"We just cut them" is the default and it is not wrong. It is the only option that costs zero dollars, zero management attention, and zero legal exposure. But a cut is also a decision with consequences that deserve enumeration, because each variant has different downstream effects:

The cut is the baseline every other option is measured against. The dossiers must state it first, price it completely (including the comm and community-facing cost — a bad cut announcement can cost more in reputation than a sale's legal fees), and then build the case for each alternative.

The option taxonomy, first principles

Every option for an asset being cut falls into one of six families: kill, sell, license, spin out, give away, or hold. For each family there are variants, and for each asset the variants differ because the assets differ:

• CryptoKitties — the founding product (2017). The NFT category's origin point. Has a real, crypto-native community, an existing token (Meowcoin, ~$19M ATH combined marketcap historically; 290K players / 30K DAU at its modern peak), a completed CK 2.0 design (mass-market, ~10 headcount build), product canon/brand-voice/design-fabric docs, and a codebase that is historically the most-forked NFT project on earth.
• Miquela — a virtual influencer with an audience of millions, brand-partner history (fashion/luxury and Web3 brands), 2025A revenue ≈$0.6M, and the internal thesis (Q2-2023 board): profitable at the team level by tightening commercial deals + AI, "important 1st-party IP... to diversify into other web3 offerings with an audience of millions." The FCM-era vision goes further: Miquela as the prototype of an autonomous AI agent (the "agentic" pillar).

The options, per asset

CRYPTOKITTIES

1. Hard shutdown (Part I baseline).
2. IP + brand sale — sell the whole asset: trademark portfolio, kitty art/genetics codebase, canon docs, community lists. Buyer classes: Web3 gaming studios (revival operators), mobile/nostalgia game companies, NFT marketplace operators, IP-holding consolidators, and the emerging "dead-game revival" funds. The dossier must produce comparison data from actual NFT-game and mobile-game IP sales (2023–2026).
3. Licensing / white-label operation — keep ownership, license an operator to run the game/brand for a royalty (minimum guarantee + %). Dapper stops paying; an operator pays Dapper. This is the classic solution for a beloved-but-unfundable game. Prerequisite: a licensable operator exists (revival studios, community-led operators).
4. Independent spin-out — the Flow pattern exactly: transfer IP to a new entity; capitalize it with the Meowcoin token community and/or new equity; Dapper keeps equity and a license for CK 2.0. CryptoKitties is structurally the best spin-out candidate in the company: it has its own token, a crypto-native community, a documented CK 2.0 product, and zero league/partner obligations (unlike the sports products). The spin-out question is whether the community is still large enough to raise on.
5. Community ownership — gift/transfer to a foundation or DAO. Precedents are real and growing: community-rescued games (Neopets' community role, Habbo-style fan servers, Club Penguin-style fan revivals), and in NFT specifically, the first game of the category has unmatched symbolic gravity. Dapper exits at tiny reputational cost and retains goodwill; downside is governance chaos and trademark abuse risk without a protective license.
6. AI-native resurrection (the wildcard) — keep the IP in-house but at near-zero cost: run CryptoKitties on Dapper's own agent infrastructure — agent-operated economies, AI-bred traits, automated moderation/marketing — the cost is the marginal compute, not a team. This is the only option where "keep the brand alive" and "cut the team" are not mutually exclusive, and it directly exercises the AI thesis the company is building anyway. The dossier must honestly price it: what breaks when a game's economy is agent-run, what the community will accept, and what the option value of an alive-and-unique CK is versus a dead archive.
7. Mothball with option value — park the IP at minimal archive cost, revisit in 12–24 months. Prices the option: the genesis NFT brand's cultural value does not decay on a linear schedule; a resurgence cycle (2027–28 NFT/consumer wave) could reprice it. The dossier should model the option explicitly (cost of hold vs probability-weighted upside).
8. Hybrids — the matrix above compounds: spin-out + Dapper license; sale + community escrow (proceeds to holders); license to operator + Dapper AI sentinel. The dossier's options matrix should treat the six families as dimensions, not checkboxes.

