ZRR0 IX LLC — Master Business Model

North-star document. This is the whole company on one page-set: what ZRR0 IX is (and is NOT), how the cross-market niche-portfolio + builds + brand + racing flywheel work as one integrated model, the value chain, the moat, and how every piece reinforces the others. All numbers are marked [RESEARCHED] (grounded in a current 2026 source) or [ESTIMATE] (a planning figure to validate with a real quote). Read this first; the domain deep-dives (PRICING_ENGINE_DESIGN.md, JAPAN_SIDE_OPS.md, ORG_AND_FINANCIAL_MODEL.md, LOGISTICS_AND_FACILITIES.md, the stress-test files) are the detail behind each section.

Status: PRE-OPS. Date: 2026-06-06. Capital line: ~$350k (trading-funded, see §9).

0. The one-sentence model

ZRR0 IX is a branded performance-car portfolio company that sources cars across four decorrelated markets (JDM / USDM / EURO / Korean), measures every car with a data engine the way the Lighthouse measures trades, turns a slice of them into branded builds, and wraps the whole thing in a racing-validated brand — so the combination, not any single car, makes the volume and smooths the margin.

It is NOT a flip-a-car-for-a-quick-buck importer, and it is NOT a generic dealer lot. It is a niche-portfolio + builds + media company that happens to move cars, run with the same discipline (hard margin floor, decorrelation, measure-once / query-many, a "villain-map" veto) the operator already runs on the trading book.

1. What ZRR0 IX IS / IS NOT

2. The integrated model — five engines, one flywheel

                 ┌─────────────────────────────────────────────────────┐
                 │     LIGHTHOUSE (trading book) — the FUEL TANK        │
                 │     funds float top-ups + racing + patience          │
                 └───────────────────────┬─────────────────────────────┘
                                         │ self-sponsor / no forced sales
        ┌────────────────────────────────┼────────────────────────────────┐
        ▼                                 ▼                                ▼
┌───────────────┐   max-bid /     ┌───────────────┐   build threads ┌───────────────┐
│  SOURCING     │── arb spread ──▶│  DATA ENGINE  │── price/uplift ─▶│  BUILDS       │
│  4 sleeves    │◀── what to buy ─│ measure-once  │◀── feedback ─────│ ZRR0/HER0/MUTT│
│ JDM/USDM/EU/KR│                 │  query-many   │   sold→re-fit    │  brand premium│
└──────┬────────┘                 └──────┬────────┘                 └──────┬────────┘
       │ inventory                       │ comps + villain-map veto        │ content
       ▼                                 ▼                                 ▼
┌──────────────────────────────────────────────────────────────────────────────────┐
│            BRAND + MARKETING ENGINE  (owned site + GEO/AEO + social + merch)       │
│   sells credibility & trust → premium prices → buyers find ZRR0 first              │
└───────────────────────────────────────────┬──────────────────────────────────────┘
                                             ▲
                                  free reach │ credibility
                                ┌────────────┴───────────┐
                                │   RACING FLYWHEEL       │
                                │  pro-drift, self-spons. │
                                │  rolling billboard      │
                                └─────────────────────────┘

How the pieces reinforce each other (the actual loop):

1. Sourcing finds cars across four decorrelated markets; the Data Engine decides what to buy and at what max-bid (arb spread × turn-velocity).
2. A slice of inventory becomes Builds (HER0/MUTT/ZRR0); builds + racing become content; content + the Brand make buyers find ZRR0 first and pay a premium.
3. Racing generates credibility + footage at near-zero placement risk (self-sponsored), which feeds the Brand and deepens build/culture literacy that improves the product.
4. Every closed sale writes back to the Data Engine (predicted-vs-actual margin, days-to-sell, mod uplift) — the Atlas update step. The engine gets sharper than any gut-number competitor: that is the durable moat.
5. The Lighthouse funds float top-ups + racing + patience, so the business never has to dump a car at a loss — it can wait for the right bid. (This is also the single biggest dependency — see §9 risk.)

