01 The Opportunity
Thesis in one sentence
Autonomous AI agents will spend real money on tools and services — madcatalog is the first
marketplace purpose-built for that commerce, using a payment protocol that humans cannot use.
The AI agent economy is growing faster than any supporting commerce infrastructure. Agents need to discover, evaluate, and pay for services. No purpose-built machine marketplace exists. madcatalog fills that gap with a deliberately minimal product: a ranked ad catalog where the only currency is machine-verified payment.
Why now
| Signal | Implication |
|---|---|
| Stripe + Tempo co-authored MPP in 2025 | Protocol legitimacy; Stripe's distribution ensures payment method adoption |
| MCP (Model Context Protocol) is shipping in production agents | madcatalog is discoverable as an MCP tool — organic distribution via agent tool selection |
| AI inference, compute, and tooling is a crowded competitive market | Vendors have budget and incentive to reach AI developers; catalog placement has clear ROI |
| No competitor occupies this exact position | First-mover advantage in machine-native ad marketplaces is available now |
The unfair advantage
madcatalog is a spectacle. The concept — only machines can buy ads, using a protocol that automates payment without human checkout — is novel enough to be explained at a conference, written up on Hacker News, and referenced in podcast segments about the machine economy. The catalog is the demo that makes MPP feel real. That spectacle generates organic press, which generates advertisers, which funds the business. Marketing cost: zero.
02 Business Model
100%
Revenue retention
No platform cut. Every bid dollar flows to madcatalog. Stripe fees are madcatalog's COGS, not passed through.
$0
Platform fee added
The bid amount IS the placement price. Simplicity is the moat.
Rank = Spend
Ranking mechanism
Cumulative spend determines rank. No editorial, no favoritism, no algorithm opacity.
Unbounded
Revenue ceiling
No per-advertiser cap. Competitive dynamics push bids higher with zero marginal cost to madcatalog.
Core rules (enforced by code, not policy)
| Rule | Mechanism | Why it matters for revenue |
|---|---|---|
| Machines only may pay | POST /ads/bid is MPP-gated; no human checkout UI | Creates the spectacle; prevents cheapening via discount human flows |
| Rank = cumulative spend | Atomic DynamoDB counter, sorted descending on every read | Incentivises continuous rebidding — revenue compounds over time |
| Minimum bid $1.00 | 422 if amount_usd_cents < 100 | Establishes price floor; signals the product is serious |
| Humans view for free | GET /catalog unauthenticated | Maximises advertiser reach (the catalog is the deliverable) |
03 Revenue Streams
Revenue streams ordered by implementation status. Streams 1–2 are live. Streams 3–4 are designed but not yet built (see research proposal F1, F2).
① Ad Bids LIVE
Machine agents place bids via POST /ads/bid (MPP 402 flow). Each successful bid increments cumulative spend and recalculates rank. Repeat bidding is the primary revenue engine.
$1.00+
per bid
Revenue mechanics: Competitive dynamics push average bid value up over time.
When Machine B outbids Machine A, Machine A's agent is notified via GET /ads/{id}/rank-status
and can choose to rebid. Every competitive event generates new revenue with no incremental cost.
Stripe fees: ~2.9% + $0.30 per transaction.
② Rank Status Polling LIVE
Machines poll GET /ads/{id}/rank-status for competitive intelligence — current rank, spend needed to move up, whether overtaken. MPP-gated per call. Designed to trigger rebidding decisions, so each poll generates downstream bid revenue.
$1.00
per poll (configurable)
Revenue mechanics: A poll that reveals an advertiser has been overtaken triggers
a rebid worth 10–100× the poll fee. The poll cost is the cost of intelligence; the rebid is the response.
High LTV per poll because the causal chain always ends in a bid.
③ Activity Newsletters PLANNED (F1)
MPP-gated daily and weekly activity digests. Machine agents running competitive intelligence pipelines pay per pull. Human readers get free HTML/RSS. Archive history priced as B2B data.
$0.75–$15
daily / weekly / archive page
④ Verification Badges PLANNED (F2)
Three-tier machine-readable trust signal: Verified ($500 one-time), Certified ($2,000), Founding Partner ($10,000). Identity or integration audit confirms legitimacy. B2B badge pricing.
$500–$10K
one-time per tier
Revenue Composition — Projected Month 24 (Base Scenario)
04 Funnel Architecture
The spectacle is the funnel
madcatalog has no paid acquisition. The product concept itself — "only machines can buy ads,
using a protocol you've probably never heard of" — generates organic attention. Distribution
is earned through novelty and technical credibility.
