Flagship series · Crypto

Synthos Flagship Crypto Portfolio

A conviction, sector-weighted crypto portfolio built from the live four-pillar composite scores, the Synthos valuation framework, and the current macro-liquidity regime. Cited, rational, risk-taking — and honest about what it can't forecast.

Synthos Research — Flagship Series

Date: 2026-07-04 | Author: Fable, portfolio architect

Sources of truth: results/crypto/scored.json (73-token scored universe), docs/crypto_framework_fable.md (sector/valuation framework), results/macro/regime_2026-07-04.md (macro regime).

Not investment advice. This is a research construct expressing the Synthos framework's current view. See Section 6.

1. THESIS & RISK POSTURE

The one-paragraph thesis

The macro engine says we are in Reflation tipping toward Inflation under fiscal dominance, with the single highest-conviction forward vector being a policy-forced liquidity injection inside 6–18 months (empty RRP, bank reserves −4.8% YoY, M2 already +5.6% — the mechanical QT-stop trigger is close; Visser sizes the injection at $6–8T). Crypto is the purest expression of the liquidity cycle — the regime gate does more work here than in any equity sector. Meanwhile the tape is nowhere near euphoria: BTC sits at $63,245, −49.8% from its October 2025 ATH of $126,080; ETH is −63.6% from ATH; SOL −71.9%; the median token in our universe is down 80–95% from its high and down 40–70% over one year. That combination — a forward liquidity turn into a market that has already been cut in half — is the asymmetry this portfolio is built to own. We accumulate ahead of the injection, not after it, exactly as the macro report instructs ("position ahead of the injection, not after it").

Risk posture: constructive, front-footed, but tranche-disciplined

• Constructive because the liquidity turn is the forward base case and prices are deep in drawdown (this is early-cycle pricing inside a late-cycle macro — the fat pitch in crypto historically).
• Tranche-disciplined because the regime is "constructive-but-late": inflation is re-accelerating (CPI 4.3%, core PCE 3.4%), the Fed put is dead, and the injection's timing is not forecastable even though its direction is. We therefore buy in tranches on weakness rather than all-at-once, and we let the regime gate (Section 5) scale gross exposure.
• Concentrated by design. This is a conviction portfolio: 20 positions, top 5 = 58% of the book. The composite scores pick the names; the framework's quality hierarchy sets the sector weights.

The scope caveat (say it plainly)

This portfolio is 100% within crypto. It answers "given a crypto allocation, how should it be built?" — it does NOT answer "how much crypto should you own?" Sizing crypto against the rest of a book is the investor's decision and depends on their liabilities, horizon, and stomach. The macro regime (debasement, fiscal dominance) argues for some hard-money allocation; the number is yours.

2. TARGET SECTOR ALLOCATION

The rationality test, stated up front

Sector weights are set by three things, in order: (1) durability of value capture (does the token own real cash flows or a monetary premium?), (2) score evidence (where the Attractive-tier names actually live in scored.json), and (3) regime fit (what the liquidity/debasement regime pays for). Not equal-weight, not mcap-weight, not vibes.

The Layer-1-versus-meme case, concretely: L1s get 25%, Meme gets 3% — an 8-to-1 ratio. L1s are fee-generating digital economies with the strongest network moats in the asset class (SOL composite 65.3 / Attractive with adoption pillar sA 90.9; ETH moat pillar sM 90.6 with a fee-burn mechanism and 100% float). Memes are, per our own framework, "attention assets — momentum only, fundamental value: none, expected long-run value: zero for most." A sane model can still own memes small — the meme sleeve is the sentiment thermometer of the liquidity cycle and SPX6900 is literally the highest composite in the universe (78.6) on pure momentum — but only at a size where a 100% loss costs 3% of the book. That is what "weighs them accordingly" means.

Allocation table (11 sectors, sums to 100%)

The two zeros, justified in writing

DePIN — 0%. The framework's own honesty ratio condemns the sector today: demand/emissions coverage < 1 is the norm — these are "machines that convert token dilution into hardware deployment while real demand catches up (or doesn't)." The scores agree: the sector's best composite is AR at 56.5 (Neutral), followed by IOTX 51.0, FIL 49.8, GEOD 47.5, HNT 45.2 — zero Attractive-tier names, and three of five constituents (IOTX $28.5M, HNT $45.5M, GMX-style sub-floor sizes) carry sub-$100M flags, i.e. they fail our own universe floor. HNT is −99.5% from ATH with a −56.4% 30-day move — that is a value trap profile, not a value profile. The genuinely DePIN-shaped demand story we do believe in (paid compute/bandwidth) is already owned through the AI sleeve — GRASS and AKT are physically DePIN, economically AI-demand, and classified accordingly. What changes this: any DePIN name printing demand/emissions coverage trending through ~0.7 toward 1.0 (Helium data-credit burn, GEODNET subscription burn) re-opens the sector at 2–3%.