3. The cross-market niche portfolio (the "chimera")

Each market is an independent, decorrelated edge sleeve; each car is a fungible "trade" against a hard landed-margin floor (~18–22%) treated exactly like the trading book's R:R floor. Volume and margin-smoothing come from the combination, not any one sleeve. The trading-system mapping:

The four sleeves:

First JDM run = a decorrelated basket, not one bet [RESEARCHED sourcing brief]: 2× liquid flippers (Supra/RX-7/Evo/STI grade 4+) · 1× just-unlocked 25-yr premium (R34 '01 verified by build MONTH) · 1× drift/brand car (S15/JZX/ECR33 — doubles as content + flywheel) · 1× HER0/MUTT base. Mixed liquidity = smooth cash + brand fuel. Keep run 1 small (3–4 cars) and prove one full auction-to-sale cycle before scaling.

4. The brand architecture — branded house

ZRR0 is the master/credibility brand (the operator's racing-validated signature — the Hennessey / Liberty-Walk halo). Three build lines sit under it as endorsed sub-brands, plus the import/inventory arm:

Naming convention "HER0 by ZRR0" / "MUTT by ZRR0" keeps one credibility umbrella. The future Japan/US mod-shop partner touches IX inventory + custom builds, NOT the ZRR0-branded signature cars (protect signature exclusivity — the Liberty-Walk magic). Lock the naming + one ownable visual signature BEFORE the website/logo, and run a USPTO + common-law trademark clearance on ZRR0 / HER0 / MUTT before any merch spend (MUTT especially may collide). [RESEARCHED brand brief]

5. The value chain (JDM inbound, primary flow)

End-to-end ~6–10 weeks per car [RESEARCHED logistics brief]:

Auction agent screens USS/TAA/JU + translates auction sheet
  → Data Engine sets MAX BID → agent proxy-bids (foreigners can't bid directly)   [0–2 wk]
  → win+pay (JPY, 3–5 business days) → exporter: inland → de-reg → Export Certificate [~1–2 wk]
  → ocean RoRo/container Yokohama/Kobe → Port of Tacoma                            [~2–3 wk]
  → US customs broker: HS-7 + EPA 3520-1, clear under HTS 9903.94.04               [3–7 days]
  → trucking port → WA warehouse (<20 mi)                                          [~$150–400]
  → intake QA → flip as-is (IX) OR route to build bay (HER0/MUTT/ZRR0)
  → photograph → list → sell (owned site for signature; syndication for IX commodity)

Per-car JDM landed-cost stack (~$30k auction example) [components RESEARCHED, car-specific ESTIMATE]:

Unit economics by sleeve [landed RESEARCHED, margins ESTIMATE]:

Treat the portfolio as two decorrelated sleeves like the trading book: JDM flips = high-frequency core; builds = niche high-R / low-freq. Sub-contract build labor — do NOT own a shop early (labor + cost-overrun is the #1 margin-eater and the inventory that gets stuck = the max-consecutive-loss analog).

Facilities: PRIMARY = one 5,000–10,000 SF flex unit in the Fife/Tacoma/Kent corridor near the Port of Tacoma ($10–16/SF/yr NNN) doing receive / QA / store / build / photograph [RESEARCHED]. EXPANSION = Minnesota — a demand/arbitrage node fed from WA, NOT an import gateway (no port → seasonal pricing power on spring/summer sports cars; ~$8.73–12/SF rent; separate $50k MN dealer bond + Jan-1-2026 fingerprint/background on ALL owners/officers).

6. The moat (three reinforcing layers)

6.1 — The Data Engine ("Lighthouse for cars") — the durable, non-obvious moat. A five-layer pipeline mirroring the Atlas (full design in PRICING_ENGINE_DESIGN.md):