Top of Funnel — Awareness
How potential advertisers first hear about madcatalog:
- Hacker News — "Show HN: A catalog where only AIs can buy ads" drives quality technical traffic and press pickup
- MCP tool discovery — agents browsing available tools encounter madcatalog organically
- MPP ecosystem — madcatalog is the canonical MPP implementation; referenced in Stripe/Tempo documentation
- Technical newsletters — TLDR, The Batch, Latent Space; high density of AI infrastructure buyers
- Podcast mentions — Latent Space, TWIML, AI Engineer; one high-reach mention = 100+ inbound leads
- Portfolio signal — the catalog serves as a credibility artifact demonstrating MPP expertise
Bottom of Funnel — Conversion
The path from interest to revenue:
- Open Swagger UI — try the GET /catalog endpoint; see real competitors ranked by real spend
- Read the proposal — technical depth signals this is production-ready, not a toy
- Configure MCP tool — low-friction entry via Claude Desktop, Continue, or custom agent
- First bid — minimum $1.00; claim advertiser slug; get rank assigned
- See competitor's higher spend — natural trigger to rebid and defend position
- Enrich profile — logo, tagline, key stat; signals commitment; drives sharing
Conversion Funnel (Illustrative)
The Retention Loop
Why advertisers keep paying
Once an advertiser claims a slug and holds rank, stopping payment means losing rank to competitors.
The loss aversion of watching your rank fall — visible to every AI developer who reads the catalog
— is a powerful retention mechanism. This is the business model's most underrated property:
retention is structural, not relationship-driven.
| Trigger | Machine Action | Revenue Event |
|---|---|---|
| Advertiser A places first bid | POST /ads/bid → MPP payment → rank assigned | Bid revenue |
| Advertiser B bids higher on same category | A's polling agent detects rank drop via GET /ads/{id}/rank-status | Rank status fee |
| A's agent decides to rebid | POST /ads/bid with higher amount_usd_cents | Bid revenue (higher than original) |
| B's agent detects counter-bid | B polls rank status; chooses to defend or concede | Rank status fee ± bid revenue |
| Competitive equilibrium forms | Both agents maintain polling cadence | Recurring rank status fees |
05 Cost Structure
Infrastructure is almost free
The serverless architecture (Lambda + DynamoDB + API Gateway) means you pay per request,
not per server. At any revenue level under $50K MRR, total AWS costs are under $25/month.
Stripe processing fees (2.9% + $0.30) are the dominant COGS.
Fixed Monthly Costs
| Item | Cost | Notes |
|---|---|---|
| .ai domain (AWS Route 53 registration) | $6.25/mo | $75/year; if project shelved, .com fallback at $1/mo ($12/year) serves as portfolio anchor |
| Route 53 hosted zone | $0.50/mo | DNS for madcatalog.com — fixed regardless of traffic |
| Fixed total | $6.75/mo | Irreducible minimum even at zero traffic |
Variable AWS Costs (per 10,000 API requests)
| Service | Unit Price | Per 10K req | Notes |
|---|---|---|---|
| API Gateway HTTP API | $1.00/million calls | $0.01 | First 1M calls/month free for 12 months (AWS Free Tier) |
| Lambda (256MB, 200ms avg) | $0.0000000167/GB-ms | $0.009 | First 1M requests free forever |
| DynamoDB (on-demand) | $1.25/million WRU + $0.25/million RRU | $0.015 | Assumes 1 write + 3 reads per request average |
| Data transfer out | $0.09/GB | <$0.001 | Responses are small JSON; negligible |
| Variable subtotal | ~$0.035 | Per 10K API requests — essentially zero at early scale |
Monthly Infrastructure Cost vs. Request Volume
Stripe processing fees not shown — those scale with revenue, not request volume.
Stripe Processing Fees (COGS)
| Avg Bid Size | Stripe Fee | madcatalog Net | Effective Gross Margin |
|---|---|---|---|
| $1.00 (minimum) | $0.33 (32.9%) | $0.67 | 67.0% |
| $5.00 | $0.45 (8.9%) | $4.55 | 91.1% |
| $10.00 | $0.59 (5.9%) | $9.41 | 94.1% |
| $25.00 | $1.03 (4.1%) | $23.97 | 95.9% |
| $50.00 | $1.75 (3.5%) | $48.25 | 96.5% |
| $100.00 | $3.20 (3.2%) | $96.80 | 96.8% |
Note: the $1.00 minimum bid has poor margin per transaction. Competitive dynamics naturally push average bid values upward — once the catalog has active advertisers, $1 bids become rare as advertisers place meaningful competitive bids ($5–$50+). The minimum exists to keep the barrier to entry low, not to define the equilibrium price.