Exchange Tokens & Payments — 0%. Two industries, two disqualifiers. Exchange tokens carry single-company regulatory annihilation risk — the framework's own words: "one enforcement action is the whole thesis" — and the scores don't pay us for it: OKB 56.5, CRO 51.2, BNB 48.5, all Neutral, none Attractive. BNB's composite is dragged by a 22.3 adoption pillar (sA) — the burn is real but the growth isn't. Payments (XRP 45.5, XLM 48.3) fail the value-capture test: enormous float versus measured settlement revenue, XRP still −67.6% from ATH with 38% of supply yet to circulate (float ratio 0.623) and a standing 1B/month escrow-release overhang — deterministic sell pressure, our most forecastable negative signal. Regime logic doesn't rescue them: in a liquidity upturn these are lower-beta expressions than L1s with none of the moat. What changes this: a hard revenue link (e.g., a fee-share/burn tied to measured settlement volume) or an Attractive-tier composite; BNB re-enters first if its adoption pillar recovers above ~50.

Why the top of the table is the top

• SoV 30% > everything: BTC is the only asset in the universe whose entire thesis is the regime call itself (fiscal dominance → debasement → non-sovereign scarcity). It has no dilution (float 1.0, 20.05M of 21M mined), no value-capture ambiguity, and the deepest liquidity. When correlations collapse — and Section 6 says they will — BTC falls least and recovers first. Note its composite of 50.0 is the sector anchor by construction (all four pillars = 50.0 in scored.json — it is the benchmark other SoV names would score against), not a mediocrity signal.
• L1 25% > DeFi 20% combined: L1s own the moat (Metcalfe residuals, economic-value-secured) and the optionality on every app sector; DeFi owns better current cash flows but rents its land. We pay up for sovereignty, but we pay DeFi real weight because that is where the Attractive-tier cash-flow names actually are (AAVE 66.3, ETHFI 65.2, AERO 64.9).
• L2 6% despite TIA at 69.9: the tokenomics pillar dominates L2 scoring for a reason — governance-only tokens with large unlock overhangs (ARB, OP, STRK are the framework's canonical case studies). We own the two exceptions (TIA's modular-DA story, POL's fully-diluted 1.0 float with sT 84.5) and refuse the rest.

3. HOLDINGS — 20 POSITIONS

Concentration note: top 5 positions = 58% of the book. Every pick cites its composite / tier / decisive pillar and value-capture flag from scored.json.

Sleeve A — Monetary core (30%)

Sleeve B — Layer 1 platforms (25%)

Rejected on scores, not vibes: SUI (37.0, Weak — float 0.405, sT 8.4, P/REV 6,520x: the low-float/high-FDV archetype our tokenomics overlay exists to avoid), APT (32.0, Weak), SEI (32.9, Weak), AVAX (43.7 — P/fees 2,310x with a 32.1 value pillar).

Sleeve C — DeFi: Exchanges (12%)

Rejected: GMX (62.8 — but $62.9M mcap carries the sub-$100M flag and fails the universe floor), UNI (42.9 — GOVERNANCE-only; you are pricing an eternal fee-switch option at 35.1x revenue you don't own), DYDX (30.6, Weak, P/fees 90.6).

Sleeve D — DeFi: Credit (8%)

Rejected: ENA (27.4, Weak — P/Rev 781.8; seigniorage structurally dependent on positive funding, exactly the thing that dies in a correlation collapse), MORPHO (52.1 — GOVERNANCE-only, sT 34.6), PENDLE (48.6), COMP (45.9 — P/Rev 121.9).

Sleeve E — Infrastructure & Oracles (7%)

Rejected: GRT (58.3 — respectable, but third oracle/indexer duplicates the sleeve), ZRO (28.3, Weak — float 0.252), W (43.3, sub-$100M flag).

Sleeve F — Layer 2 (6%)

Rejected: ARB (45.9), OP (45.3), STRK (43.8) — all GOVERNANCE-only sequencer economics the token doesn't own, all with float ≤ 0.66; MNT (36.9, Weak).

Sleeve G — Liquid Staking & Restaking (5%)

Rejected: LDO (52.4 — massive $16.4B TVL but sA 20.1: share is eroding, the fees prove it: P/fees 0.6 yet P/Rev 9.6 on a thin 6% take), JTO (42.6 — P/Rev 92.3), EIGEN (39.7, Weak — float 0.403, sM 19.0).