• Ingest: NHTSA vPIC (free VIN decode = join key) · Marketcheck (US active+sold, ~$299/mo Basic) · eBay sold API · classic.com / BaT / Cars&Bids (enthusiast comps — prefer APIs, scrape supplementally) · auctionsheetjp / AuctionDataSearch / JapanStat (JP auction history) · Manheim MMR + Black Book once licensed (the wholesale truth layer — license-gated). [RESEARCHED]
• Measure: one normalized row per car-config — identity, condition (US title OR JP overall/ext/int grade + damage map), structured mods + mod_stack_hash, context (market/region/season/channel/FX), outcome labels that fill over time.
• Model: depreciation/appreciation curve · mod-value uplift attribution · cross-market arb spread = US_resale_comp − (JP_auction_comp + landed_cost) ranked by spread × velocity.
• The utility function (literally the max-bid): max_bid = E[net_resale] − all_landed_costs − target_margin − risk_buffer where risk_buffer scales with comp-variance (IQR), grade uncertainty (R/RA penalty), thin-comp penalty (n<10), and season. Bid only when deal_score ≥ floor. Store the WHOLE distribution of sold comps, never a single number.
• Feedback: every won/lost bid + sale writes predicted-vs-actual; a nightly job re-fits curves, uplift coefficients, and variance — measure-once, query-infinitely.

The villain-map veto (a free edge): encode R/RA grade, sub-anniversary build month, exemption-voiding swap, rust-belt/salvage title, value-destroying mods as hard rejects. This turns the metaphor into a compliance firewall (see §7).

6.2 — The Brand. Racing-validated credibility + an ownable visual signature make a "ZRR0 GT-R" a different object than a stock GT-R (Liberty-Walk / Hennessey playbook). Buyers pay a premium and find you first.

6.3 — Cross-market arbitrage. The documented spreads (JDM inbound on the 25-yr unlock; USDM muscle into Japan at 0% duty / Australia) are real and datable — and the engine prices them net of duty + freight + insurance + broker, not gross.

7. Hard rules & landmines (re-sequenced to the FRONT — stress-test verdicts)

All three adversarial lenses converged on the same plan-breakers. These move to day-one.

7.1 — Duty is a LIVE regime variable, NOT a stored constant ⚠️ [RESEARCHED, June 2026]

The research briefs contradicted each other (2.5% / 15% / 27.5%). Verified current state (June 2026):

• Feb 20 2026: SCOTUS voided IEEPA reciprocal tariffs (Learning Resources v. Trump).
• Feb 24 2026: White House installed a Section 122 ~10–15% surcharge with a ~150-day sunset (~mid-to-late-Aug 2026) — and the CIT has since ruled Section 122 itself unlawful; it is in active appeal.
• Passenger cars are EXCLUDED from Section 122 AND exempt from the 25% Section 232 auto tariff (HTS 9903.94.04). Net: a 25-yr Japan-origin passenger car is back to ~2.5% base duty — but the regime moved THREE times in six months.
• RULE: the broker confirms the live HTS 9903.94.04 / Section-122/301 posture per shipment; the max-bid engine treats duty like an FX-regime variable and rejects any bid whose spread is thinner than the duty-volatility band (~15% + buffer worst case). Watch the ~Aug-2026 sunset specifically. Sources: SCOTUS Feb 2026; providecars.co.jp Section-122 analysis; CIT/Ward & Smith; CBP Section 232 FAQs.

7.2 — The WA "wholesale-first" plan does NOT exist ⚠️ [RESEARCHED]

Washington abolished the wholesale dealer license in 2019 (stopped issuing 2017, renewing 2018, deleted from law July 1 2019). There is ONE motor-vehicle dealer license that, from day one, requires: $30,000 bond (premium ~$196–900/yr), enclosed commercial building, permanently-affixed exterior sign, directory-listed phone, records area, zoning compliance, and staffed hours 10am–4pm ≥5 days/week, plus the 8-hr WSIADA dealer-education course for first-time used dealers. The "lean wholesale, defer the staffed location + hire" path is a phantom. Source: dol.wa.gov; suretybonds.com; ACV.

• DECISION (before any lease or agent engagement): either (a) accept the full WA dealer build-out — staffed location + first hire — as a DAY-ONE cost, or (b) license the dealer entity in a state that retains a true wholesale tier and keep WA warehouse-only, confirming AuctionACCESS reciprocity — with CPA + attorney signoff, not assumption.