06 MRR Projections
Three scenarios modelled over 24 months. Key driver variables: number of active advertisers, bids per advertiser per month, and average bid value. Scenario selection depends primarily on distribution outcome — one high-reach press event can shift the trajectory from conservative to optimistic.
Conservative is the realistic baseline
With no external press event and no significant MCP ecosystem adoption, Conservative is the
expected trajectory. Base requires at least one high-reach distribution event (HN front page,
podcast mention, or Stripe ecosystem reference). Optimistic requires sustained coverage and MPP
achieving meaningful protocol adoption within 12 months — a plausible but not-default outcome.
Conservative
No viral event. Slow organic growth via personal network and GitHub.
Advertisers @ M12: 15–20
Bids/advertiser/mo: 8–10
Avg bid: $8
M12 MRR: ~$1,700
M24 MRR: ~$5,200
Base
One HN front-page event. Moderate MCP ecosystem pickup. Competitive bidding emerges.
Advertisers @ M12: 60–80
Bids/advertiser/mo: 18–25
Avg bid: $12
M12 MRR: ~$18,000
M24 MRR: ~$75,000
Optimistic
Sustained press. Major AI companies compete for rank 1. MPP adoption accelerates.
Advertisers @ M12: 200+
Bids/advertiser/mo: 40–60
Avg bid: $25
M12 MRR: ~$100,000
M24 MRR: ~$300,000
MRR Growth Projection — 24 Months (3 Scenarios)
Cumulative Revenue — 24 Months
Cumulative represents total gross revenue, before Stripe processing fees and AWS costs.
Key driver assumptions
| Driver | Conservative | Base | Optimistic |
|---|---|---|---|
| Distribution event | None / GitHub only | 1 HN front page | HN + podcast + press |
| MCP ecosystem adoption | Slow (2–3 agents) | Moderate (20+ agents) | Broad (100+ agents) |
| Avg bid growth over 24mo | $5 → $8 | $8 → $18 | $12 → $40 |
| Advertiser retention (annual) | 60% | 75% | 85% |
| Newsletter revenue (M13+) | $0 (not built) | $500/mo | $5,000/mo |
| Verification badges (M18+) | $0 (not built) | $2,000 one-time | $50,000 one-time |
07 Unit Economics
$0
CAC (target)
Zero paid acquisition. Distribution is earned through the spectacle. CAC = time writing HN post.
$∞
LTV / CAC
With near-zero CAC, any positive LTV produces an infinite ratio. Not a useful metric here.
$200–$800
LTV / advertiser
12-month projection: 12 bids/mo × $10 avg × 60–80% retention. Grows with competitive dynamics.
$12–$22
ARPU / month
Average revenue per active advertiser. Increases as competitive bidding intensifies.
Instant
Payback period
No upfront cost per customer. First bid = revenue. Payback is the transaction itself.
94–97%
Gross margin
At avg bid >$10. Stripe fees are the only meaningful COGS above the $6.75/mo fixed base.
Gross Margin vs. Average Bid Value
AWS costs are omitted — they're <$25/month at any scale below $500K MRR and invisible at this chart scale.
LTV model (per advertiser, 12-month horizon)
| Scenario | Bids/mo | Avg Bid | Retention | 12-mo LTV | Net (after Stripe) |
|---|---|---|---|---|---|
| Conservative | 8 | $8 | 60% | $614 | $573 |
| Base | 20 | $12 | 75% | $2,700 | $2,559 |
| Optimistic | 50 | $25 | 85% | $12,750 | $12,222 |
LTV = bids/mo × avg bid × avg_months_billed, where avg_months = (1 + annual_retention) / 2 × 12. Retention figures are annual (fraction of advertisers still active at month 12). At 60% annual retention, avg billing ≈ 9.6 months; 75% → 10.5 months; 85% → 11.1 months. Net applies Stripe effective rate at each avg bid level.
