06 · Business Model & Monetization
Tax Labs · Strategic Design & Analysis
Confidential · Aug 2026 · v0.1

Business Model & Monetization

Tax Labs runs three revenue levers at once — usage covers the cost of doing the work, licensing makes revenue predictable, and equity captures the upside of every vertical it helps build. Blended and anchored on a co-build Studio tier, this turns an infrastructure business into a cash-flow-positive venture studio: transaction fees pay the engineers; equity is the return.

~€35
Cost-to-serve / truck / yr at 500 trucks — falling to ~€11 at 10k 9
50 → 89%
Transaction gross margin as the shared platform scales 9
15–20%
Equity per co-built vertical — the Studio-tier anchor 1
€1.5M
Year-5 recurring cash revenue at ~89% margin Lo
~$45M
Expected equity value across 10 verticals (illustrative) Lo
85–95%
Cost saved vs a vertical building the stack itself 16
The recommendation, in one line

Adopt a three-tier blended model anchored on the Studio tier: 15% equity + cost-coverage usage fees for new co-built verticals; a pure Platform licence (usage + subscription, no equity) for operators who already have a team; and an Enterprise white-label for incumbents. Usage keeps the lights on, licensing floors the revenue, and equity is where the money is made.

6.1 The three-lever model

Each lever does a different job. Usage recovers the marginal cost of every document, filing and bank read. Licensing turns lumpy usage into a predictable floor. Equity converts Tax Labs from a vendor into a co-investor aligned with each vertical's long-run success. The art is the blend — and which lever leads depends on the tier.

A · Usage — covers cost

Per-document, per-filing, per-bank-read. Scales with work done; passes variable cloud + API cost through at a margin. Defensible even to skeptics: "you pay only when we do work." The model of Stripe, Twilio, AWS.

covers fixed + variable cost

B · Licensing — predictability

A monthly platform fee covers access to auth, RAG-over-legislation, confidence scoring, the HITL queue, dashboards and white-labelling. It floors the economics even in low-volume months and gives revenue visibility.

revenue predictability

C · Equity — upside & alignment

Tax Labs is the technical co-founder. For early verticals without a tech team it takes 15–20% in exchange for building and running the stack — the in-kind value of a 2–4 person team costing €300–600k/yr.

upside + alignment

USAGE per-doc · per-filing · per-read LICENSING monthly platform subscription EQUITY 15–20% of the vertical entity Covers cost-to-serve fair to every operator size Revenue predictability floors low-volume months Upside + alignment grow the vertical, not just bill it Each lever aligns Tax Labs with a different party — usage with the payer, licensing with operational stability, equity with the long-run outcome.
The three levers are not alternatives — they are complements, each carrying a distinct economic role. Tier design chooses which lever leads.
So what

No single lever is right for every customer. Usage alone under-captures a vertical that scales to a €100M exit; equity alone scares off operators who already have a team. The blend — tuned per tier — is what lets Tax Labs serve the whole market without leaving money, or alignment, on the table.

6.2 Packaging — three tiers

One platform, three commercial wrappers. Studio is the recommended default for new verticals; Platform for existing operators; Enterprise for incumbents.

Tier 1 · Platform no equity

Who: operators with a team switching infrastructure or adding verticals.

€1,500/mo base (€1,200 annual)
€0.80 / invoice · €2.00 / filing
€0.30 / bank-account/mo · €1.50 / HITL doc
RAG ≤10k queries included · 48hr SLA

Tier 2 · Studio recommended

Who: entrepreneurs launching a new vertical with no tech team. Tax Labs is the CTO.

€0/mo licence (waived 24 mo)
Usage at cost + 20% (cost-coverage only)
Full technical build + weekly advisory
15–20% equity · 3-yr vest, 1-yr cliff

Tier 3 · Enterprise white-label

Who: fuel-card, payroll & HR platforms wanting the stack under their own brand.

€60k–€250k/yr licence (custom)
Per-unit rate 40–60% below Tier 1
Dedicated infra · 99.9% SLA · 4hr response
White-label rights · no equity

ElementTier 1 · PlatformTier 2 · StudioTier 3 · Enterprise
Price mechanicLicence + usageEquity + cost-coverage usageAnnual licence + volume usage
Platform licence€1,200–1,500 / mo€0 (waived 24 mo)€60k–250k / yr
Per-invoice processing€0.80cost + 20%40–60% below T1
Per-filing€2.00cost + 20%negotiated
Technical build & advisory✓ included~ integration team
Equity taken✗ none15–20%✗ none
Lead leverUsage + LicenceEquityLicence

Studio migrates to Tier 1 at a 20% "portfolio company" discount after 24 months. Tier benchmarks anchored to Mindee €49–€649/mo 13 and Veryfi's $500/mo minimum 14. Md

So what

The same infrastructure is monetised three ways so no customer is turned away. New verticals default to Studio (equity-led); established operators buy Platform (usage-led); incumbents license Enterprise. One build, three markets.