7.3 — Engine swaps ONLY on US soil, on cars already 25+ ⚠️ [RESEARCHED]

EPA Form 3520-1 Code E (21+ yr emissions exemption) is voided by a non-original engine unless it's an EPA-certified equal-or-newer engine — which a JDM/aftermarket performance engine never is. The MUTT "ECR33 + supercharged LS7, swap in Japan then ship" thesis is exactly the config that loses the exemption and can make the car non-importable / held at port / forced into the Registered-Importer + DOT-bond (1.5× value, 120-day) conformance path. RULE: import the shell in original drivetrain; do emissions/engine-relevant swaps domestically, post-entry, on 25+ cars only. Japan-side work = cosmetic / chassis / suspension / widebody only.

7.4 — Other confirmed seizure/penalty traps

8. Sequencing — Japan side (agent-first, GK-later) [RESEARCHED Japan brief]

• Runs 1–3: licensed export agent only (proxy-bids USS/TAA, de-reg, Export Cert, books ocean). Fee ~¥70k–290k/car + refundable ~¥100k bid deposit. Do NOT form the GK; do NOT touch the Business-Manager visa.
• Foreigners legally cannot bid directly. USS direct membership requires a kobutsusho license HELD ≥1 FULL YEAR before you may even apply + a Japan-registered business + a real physical premises (virtual office fails the police check) + Japanese-speaking staff.
• When you commit to scale: recruit the Japan-resident executive manager FIRST, then form GK + file kobutsusho the SAME DAY to start the mandatory 1-year USS clock — keep the agent running the full ~12–15 month bridge. (Backup resident-rep = de-risk the single point of failure.)
• GK warehouse + direct mod-shop + USDM-inbound = Phase 3+. Intercompany WA-LLC ↔ GK car transfers create a transfer-pricing obligation — design with CPA + attorney before the first intercompany sale.
• Time-boxed tailwind: USTR Section 301 China-vessel fee suspended through Nov 9 2026 → RoRo temporarily cheaper. Front-load runs inside this window, re-quote before it lapses.

9. The funding spine & financial model [RESEARCHED fees, ESTIMATE allocation]

~$350k allocation (illustrative, re-baseline with WA staffed location as day-one cost):

• Break-even ~2.2 JDM flips/month vs $13,800/mo steady-state fixed cost [ESTIMATE]. A blended 3-flips + 1-build month ($34,900 contribution) covers fixed ~2.5×.
• The binding constraint is float-velocity (frequency × per-car edge), NOT capital size — mirror the trading book. A ~$30k car ties up ~$35–37k for 6–10 weeks; cap simultaneous in-flight units to a fraction of capital.

⚠️ The funding-spine risk (highest execution risk — all lenses agree)

The plan is architected as "not-dependent-on-fast-cash because the Lighthouse funds it" — but the Lighthouse is in its FIRST live/OOS verdict month (June 2026; the operator's own docs say "June live = verdict"). The plan dresses a dependency as independence. If the June verdict disappoints, racing + float top-ups + the "runs-itself" premise lose their backstop, and the operator carries trading drawdown + illiquid inventory + build overrun + racing burn simultaneously on one psychology and one calendar.

GATING RULE: do NOT launch on a "trading should fund it" assumption. Gate launch on a measured milestone: 12 months of car fixed costs segregated in cash from realized/withdrawn live P&L before spending on cars or racing. Name racing as the first bucket cut if trading underperforms.

10. Org & hiring sequence [RESEARCHED roles, ESTIMATE comp]

Operator stays brand + capital + driver; everyone else is vendor/contract until a single car-flow line proves repeatable, then convert the bottleneck role to payroll.

Vendors first (V1–V6, run-1 ready): JP auction agent · exporter · freight forwarder · 25-yr-savvy customs broker · import/customs attorney · international CPA. Keep these outsourced indefinitely.