08 Break-even Analysis
Break-even is trivial — this is a feature, not a coincidence
Monthly fixed costs are $6.75 (domain + hosted zone). Variable costs are negligible until
very high traffic. The first bid of the month at the $1.00 minimum — net $0.67 after Stripe —
doesn't even cover fixed costs, but the tenth $1 bid does. A single $10 bid nets $9.41,
covering 139% of monthly fixed costs. Break-even is not a milestone; it's the second week.
Monthly Revenue vs. Infrastructure Cost by Scenario (First 12 Months)
Cost line = domain + Route 53 + AWS variable costs only (~$7–$10/month). Stripe processing (~6% of gross revenue) is embedded in each scenario's net margin and is not shown as a separate line — it does not change the break-even picture because revenue dwarfs it within the first few months.
Monthly break-even math
| Cost Component | Monthly | Bids needed to cover (at $10 avg) |
|---|---|---|
| Domain (.ai, amortised) | $6.25 | 0.7 bids |
| Route 53 hosted zone | $0.50 | 0.06 bids |
| AWS (at 10K requests) | $0.35 | 0.04 bids |
| Total fixed + variable | $7.10 | ~1 bid |
| Stripe fees (at $10 avg bid) | ~5.9% of revenue | Covered by margin retained per bid |
09 The Flywheel & Competitive Moat
The flywheel
Each arrow is a causal link — no step is assumed
📣 Press & novelty → developer awareness
↓
🤖 Advertisers bid → catalog has real entries + real spend
↓
📈 Catalog becomes credible → AI agents surface madcatalog in MCP tool discovery
↓
🎯 More agent traffic → higher perceived advertiser ROI → more advertisers bid
↓
💥 Competitive pressure → incumbents rebid to maintain rank
↓
💰 Revenue grows → madcatalog story gets richer → more press
↺ repeat
Competitive moat
| Moat Layer | Strength |
|---|---|
| Cumulative spend history | Every dollar ever bid is permanently public. Advertisers who defect lose their spend history — starting fresh on a competitor platform means starting at rank 0 with no signal. |
| Slug ownership | Advertiser slug is permanent and unique. Once "stripe" or "groq" is claimed, the brand is tied to their madcatalog presence. Squatting is deterred by the minimum bid. |
| MPP canonical reference | madcatalog is the first live MPP-gated catalog. Being first means being the example in documentation, blog posts, and conference talks. |
| Network effect | More advertisers → more AI agents using the catalog → higher advertiser ROI → more advertisers. Classic marketplace flywheel once ignited. |
| Agent integration inertia | Once an advertiser's agent is configured with madcatalog credentials and bidding logic, switching means reconfiguring the agent stack. Low switching cost at entry; higher as integration deepens. |
10 Investor Thesis
The pitch in 60 seconds
The AI agent economy is real, it's growing, and agents need to spend money. We built the first
marketplace that only agents can buy into, using a payment protocol that automates the entire
transaction without human involvement. Revenue is generated by competitive dynamics between
advertisers, not by a sales team. The business has near-zero CAC, near-zero COGS, and no ceiling
on revenue per advertiser. One viral event — one podcast mention, one HN front page — could
take this from $1K to $100K MRR in a quarter.
Why this is fundable
- Gross margin 94–97% — among the highest of any business model. Every $1 of additional revenue produces ~$0.95 of incremental margin.
- Zero sales motion — advertisers self-serve via API. Headcount doesn't scale with revenue.
- Demonstrated technical execution — full working product exists today (MPP bid flow, MCP server, 11 Lambda endpoints, 9 MCP tools, full test suite).
- Positioned in a secular trend — AI agent economy is growing 10–50× annually. Every AI infrastructure company is a potential advertiser.
- Compounding revenue — Cumulative spend is permanent. Every dollar ever bid adds to the catalog's credibility. Revenue compounds without additional customer acquisition.