6.3 Unit economics — one vertical customer

Reference: the EU truck fuel-VAT vertical (Studio tier), at 500 trucks. Each truck recovers ~€6,000/yr; the operator charges a 15% success fee 7 → €900 gross revenue/truck. Tax Labs sits underneath at <8% of that.

€6,000
Recovered / truck / yr (end-customer value) 6
120
Documents processed / truck / yr (24 batches × 5) Md
~€35
Tax Labs cost-to-serve / truck / yr at 500 trucks Md
Cost line (per truck / yr)Unit costQty€ / truckSource · conf
LLM inference — invoice extraction~$0.003/inv120~€0.36Claude Haiku ($1/$5 per M tok) 11 Hi
OCR / document preprocessing€0.005/pg240 pg€1.20Mindee Pro rate 13 Hi
Open-banking reads€0.50/mo12€6.00Tink standard 10 Md
Tax-authority API / filing~€0.50/filing2€1.00AEAT free; wrapper compute Lo
Compute (serverless) + storage~€0.12AWS Lambda + S3 15 Hi
RAG / legal queries€0.002/q50€0.10Embedded token cost Md
HITL review (10% escalation)€1.50/doc12€18.00Freelance reviewer, 5 min/doc Md
Allocated platform fixed costshared€8.00§6.6 allocation Md
Total cost-to-serve / truck / yr~€34.80at 500 trucks

Key insight: HITL (human review) is ~52% of cost-to-serve — the single biggest line, and the one that automation compresses most as training data accumulates.

Tax Labs revenue / truck (Studio)

Stream€ / truck / yr
Usage fees (cost + 20% margin)~€41.76
Licence allocated (post Y2)~€29
Total cash / truck / yr~€70–100
+ Equity upside15–20% × exit

Cash gross margin ~20–40% early (HITL-heavy), scaling to 60–70% as escalation drops. Equity is the real return.

The full truck-economics stack

€6,000  recovered / truck / yr
€900  operator fee (15%)
├ Tax Labs usage: €42–71
├ Tax Labs licence: ~€29
├ operator S&M / ops: €150–300
net €500–700 to operator ✓
Tax Labs: ~€70–100 cash + equity upside

Tax Labs takes <8% of the operator's revenue in cash — well within the 10–15% infra-COGS benchmark — and leaves a strong net contribution to the operator.

So what

At the level of one truck the model is unremarkable — a few tens of euros of cash. The point is that this repeats across thousands of trucks and dozens of verticals on shared fixed cost, and every one carries an equity option. Unit economics are the floor; leverage and portfolio are the return.

6.4 Operating leverage — the cost curve

Shared fixed cost is amortised across every truck on the platform, and HITL escalation falls as cross-vertical data trains the models. Cost-per-truck collapses; margin approaches SaaS quality.

€40 €30 €20 €10 €0 COST / TRUCK / YR 5002,0005,000 10,00020,000+ TRUCKS ON PLATFORM (all verticals, log scale) Build-it-yourself ≈ €200 / truck (off-scale ↑) €35€25€17€11€8 50% gm64%76%84%89% Tax Labs — shared platform the gap = the economic moat
Two forces compound: fixed cost per truck falls from €8 → €0.40, and HITL escalation drops 10% → 2% as models train on cross-vertical data. A single vertical is stuck near €200/truck; the shaded gap is what a shared platform captures. Illustrative model.
Trucks on platformFixed cost / truckHITL rate → costTotal / truck / yrGross margin on €70 cash
500€8.0010% → €18.00~€35~50%
2,000€3.007% → €12.60~€25~64%
5,000€1.505% → €9.00~€17~76%
10,000€0.803% → €5.40~€11~84%
20,000+€0.402% → €3.60~€8~89%

Illustrative operating-leverage model. Md — unit costs are sourced; escalation-rate improvement and fixed-cost allocation are internal estimates.

So what

This is the whole argument in one curve. A single vertical at 500 trucks runs ~50% margin; a platform serving 10,000 trucks across five-plus verticals reaches 84–89% — SaaS-quality economics — because fixed cost becomes essentially free and HITL automation compounds with data. No single-vertical operator can reach the right-hand side of this chart.

6.5 Portfolio equity — venture-studio optionality

Small equity across many verticals = recurring cash plus venture upside. Even with modest per-vertical odds, the portfolio expected value is material. Illustrative model — venture outcomes are estimates.

Per-vertical expected value Lo

10 verticals over 5 years · Tax Labs holds ~15% (post-dilution) · outcomes below.

ScenarioExitTL 15%PEV
Material exit$30M$4.5M30%$1.35M
Scale$100M$15M20%$3.00M
Modest$2M$0.3M40%$0.12M
Failure$0$010%$0
EV / vertical$4.47M

The combined picture Lo

~$45M
Expected equity value across 10 verticals
€1.5M
Year-5 recurring cash (usage + licences)
~89%
Cash gross margin at 20,000 trucks
$20–30M
NPV of equity portfolio (15% disc, 6yr)

The infrastructure revenue funds the engineering team; the equity portfolio is the return on that investment. Tax Labs is, in effect, a cash-flow-positive venture studio — recurring revenue underneath, portfolio optionality on top.

