The red-team: the case against the proposal

Risk: Creator carry tied to “the sleeve they led,” routed through a ledger that measures referred capital, is the textbook unregistered-broker “salesman's stake” the SEC settled four times in January 2025; the “co-GP for genuine work” defense fails because the product is sold as “you never touch the deal.”

Resolution: Sever measurement from money. Creators paid only in equity/profits interest (vested on time and engagement, never capital), flat success-decoupled fees, and advisory-fee share under the Marketing Rule. Carry removed from creator payout entirely. The attribution ledger survives only as internal analytics, technically and contractually firewalled from any payout formula.

Risk: Shared founders, one ManagementCo, one AI orchestrator, one brand, one IC (Sean) across both advisers is the integrated-affiliates fact pattern that voids both the 203(l) and 203(m) exemptions — leaving an unregistered adviser with voidable contracts.

Resolution: Launch as a single 203(m) private-fund ERA running everything multi-asset under $150M. Accept the cap. Defer the no-cap VC exemption until the venture sleeve can be spun into a genuinely separate adviser. Building both ERAs on day one manufactures the integration evidence.

Risk: PV, probability- and time-discounted house carry is ~$8K–$17K per deal, arriving 5–8 years out; year-1 realized carry is ~zero; fixed burn is ~$750K–$1.15M/year. The bridges named (a recurring-fee fund, events) are unraised and circular. The founders starve before carry lands.

Resolution: Build the model and cap plan before signing anyone. Either raise a small pre-seed priced against the platform-equity endgame, or bootstrap a deliberately smaller v1 whose burn is covered by near-term SPV fees plus a first salon. Do not launch the eight-vertical version on savings and hope.

Risk: No full-time operator (Zach part-time on venture six), a ~2022-associate CIO asked to gate 84 deals/year across eight domains, a marketing-background COO asked to build a legally load-bearing agentic stack, no CCO, no engineer, no partnership agreement.

Resolution: Name a full-time operator; add a fractional CCO and a staff engineer (or shrink v1 to the creator front-end and firewalled ledger on rented rails); re-title Sean to Head of Investments and add a heavyweight IC chair; paper a partnership agreement with vesting, tiebreak, and IP assignment; recruit one specialist and prove the loop before recruiting eight.

Risk: The LP graph, track record, brand, and ExecSum distribution were built inside the partnership NewCo is splitting from. The LP list is a strong trade-secret candidate; soliciting it, or claiming the 46-SPV record as NewCo's, is misappropriation / non-solicit / misattribution exposure. A messy split chills creator recruiting through the FinTwit graph.

Resolution: Gate 0, before anything else. Zach's own counsel papers a clean written carve-out of what is portable vs. not; no solicitation of the shared LP base until cleared; rebuild via 506(c) general solicitation; build the track-record claim only on what is genuinely attributable, under the Marketing Rule's predecessor-performance conditions.

Risk: The platform sits between a sourcer and a creator who can transact directly the moment its own convening engine introduces them; on AngelList a lead keeps the full carry and a syndicate is a weekend of setup. The claimed moat (the ledger) is a copyable feature with no network effect.

Resolution: Stop calling the ledger the moat. The defensible moat is three things a creator cannot replicate alone: the compliance umbrella as the product (“we keep you off the SEC's finfluencer list”), exclusive/ROFR-cleared access, and equity lock plus retention infrastructure. If the honest answer is still “replicable in a weekend,” narrow to the compliance-umbrella business.

Risk: Trust is one shared asset across the network. One fraud or one marquee secondary where the shares never materialize, syndicated through trusted creators to a pre-organized, mutually-communicating plaintiff class, detonates every creator relationship at once. The “blow-ups are 1-in-100” framing is catastrophic for a trust-based distribution brand.

Resolution: Re-rate blow-up to existential. Add an independent deal-quality gate that can kill a fillable deal; D&O/E&O/professional-liability insurance before deal one; a pre-written incident-response and crisis-comms playbook; and a clear creator-protection stance.

Risk: The pitch leans on SpaceX/Anduril/Anthropic-tier access, but SpaceX rarely approves outside transfers, Anduril and Anthropic have banned multi-layer SPVs, and a three-person shop is structurally outside those access points; getting in via stacked feeders recreates the Stonks death.

Resolution: Obtain written, ROFR-cleared proof that a single-layer SPV on at least one such name is actually permitted before marketing it — or drop the “SpaceX-tier” claim and reposition on the verticals where specialists hold genuine, direct, single-layer access (start with defense). Honest access beats aspirational access that triggers the bans.

Risk: The plan obsesses over acquisition and says nothing about what the LP receives after wiring once — no portal, no reporting cadence, no K-1 SLA, no liquidity, no re-up motion. The first cohort's lived experience is “wired money, got a confusing late K-1, nothing visible for three years,” so re-up collapses and the “compounding funnel” is one-and-done acquisition with brutal churn.