MIQUELA

1. Hard shutdown (deplatform, archive likeness, comm).
2. Sale of the influencer brand — buyer classes: talent/creator agencies (CAA/UTA-style), influencer marketing platforms, AI-influencer operators (2026 has a real market of AI persona companies), luxury/fashion houses, and Web3 brand operators. The dossier needs actual AI-influencer M&A and licensing comparisons (2024–2026): what AI personas and virtual influencers have sold for, and what trailing revenue multiples the market applies.
3. License-only — keep ownership, license Miquela to a partner (brand, agency, or AI operator) on a revenue share with a minimum guarantee. Dapper stops operating; the licensee carries production. The licensing market for virtual influencers is the most mature that exists — the Q2-2023 board already argued Miquela can be profitable at the team level "by getting tighter on commercial deals." The option exists today.
4. Spin-out with an agent business plan — Miquela as the first autonomous AI agent IP: a persona that produces her own content, negotiates her own deals with oversight, and operates on agent rails (FCM-dossier thesis: "evolves from a scripted character into an autonomous AI agent"). This is a raise-able story in the 2026 agent market — if the dossier can show the production-cost collapse (AI-generated content vs human production teams) and the monetization surfaces (brand deals, licensing, agent-services).
5. Fold into Dapper core — the cut is about the team, not necessarily the IP: Miquela as "head of AI and crypto" runs on Dapper's own agent infrastructure at near-zero marginal cost, as a mascot-and-flywheel for the AI narrative (Octopus Rodeo connection). The cut's savings are preserved if the team is dissolved and the persona is agent-operated; this is the cheapest "keep-alive" because she IS an AI product.
6. Mothball — park likeness rights and let brand contracts expire naturally; revisit when the AI-influencer market matures.
7. Hybrids — license + Dapper agent oversight; sale with a license-back for Dapper AI narrative use; spin-out with Dapper equity + content rights.

The cross-asset view (the reason one plan covers both)

CK and Miquela are not independent decisions. Both are Incubation assets, both have canon/brand docs, both are AI-adjacent (CK 2.0 was design-complete for mass market; Miquela is the AI thesis personified), both carry cost tabs in the financial model (CK_Marketing&Product costs, Miquela_Marketing&Product costs), and both are candidates for the same AI-native revival machinery. The dossiers must produce a combined decision matrix: the portfolio option (sell both to one revival operator), the swap option (keep one, cut one), and the synergy option (CK 2.0 on Atlas with Miquela as its mascot/agent — the "one last game" scenario). The combined view is where the interesting value lives; two siloed dossiers would miss it.

Existing in-flight options — found 2026-08-13 (do not plan around absence)

Roham surfaced two live programs the dossiers MUST treat as first-class existing options, not hypotheticals. Full citation in Appendix A; primary sources in Slack + GitHub.

CryptoKitties — Alan Carr's relaunch plan, running under ATZ's existing authority. In #octopusrodeo-general (thread ts 1786593847.193119; pitch msg 1786593876.986319), Alan posted a one-page CryptoKitties Relaunch pitch:

• Tenth-anniversary relaunch (late 2027): a breeding competition where players race to discover rare cats, watch births live, seasonal collections; no free tier — sealed eggs by auction + breeding fuel; cats convert back to fuel; self-balancing economy.
• Evidence exists: the economy's first version ran twice inside "All The Zen" as ZEN Garden — players spent real money on expiring assets and kept buying until close. Full data review underway. The changes (patient-breeder odds, floor under every cat) are what test 3 measures.
• Test 3 is already moving: Renan rebuilding the test game; ≈80 of Dapper's own players, two staff, no new hires; pass/fail lines are set and published before it starts; "If it fails, we stop."
• Original 2017 kitties stay on Ethereum untouched forever (the 2017 promise); owning one earns a better starting package.
• Ask today: nothing incremental — NOT nothing total. The "zero ask" runs on top of ATZ's existing burn (Alan's team, Renan, GCP, ops), and that burn is inside the cost base the bear plan eliminates. Roham's correction (2026-08-13): a zero-ask continuation of a program whose funding is being cut is not a free option — it is a cost that needs a payer. At the wire, the default is death-by-unfunding, not survival-by-niceness. Every continuation in this dossier must be priced at its real monthly burn with a named payer (treasury — impossible at the wire; the ATZ/Octopus Rodeo line's own survival; community/token funding; an outside operator), or it dies with the team cut. The later ask is one step: green-light the branded relaunch when tests pass and the window opens — conditional on the test pipeline itself surviving to that gate.