Illustrative portfolio model, independent outcomes assumed. Equity value depends on venture outcomes years 4–8 and is the lowest-confidence figure in this dossier. Lo

So what

Any one vertical may fail — 10% of the model assumes exactly that. But across ten independent bets, small stakes compound into a portfolio worth more than the entire recurring business, on top of that recurring business. That optionality is what a pure-SaaS pricing model can never capture.

6.6 Cost structure & the build-it-yourself counterfactual

Fixed vs variable split

Cost categoryEarly
~1k trucks
Mature
~10k trucks
Platform engineering€200k€350k
AI / RAG infra€30k€50k
Tax-authority connectivity€40k€80k
Cloud base€15k€30k
Legal / compliance€20k€40k
Fixed / semi-fixed~€305k~€550k
Variable (LLM·OCR·OB·HITL)~€13k~€130k
Total~€318k~€680k
Cost / truck€318€68

Fixed cost dominates early and is shared across all customers — the source of the operating leverage in §6.4. Md

The single-vertical counterfactual

What a vertical operator spends to build the same stack alone, Year 1:

Build-it-yourself lineYear 1
CTO hire€120–180k
2 engineers€160–240k
OCR · LLM · cloud€30–50k
Tax-authority integrations (one-time)€100–200k
Total Year-1 tech cost€430–770k
The pass-through

Tax Labs charges a 500-truck vertical €35–50k/yr cash + equity for an equivalent, better, shared-data stack. Against €430–770k build cost, that is an 85–95% cost reduction 16 — the operator's core reason to buy.

A 10-vertical platform spends ~€680k/yr total vs €4.3M–7.7M if each of 10 verticals built independently. Tax Labs captures ~€1–2/truck in platform margin and passes ~€120–140/truck in savings to each operator. Md

6.7 Alignment, defensibility & handling equity resistance

Why the blend is defensible
  • Incumbents can't copy it. Eurowag, VAT IT, Taxback are service businesses; re-platforming as a multi-tenant API cannibalises their own service margins.
  • Pure-API players can't either. Stripe/Plaid lack tax-authority connectivity, EU-country regulatory depth, and HITL expert review.
  • The data network effect compounds. Cross-vertical fraud signals and confidence models trained on hundreds of thousands of claims are unreproducible by any single vertical. Platform §5
Handling customers who resist equity

"We don't want to give up equity." Three structured responses:

  • Pure Platform tier: no equity, higher usage rates, 100% ownership retained — but no co-build or advisory.
  • Delayed warrant: equity vests only if the vertical reaches €500k ARR — removes founding-day dilution anxiety, aligns on growth.
  • Cash buyout option: operator can buy out the stake at a pre-agreed formula (e.g. 3× revenue) within 3 years.

The message: equity is for co-builders, not for platform users. Want infrastructure? Pay for it. Want a technical co-founder? Share the upside.

Fallback for resisters who still want alignment: modified Tier 1 at €2,500/mo + a 5% success-fee kicker on recovered amounts — alignment without dilution.

Decisive recommendation & the single biggest risk

Default to Studio (15% equity + cost-coverage usage) for new verticals; Platform for existing operators; Enterprise for incumbents. This captures upside asymmetrically, aligns Tax Labs with each vertical's success rather than its API volume, and builds a venture-studio cash flywheel. Biggest risk: downstream funding dilutes Tax Labs's stake — mitigate with pro-rata rights, clear vesting & anti-dilution terms, and a pre-agreed conversion to pure Tier 1 pricing if a large round makes the equity relationship complex.

Sources: 1 Hexa studio deal terms — hexa.com/start/deal (hi). · 2 Venture-studio equity models — TFSF Ventures (md). · 3 Pioneer Square Labs model — BMC (md). · 4 Antler residency terms — RocketDevs (md). · 5 Fractional / co-founder CTO equity — Pangea.ai · DigitalDefynd (md). · 6 Recovery per truck ~€6,000 — Market §3 / Eurowag (md). · 7 Success-fee ranges (10–25% VAT; 5–12% excise) — MyTaxRebate · VAT IT (md). · 9 Cost-to-serve & operating-leverage model — internal, built on lines below (md/lo). · 10 Open-banking €0.50/user/mo — Tink via Finexer · Plaid (md). · 11 LLM inference — Claude · OpenAI pricing (hi). · 12 Stripe / Twilio usage precedents — Stripe · Twilio (hi). · 13 OCR per-page & SaaS tiers — Mindee (hi). · 14 Invoice OCR / min. commit — Veryfi (hi). · 15 Compute + storage — AWS Lambda / S3 (hi). · 16 Build-it-yourself counterfactual & savings — internal cost model (lo).
Previous← 5 · Platform Architecture
Tax Labs · Confidential06 · Business Model