAT classification |
Pattern: every vertical shares an identical technical requirement stack — OCR/IDP, tax-authority connectivity, RAG over legislation, open banking, inference gateway, confidence scoring + HITL, and cross-customer intelligence. The tax logic is the niche; the technology is the commodity.
Incumbents are vertically siloed and serve the head of each market. None offers white-label, cross-vertical infrastructure. A vertical operator that builds this alone spends €500k–€2M and 18–24 months before shipping a production-grade version. Tax Labs delivers it on Day 1 — plus a cross-vertical data network effect no single operator can replicate.
Tax Labs is not a participant in these markets — it is a buyer and integrator of them. Their scale and trajectory validate the thesis: the ingredients are large, growing fast, and commoditising, so the durable advantage is the orchestration layer above them that only platform scale can fund.
| Enabling market | 2024 size | CAGR | How Tax Labs uses it | Conf. |
|---|---|---|---|---|
| IDP / OCR Mindee · Veryfi · Klippa · ABBYY · Textract | ~$2.5B global | 27–34% | Document ingestion backbone; fine-tuned per vertical doc type | md |
| Open banking (EU) Tink · TrueLayer · GoCardless · Yapily | €9.6B EU | ~24% | Transaction verification, card-to-receipt match, pre-financing | md |
| RegTech (EU) reporting · AML · KYC | €4.6B EU | 15–20% | Regulatory reporting & compliance filing layer | hi |
| Tax IT software (EU) Vertex · ONESOURCE · Sovos · Avalara | $3.9–5.2B EU | 7–10% | Filing layer underlies — not competes with — customer products | md |
| E-invoicing / ViDA infra structured invoice exchange | ~€1–2B EU | 25%+ est. | Structured invoice ingest & authority connectivity | lo |
| AI inference / LLM extraction · RAG · scoring | multi-$B global | fast | Intelligence layer; inference cost fell >90% (2022→2025) | hi |
Every ingredient Tax Labs needs is a fast-growing, commoditising market — IDP compounding at ~30%, open banking at ~24%, inference costs down >90% since 2022. At platform scale, Tax Labs buys each capability cheaper per unit than any single operator, then adds the one thing money can't buy alone: cross-vertical intelligence. Cheap, improving inputs make the platform's unit economics improve without a pricing change.
Tax Labs is a B2B platform: its revenue comes not from taxpayers but from vertical operators. We size bottom-up (operators × platform revenue per operator) and cross-check top-down (share of the EU fee market). The bottom-up model is more defensible at this stage.
Tax Labs sells to vertical operators — existing agents plus the new startups the platform itself makes possible. Across six verticals: ~110–205 active today, plus ~95–160 enabled new entrants = ~205–365 total addressable operators md.
| Vertical | Active | New | Total |
|---|---|---|---|
| Trucking | 30–50 | 20–30 | 50–80 |
| SME VAT | 20–40 | 30–50 | 50–90 |
| Posted workers | 10–20 | 10–20 | 20–40 |
| Customs | 15–25 | 10–15 | 25–40 |
| T&E VAT | 15–30 | 15–25 | 30–55 |
| E-commerce | 20–40 | 10–20 | 30–60 |
| Total | 110–205 | 95–160 | 205–365 |
A hybrid model — usage (€0.10–0.50/document), licensing (€2k–10k/month), and equity — blends to ~€80–120k/yr per early-stage operator. Compounding operator count drives the ramp lo.
| Year | Operators | Avg rev | Platform revenue |
|---|---|---|---|
| Y1 · 2026 | 1–3 | €40–60k | €40–180k |
| Y2 · 2027 | 4–10 | €60–80k | €240–800k |
| Y3 · 2028 SOM | 10–20 | €80–120k | €0.8–2.4M |
| Y4 · 2029 | 20–40 | €100–150k | €2–6M |
| Y5 · 2030 | 40–80 | €120–200k | €4.8–16M |
ViDA tailwind (§3.5 comps) accelerates Y4–Y5 as thousands of businesses digitise invoicing.
The honest number is not the €500M–1B ceiling — it's the €640k–1.2M SOM at 8–15 operators by Year 3, scaling to €4.8–16M at 40–80 operators by Year 5. At €5–10M ARR with 30–50 operators, Tax Labs is Series-A-compelling; at €20–50M ARR, acquisition-grade. Revenue estimates are modeled, not tested — the pilot's job is to validate pricing and deal structure.
Infrastructure-layer companies serving fragmented vertical markets share a consistent playbook and an attractive exit profile. One of them is a direct EU comparable.
| Company | Domain | Exit / valuation | Customers | Lesson for Tax Labs |
|---|---|---|---|---|
| Stripe | Payments infra | ~$65B (2023) | millions | One horizontal serving many verticals beats any single vertical |
| Plaid | Bank data (US) | $13.4B raise (2021) | ~8,000 apps | The data layer captures more value than the apps it enables |
| Tink EU comp | Open banking (EU) | €1.8B (Visa, 2022) | ~150+ platforms | EU-native financial-regulatory infra exits at €1B+ scale |
| Mambu | Cloud banking core | $5.5B (2021) | ~200 fintechs | A B2B platform reaches unicorn status with a few hundred customers |
| Toast | Restaurant SaaS | ~$13B (IPO) | 120k locations | Deep vertical infrastructure compounds into a large business |
| Twilio | Comms infra | $60B+ peak | dev ecosystem | Usage-based pricing scales with each customer's own growth |
Tink already proved the thesis. An EU-native open-banking aggregator connecting 3,500+ banks to ~150 platform customers sold to Visa for €1.8B in 2022 — before reaching US-scale density. Tax Labs targets an adjacent, equally deep, and far less-served layer: EU-native regulatory infrastructure connecting tax authorities, banks, and document flows to recovery operators. Same customer profile, same monetisation, same regulatory tailwinds — a category no incumbent occupies.
A ~€40B recoverable base, six verticals of underserved operators, enabling markets compounding at 15–34%, and a proven €1.8B EU-infrastructure exit converge on one conclusion: the "tax-recovery infrastructure" category is real, sized, and open. The near-term prize is a defensible €640k–1.2M SOM; the long-term prize is the layer beneath an entire industry.