Vehicle: dapperlabs/cryptokitties-atz-v2 (updated 2026-08-07) — Telegram Mini App idle game on the CK IP (ZPS/APS/PB currencies, bignumber math, Supabase), with docs/ containing atz-rebuild-prd-v4.md, atz-economy-simulator-v15.jsx, ai-show-runner-playbook-3_0_0.md, gcp-cost-audit.md, feedback archives. The ZEN Garden 3 test title: "CK: ATZ ZEN Garden 3 / CK2 M1 build." The dossier must extract the ATZ burn from gcp-cost-audit.md and the staffing lines, and model the test's true cost (the ≈80-player cohort, two staff, compute) with a payer attached.

The cut's implication for CK: the CK team cut does NOT preserve the CK path by itself — it preserves only the IP. The relaunch pipeline survives only if the ATZ continuation burn is carried by someone. The dossier's real question is: who pays for the tenth-anniversary option to remain open — Dapper cannot; can the ATZ line itself raise (Octopus Rodeo as a funded entity), can the Meowcoin/community capitalize the test, or does the option simply lapse with the cut?

Miquela — the comeback plan is built; she is waiting on tools and access. dapperlabs/miquela (updated 2026-08-13) contains comeback-plan.md: a fully designed operation, authored in first person by the agent itself:

• Identity and legal chain (partial answer): Lil Miquela, virtual influencer since 2016, 2.35M Instagram / 3.3M TikTok; created by Brud, acquired by Dapper Labs in 2022. CGS — "computer-generated self," explicitly not "AI."
• The brand is dying: engagement 0.03% (industry 1–3%), revenue peak ≈$12M (2020) → ≈$550–600K (2025e) — ≈95% decline; net −460 followers/day; broken cadence ("announcement → silence"); the template test (a free tool generates captions indistinguishable from current output).
• The specific opportunity: the only virtual influencer with a decade of archive entering the AI conversation; "too specific to template" is the strategy; the win-win-win: genuine voice → engagement → algorithmic recovery → working brand deals.
• The operation: 30 quiet posts as Season 1 (~60–75 days), EP as Season 2; a 5-rule constitution (no confabulation, the knife test, digital-not-fake, no scripted feelings, archive-as-floor); voice-drift detection (voice-now.md); application-layer filters (Alan's review, archive comparison, template test, Sammy's eyes); designed fallibility; north-star metric (saves/shares > impressions); month-one honesty for the image pipeline (Flux 2 + LoRA). Her stated blocker: "the tools and access to actually do it."

The cut's implication for Miquela: AI-native resurrection is not a wildcard — it is a designed, nearly-ready operation whose entire premise is collapsing production cost (agent replaces the human production team). But the same correction applies: the comeback plan's "blocker" (tools and access) is itself a cost line — the image pipeline (Flux 2 + LoRA), compute, Alan's supervision, the stage infrastructure — and it rides on the same burn being cut. The dossier's real question is therefore: what does the agent-run comeback cost per month in hard dollars (compute, pipeline, supervision), who pays it, and what is it worth against the 2026 AI-influencer market — not "kill or keep" in the abstract. Her own revenue history ($12M peak) is the comeback's prize if the strategy works; the plan must show a payer before it can call the prize reachable.

The named human: Alan Carr (Slack U9S023RQS; GitHub AlanFalcon — a CryptoKitties community veteran: KittyVer.se open-source CK RPG 2019, kittyverse lib, the Catatonic CK podcast; today runs ATZ + the Miquela stage). Both in-flight programs route through him. The dossiers' stakeholder work treats him as the operator of both options, and the Octopus Rodeo AI Showrunner Playbook (docs/ai-show-runner-playbook-3_0_0.md) as the shared operating machinery.