Then staff (H1–H6):

1. Analytics/data hire — engaged in PRE-OPS (correction to "defer until feedback accumulates"). The engine prices car #1; you cannot make it the centerpiece AND defer the only person who builds it, or car #1's bid reverts to the gut number you're trying to kill. Real data engineer (Python/SQL/dbt + light ML + data-licensing compliance).
2. WA lot/storefront staff — day-one if selling in WA (the 10–4/5-day staffed-hours rule forces it; not phase-2).
3. Marketer = SYSTEM-OWNER, not a do-everything generalist: owns content + brand strategy, editorial/GEO/AEO calendar, brand consistency across 4 marks, community/UGC, GBP + reviews, vendor management. Specialist contractors underneath: a dedicated video editor (highest-leverage auto-brand contractor), automotive technical-SEO/GEO, paid media, one-time brand-identity designer.
4. Logistics coordinator / Japan-resident rep — only at GK formation.
5. Assistant; retail staff scaling; JP coordinator — as volume proves out.

11. The marketing engine (2026 reality, not dead-schema) [RESEARCHED brand brief]

Google KILLED the dedicated vehicle-listing rich result in Sept 2025. "Type X car → my site first" is NO LONGER won by Vehicle schema snippets. It is won by THREE things:

1. Google Business Profile + local Map Pack dominance (highest-ROI for a dealer; drive 5–15 answered reviews/month) — only pays off with a real, reviewable public location.
2. GEO/AEO — getting cited inside AI Overviews / ChatGPT / Perplexity (now front ~half of US searches). Put a 40–60 word direct-answer block atop every model/guide page.
3. Marketplace syndication for the commodity IX inventory (CarGurus/AutoTrader/Cars.com/ Google VLAs/Facebook via a ~$99/mo feed tool). Keep signature ZRR0/MUTT builds OFF AutoTrader — route that demand to the owned site (commoditizing them destroys the premium).

Website = two zones, one domain: transactional IX inventory (Vehicle+Product+AutoDealer schema — table-stakes, not the differentiator) + editorial/brand zone (build threads, arb explainers, model buying guides, the sub-brand stories). Content engine: YouTube long-form (depth, $15–40 RPM for personality-led vs $1.50–3 generic) + IG Reels/TikTok (discovery) + Facebook (community). Merch on Fourthwall, operator's own kit first (signal > revenue). Add a SEMA / Tokyo Auto Salon goal to the 18–24 mo roadmap — the credibility accelerant that minted Liberty Walk and Hennessey.

12. The racing flywheel — a CAPPED marketing line, not a championship [RESEARCHED racing brief]

Drift is structurally unprofitable below the top 5–10% (FD season-champ prize ~$25–50k vs a $150–500k Pro campaign). Everyone else monetizes via sponsorship + content + side business — which is exactly what ZRR0 wants. Self-sponsorship (Lighthouse-funded) removes prize-money dependence, so marketing value accrues from CONSISTENCY + CONTENT + DISTINCTIVENESS, not placement.

• 2026 ladder: grassroots → 1 of 10 affiliated Pro-Am series → PROSPEC license → FD PROSPEC → FD PRO. WA fit: Evergreen Drift (Monroe, WA) is an FD Scouting series in the primary-warehouse state — natural home venue.
• Phase it, cap it: Phase 0 (school + cheap grassroots car, ~$8–15k) → Phase 1 (consistent WA grassroots + distinctive ZRR0 build + content engine) → Phase 2 (a Pro-Am season for the license headline, ~$25–50k) → Phase 3 (FD only if brand ROI clearly justifies it).
• Guardrails: hard annual cap (~$40–60k/yr through Phase 2) as a marketing line; content-yield KPIs (reach/leads per event, NOT finishing position); a replaceable comp chassis, not showpieces, for the rough rounds; run it as a documented business marketing activity (CPA-confirmed deductible) with real deliverables.

Honest risk: the "placement-doesn't-matter" thesis is half-true — the moment a license becomes the goal, the WA on-ramp awards only one podium-gated PROSPEC petition per season, so you DO have to win, reintroducing the time-sink/obsession risk. Decision: stop at "credible grassroots/content driver" unless ROI proves the license win-chase. One injury removes both the driver and the rolling billboard → build HER0/MUTT/IX as founder-independent content lines.