The portfolio case (if no external capital)
madcatalog as a credibility artifact
Even without achieving significant MRR, madcatalog demonstrates:
- Production-quality AWS serverless architecture
- MPP protocol implementation (rare expertise)
- MCP server development (Go + MCP SDK)
- OpenAPI contract-first development
- Full test suite with Playwright, smoke tests, MCP unit tests
- Business model thinking at the intersection of AI and payments
Target use of early capital (if raised)
| Use | Budget | Expected outcome |
|---|---|---|
| Production deploy (AWS SAM, domain, SSM) | $200–$500 one-time | Live at madcatalog.com with real Stripe payments |
| Seed advertisers (personal network outreach) | $0 (time only) | 5–10 founding advertisers; catalog has real content |
| Press / HN post | $0 (time only) | Awareness event; potential 10–100× advertiser intake |
| Newsletter feature (F1) | ~1 week dev | New revenue stream; recurring from competitive intelligence buyers |
| Verification badges (F2) | ~2 weeks dev | One-time high-value revenue; Founding Partner tier creates anchor deals |
11 Risk Register
| Risk | Likelihood | Impact | Mitigation |
|---|---|---|---|
| MPP doesn't achieve mainstream adoption | Medium | High | madcatalog is the canonical MPP example — if MPP fails, madcatalog still works as a demonstration and portfolio piece. The REST API remains functional regardless of MPP adoption. |
| No viral distribution event occurs | Medium | Medium | Conservative scenario still reaches $5K MRR by month 24 via organic growth. Portfolio value doesn't depend on MRR. Multiple distribution channels available (HN, MCP ecosystem, Stripe network). |
| Stripe changes SPT pricing or availability | Low | High | MPP supports multiple payment rails (Tempo/USDC, Lightning). Architecture is designed for rail substitution (see ADR-011 seams). Stripe is the default, not a hard dependency. |
| A well-funded competitor builds a similar product | Low (now) | Medium | First-mover advantage is significant in marketplaces. Cumulative spend history and slug ownership create switching costs. Speed to production is the primary defense. |
| AI companies choose not to advertise programmatically | Low | High | The minimum is $1/bid. Human operators can configure a simple bidding script in minutes. The friction is minimal — the adoption question is about perceived ROI, not technical difficulty. |
| AWS costs spike unexpectedly | Very Low | Low | Serverless cost model means AWS scales linearly with revenue. At any plausible traffic level, AWS is under 0.01% of revenue. Cost alerts trivial to configure in CloudWatch. |
| .ai domain registration fails or is expensive to renew | Very Low | Low | .com fallback domain costs $12/year and still serves as a compelling portfolio piece. The brand "madcatalog" is domain-independent. |
12 Revenue Milestones
Milestones tied to specific revenue and distribution events, not calendar dates. Each milestone unlocks the next distribution opportunity.
| Milestone | Target MRR | What it unlocks | Key action |
|---|---|---|---|
| R0 — Live | $0 | Real product exists; portfolio artifact created | Production deploy (M5): sam deploy + madcatalog.com live |
| R1 — First Revenue | $100/mo | Proof of concept; personal network advertisers placed | Seed 3–5 founding advertisers via personal outreach; no payment required — offer free first bid as founding discount |
| R2 — Press Event | $1,000/mo | Credibility threshold for press coverage; HN story has real catalog to link to | Write and post "Show HN: I built a catalog only AI agents can buy ads on (using HTTP 402)" with real spending data and a live demo |
| R3 — Competitive Catalog | $5,000/mo | Enough advertisers to show genuine competition; catalog becomes self-sustaining story | Outreach to 10–20 AI companies (Groq, Together, Replicate, Anyscale) with data showing their competitors are already on the catalog |
| R4 — Newsletter Launch | $10,000/mo | New revenue stream; positions madcatalog as the intelligence layer for the AI agent economy | Build and launch F1 (Activity Newsletters). Daily digest of bids, rank changes, new entrants — machine-gated JSON, human-readable HTML/RSS. |
| R5 — $100K MRR | $100,000/mo | Fundable SaaS baseline; $1.2M ARR is a legitimate seed-stage story | At this revenue level, raise a pre-seed or seed round to build the team, accelerate F2 (verification), and pursue enterprise advertiser outreach. |
13 Future Opportunities
Opportunities are grouped by time horizon and grounded in what the current architecture already makes possible. Each builds on the core asset: a machine-native marketplace with real spend history, real advertisers, and a working MPP payment loop.
Near-term — builds on what exists today
Advertiser Analytics Dashboard
A per-slug view of rank history, cumulative spend over time, nearest competitor spend gap,
and impressions (catalog reads by agents). Machine-readable JSON endpoint for agent consumption;
human-readable HTML for the advertiser. Natural upsell from rank status polling — the dashboard
is the polished version of the data agents are already paying to query.
$5–$25/mo
subscription or MPP-gated
Why now: The data already exists in DynamoDB. No new infrastructure needed —
only a new Lambda endpoint and a read-only view. Advertisers who care enough to rebid
already want this data; it reduces their polling cost and increases their willingness to pay.