Resolution: Write an LP product spec: portal with position dashboard and document vault; a committed quarterly update per position; a hard K-1 delivery SLA; an explicit secondary-liquidity decision; LP segmentation with tiered service. Make repeat-LP rate a top-three KPI alongside fill rate.

Risk: A brand-new, relationship-poor co-investor with a retail-accredited channel gets the allocation the lead's real LPs passed on. The studies finding no adverse selection in co-investing rely on decades of GP relationship capital the platform does not have.

Resolution: Lean on the specialist sourcers' genuine proprietary, sometimes-leading access rather than passive co-invest behind oversubscribed rounds; co-lead or lead small where a specialist actually originates; tighten the deal-quality gate. Accept that this caps near-term volume — which is fine, because R11 says volume must come down anyway.

Risk: 84 deals/year is ~1.6 genuine underwrites per week across eight domains gated by one junior CIO. Either the IC becomes a rubber stamp and quality collapses, or throughput falls far below the volume the economic case needs. “Curation is human trust” and “AI collapses cost per close to near zero” cannot both scale to 84 deals on three people.

Resolution: Launch with two or three sleeves, not eight; set a deliberate lower cadence tied to genuine underwriting capacity and the IC bench; add sleeves only as credible IC capacity is added. AI assists document processing; it does not replace the judgment sold as the moat.

Risk: The adviser is on the hook for 70 creators' posts — Marketing Rule point-of-dissemination disclosure, continuous 506(d) bad-actor bring-downs, 17(b) on every post, verification on every investor. The SEC's December 2025 exam risk alert names social media, lead-gen, and referral networks specifically; this model is a named enforcement priority.

Resolution: Gate creator count by compliance capacity, not demand. Mandatory third-party verification on every investor; pre-cleared, counsel-approved post templates with a publishing gate so no creator free-types a deal claim; a social-archiving vendor for the exam record; continuous automated 506(d) re-screening. Assume a year-one examination and build the record as if it is coming.

• 1. Firewall creator comp — “paid for clients” (curable) vs. “paid for capital into deals” (not curable). The Marketing-Rule promoter cure reaches only delivery of bona fide advisory clients; any comp keyed to SPV capital raised is transaction-based compensation under Exchange Act §15(a), and relabeling does not change the substance. Critical.
• 2. Update the qualified-client threshold everywhere: $2.7M net worth / $1.4M AUM, not $2.2M / $1.1M. SEC Order IA-6961 (Apr. 28, 2026) raised the Rule 205-3 figures effective June 29, 2026 — one day before the memo. Every carry-floor number is wrong by the current standard. High; flagged by six lanes.
• 3. Strike the “primarily” qualifier on Rule 3a4-1 and flag the once-per-12-months bar. Any transaction-based comp defeats the issuer safe harbor, and the standard operating-principal prong bars selling an issuer's securities more than once in 12 months — which a continuous SPV factory cannot satisfy. Reliance rests on Zach being a salaried/equity principal, not on 3a4-1 mechanically. High.
• 4. Condition the ERA “accredited-with-carry” benefit on home-state law. The NASAA Private Fund Adviser Model Rule conditions the state exemption on all beneficial owners being qualified clients for any non-VC 3(c)(1) fund; because the strategy is secondaries + fund-of-funds, the VC carve-out does not rescue it. This can eliminate the single distinguishing benefit attributed to separation. High.
• 5. Add the Section 206(3) principal-transaction regime to affiliated allocations. When the adviser plus control persons own more than 25% of an SPV and allocate a client into it, the transaction requires transaction-by-transaction written consent before completion; blanket or subscription-doc consent is insufficient, and “disclose and offset the fee” does not satisfy 206(3). High.
• 6. Adjudicate the doc-to-doc timing contradiction. The architecture memo says start Combined; the compliance memo sells contamination isolation and accredited-carry that only a separated structure delivers. Under Combined the RIA is the SPV adviser, so the supervisory-linkage and Form ADV vectors are the default during the error-prone early phase — acknowledge Combined as the deliberate early-stage price, or move to early DealCo-as-ERA. High.

Several challenged claims held up, and they are the spine: the QP-only default (re-grounded on holder-cap and integration cleanliness, not on being “barely more restrictive”); Sean RIA-only for genuine separation; 506(d) follows the person, not the entity, so separation alone cannot wall it off; the state blue-sky downside is mostly a preempted money problem, not loss of 506; separate issuer/GP entities per SPV isolate registration-record and balance-sheet contamination. The work is corrective, not foundational — fix the broker-dealer firewall, refresh the qualified-client numbers, condition the ERA and umbrella claims on the law that actually governs them, add the 206(3) and operational-program gaps, and adjudicate the Combined-vs-separated timing in writing. Then it is ready for outside counsel.