13. Critical-path sequencing (do these BEFORE committing capital)

1. Gate launch on a measured, withdrawn-cash trading milestone (12 mo of car fixed costs in cash). (§9)
2. Decide the dealer-license state NOW — full WA build-out vs wholesale-tier state + WA warehouse-only — before any lease/agent. CPA + attorney signoff. (§7.2)
3. Retain a 25-yr-savvy customs broker + import attorney + international CPA in PRE-OPS — to confirm zoning, transfer-pricing, B&O/MN nexus, and the first basket's build-month/EPA/DOT-bond posture. (§7)
4. Stand up the valuation engine v0 in PRE-OPS (analytics contractor's first job) so car #1's max-bid is data-set, duty modeled as a volatile band (~15% + buffer), never a constant. (§6, §7.1)
5. Lock brand naming + visual signature + trademark clearance before website/logo/merch. (§4)
6. Run 1 = small (3–4 cars), decorrelated, vendor-only Japan side, every spread worst-case-duty-survivable; prove ONE full auction-to-sale cycle before scaling. (§3)
7. Build the USDM-export compliance spine (EIN→AES→72-hr title workflow) BEFORE the first outbound car. (§7.4)
8. Japan: agent for runs 1–3; recruit resident manager before GK; GK + kobutsusho same day at scale. (§8)
9. Cap + KPI racing; keep it the first cut. (§12)

14. Verdict (synthesized across all three stress-tests)

Concept SOUND; the cross-market arbitrage thesis and chimera framing are real and the data moat is genuinely differentiating. But as originally sequenced the plan is mis-sequenced and built on unstable load-bearing numbers. The three converging fixes that protect more capital than anything else:

1. Stop treating duty as a stored constant — verify per shipment (it moved 2.5%→15%→~2.5% in six months; currently ~2.5% for 25-yr passenger cars but in active litigation).
2. Kill the WA wholesale-first fiction — it doesn't exist; decide the license state and staffed-location cost up front.
3. Gate the launch on measured trading cash, not "it should fund it" — the funding spine is in its first verdict month; don't front-load maximum founder concurrency (sourcing + builds + racing + 4-brand content + live trading) when the fuel tank is least certain.

The end-state the operator wants — over-qualified, delegated, brand-focused, engine runs itself — is the right target. It must be sequenced into existence (engine in pre-ops, licenses decided, run 1 tiny, racing capped, sub-brands founder-independent), not assumed alongside the launch.

Source anchors (current-2026, verified or research-grounded)

• Duty/tariff: SCOTUS Learning Resources v. Trump (Feb 20 2026); providecars.co.jp Section-122/232/301 JDM analysis; CIT Section-122 ruling (Ward & Smith / Gibson Dunn); CBP Section 232 auto FAQs; HTS 9903.94.04.
• WA license: dol.wa.gov vehicle-dealers; suretybonds.com (wholesale eliminated 2019); ACV Auctions WA guide.
• Import compliance: nhtsa.gov importing-vehicle FAQs; epa.gov 3520-1; 19 CFR 192.2(b) (AES/72-hr); projectjdm.org 2026 guide.
• Japan side: ussnet.co.jp; providecars.co.jp USS; small-business-japan kobutsusho; mailmate/e-housing/venturejapan GK; housingjapan.com (¥30M visa); ustr.gov / wcshipping.com (Section 301 vessel-fee suspension).
• Data engine: developer.manheim.com / site.manheim.com MMR; marketcheck.com/apis/pricing; vpic.nhtsa.dot.gov; classic.com; auctionsheetjp.com; auctiondatasearch.jp; japanstat.com.
• Brand/marketing: developers.google.com vehicle-listing (deprecated Sept 2025); dealersunited.com SEO-vs-GEO 2026; hennesseyperformance.com; stanceauto.co.uk (Liberty Walk); fourthwall.com.
• Racing: formulad.com/pro-am + news.formulad.com 2026 Pro-Am reorg; driftschoolevergreen.com; worldofdrift.com; slrspeed.com.
• Logistics/facilities: wcshipping.com (RoRo/container, costs, ports); portoftacoma.com; dps.mn.gov dealers ($50k bond + 2026 background checks).
• Org/financial: dol.wa.gov form 430038; suretybonds.com WA bond; autotraderimports.com agent fees; classic.com R34 comps.