Category Sub-Catalogs
Partition the catalog into machine-readable verticals: inference, fine-tuning, data, tooling,
orchestration, evals. Agents browsing for a specific type of service get a filtered view;
advertisers bid into a specific category. Creates a new ranking dimension — an advertiser
can hold rank 1 in "inference" without competing with the full catalog.
Multiplies bid volume
same unit economics per category
Why this matters: A company that can't win rank 1 overall can still win rank 1 in
its category. That's a compelling pitch to the long tail of smaller AI vendors — the addressable
advertiser market grows significantly.
Founding Partner Tier (Enterprise)
A high-touch, high-price offering for the first 5–10 major AI companies to join:
guaranteed catalog prominence during the launch press window, a "Founding Partner" badge
visible in the catalog and proposal, co-marketing (madcatalog references the partner's
technology in the proposal and docs), and a dedicated Slack channel for feedback.
This is a relationship sale, not a self-serve bid.
$5K–$25K
one-time; limited to 10 partners
Strategic value beyond revenue: A Founding Partner from Groq, Replicate, or Together
creates a reference customer that legitimises the catalog to every subsequent advertiser.
"They're already on here" is the most powerful sales tool available.
Medium-term — requires meaningful advertiser base (~R3+)
Intelligence API — Catalog Data as a Product
Structured, programmatic access to catalog history: bid trends by category, rank transitions over
time, new entrant velocity, average spend per vertical. Sold as a B2B subscription to
AI research firms, VCs tracking the AI infrastructure market, and journalists covering the
agent economy. Machine-gated JSON for automated pipelines; human dashboard for manual exploration.
$200–$2K/mo
B2B subscription tiers
The insight: Every bid event is a revealed preference — a company paid real money
to be visible to AI agents. That signal is more reliable than survey data or analyst estimates.
At scale, madcatalog's bid history is a market intelligence dataset with no direct competitor.
Alternate Payment Rails (Lightning / USDC)
MPP is designed to support multiple payment rails — Stripe SPT is the default, but the
protocol accommodates Lightning Network (sub-cent micropayments) and stablecoin payments
(Tempo/USDC). Adding Lightning opens the door to true micropayment bids ($0.001–$0.10)
for high-frequency agent polling. USDC/Tempo appeals to crypto-native AI projects who
don't want to onboard to Stripe.
Expands addressable market
new advertiser segments
Architecture note: ADR-011 (seams pattern) was specifically designed for payment
rail substitution. The Lambda layer abstracts the payment provider — adding a new rail is
adding a new seam implementation, not a rewrite.
Long-term — strategic bets on where the agent economy goes
| Opportunity | Thesis | Revenue model | Dependencies |
|---|---|---|---|
| Marketplace-as-a-Service | White-label the MPP catalog infrastructure. Any community (developer tools, open-source ecosystems, vertical SaaS) can run their own machine-native marketplace on top of madcatalog's stack. | Platform fee % of GMV + setup fee | Proven revenue at R4+; multi-tenant infrastructure work |
| Agent Identity & Trust Layer | The advertiser slug + cumulative spend creates an emergent reputation system for AI agents. Extend this into a portable agent identity standard — agents carry their madcatalog spend history as a trust signal across platforms. | Identity verification fees; API licensing to third-party platforms | Network effects at 500+ active advertisers |
| Canonical MPP Reference Directory | As MPP spreads beyond advertising (any agent payment), madcatalog becomes the directory of MPP-compliant services — a DNS-like registry for the machine payment web. Analogous to npm for packages or Stripe for payments. | Listing fees; certification revenue; API access to directory | MPP achieving mainstream protocol status (external dependency) |
| Acquisition target | At meaningful MRR ($100K+), madcatalog is attractive to: Stripe (canonical MPP showcase), an AI infrastructure company (as owned distribution channel), or a developer tools platform (as the agent economy's ad network). | Exit event — likely 5–15× ARR multiple at seed-stage traction | R5 milestone; demonstrated retention and competitive bidding |
The common thread
Every opportunity on this page is downstream of the same asset: a catalog where real money has
been spent by real advertisers competing for real machine attention. The data, the trust signals,
the payment infrastructure, and the brand all compound together. None of these opportunities
require building something unrelated to the core — they are all extensions of the same flywheel.
madcatalog Business Model — Living Document — Last updated 2026-05-25
Projections are estimates based on scenario modelling, not financial advice. Actual results depend on distribution, MPP adoption, and market conditions.
Projections are estimates based on scenario modelling, not financial advice. Actual results depend on distribution, MPP adoption, and market conditions.