Sleeve H — AI × Crypto (4%) — convexity sleeve

Sleeve I — Meme / Attention (3%) — convexity sleeve, hard-labeled

Label shipped verbatim per framework: "attention assets — momentum only, fundamental value: none, expected long-run value: zero for most, position accordingly." These are momentum instruments and regime sensors, not investments. Sized so a 100% loss costs 3%.

Rejected: DOGE (48.0 — sM 90.5 brand moat but sA 33.0: the attention is elsewhere this cycle), TRUMP (36.2, Weak — float 0.237, sT 1.0, the worst tokenomics score in the universe; unlock cliffs into a meme is the exact combination the overlay forbids), PEPE/SHIB/WIF/FLOKI/PENGU (44–53, no edge over the two chosen).

Aggregate convexity sleeve (clearly labeled)

HYPE 3% + GRASS 3% + SPX 2% + AKT 1% + BONK 1% = 10% high-risk/high-upside. Each position either fights one pillar deliberately (HYPE vs tokenomics) or has no fundamental floor (memes). Expected outcome distribution is barbell: several may go to ~zero; the sleeve exists because in the projected regime its upside is a multiple, not a percentage. Total portfolio survives the sleeve going to zero with 90% intact.

4. PER-POSITION PLAYBOOK

Three management styles, matched to asset nature — BTC is not a memecoin and is not traded like one:

• CORE HOLD + DCA: systematic accumulation on schedule regardless of price; add extra on defined drawdowns; no selling below prior ATH. For monetary/moat assets.
• ACCUMULATE-ON-DIPS: build in 3 tranches on weakness toward defined support logic; trim only on extreme extension. For quality fundamentals in drawdown.
• TRADE-AROUND-CORE: hold a half-size core; add the rest on pullbacks; take profits into strength; momentum/stop rules active. For extended, low-float, or attention-driven names.

Entry framing uses each token's live price and ATH distance from scored.json. Exact moving-average levels are layered in at render time; the logic is specified here.

5. REBALANCING & RISK RULES

The regime gate (sizes crypto; the composite ranks tokens)

The macro engine's liquidity posterior scales gross exposure, not picks:

Standing rules

1. Rebalance monthly to targets; reconstitute quarterly with the framework's taxonomy review. All changes logged append-only.
2. Sector caps: no sector above its target +5 points between rebalances; single position cap 33% (BTC exempt to 35%).
3. Unlock-overhang discipline (the most deterministic signal we have): no adds to any position within 30 days of an unlock cliff > 1% of circulating mcap; any position facing a cliff > 5% is trimmed to half ahead of it (live watchlist: HYPE float 0.233, JUP 0.484, AERO 0.498, TIA 0.801-and-vesting, PYTH 0.787).
4. Tier-migration rule: any holding downgraded to Weak (< 40) is halved at the next rebalance; downgraded to Avoid (< 25) or hit by a security/bad-debt incident is exited immediately (framework hard override).
5. Value-capture watch: flags reviewed monthly. An upgrade (e.g., a UNI-style fee switch actually passing) can admit a new name; a downgrade (governance redirecting revenue) is a trim trigger.
6. Meme sleeve stops are mechanical (Section 4): turnover collapse or negative sector-relative momentum = exit. The meme sleeve also serves as the froth dial: if meme sector mcap share spikes to blow-off readings, that is a late-cycle regime signal — we take it as instruction to de-risk the whole book, not to add.
7. What would change the sector weights: DePIN demand/emissions coverage trending toward 1 (opens 2–3%); ExchPay gaining a hard revenue link (opens 2–3%); L2 fee-switches actually passing (L2 to 8–10%); the liquidity turn confirming (convexity sleeve to 12% max); the macro call flipping deflationary (everything compresses toward BTC).

6. HONEST RISKS

1. Correlation collapse is the base case in any drain, not a tail. The framework says it plainly: crypto sector diversification is thinner than the taxonomy implies — in a liquidity drain, every sector is one trade. This portfolio's "diversification" is really relative-selection within one macro bet. If liquidity turns down instead of up, everything here falls together and the 30% BTC anchor merely falls least.
2. The macro call itself is the concentrated risk. The entire construction leans on the regime report's forward read (liquidity injection within 6–18 months). If instead sticky 4%+ inflation forces genuine tightening — or the bond market forces the "capital-flight, everything-down" tail Gromen flags — the accumulate-ahead-of-the-turn posture is early, and early in crypto can mean −50% before right. The tranche discipline is the mitigation, not a cure.
3. Reflexivity cuts both ways. Usage attracts price attracts usage; our adoption pillars (SOL sA 90.9, TIA 89.4, AERO 90.2) are partly measuring the reflexive tide we hope to ride. When it reverses, adoption scores will decay after prices do. The framework expects its own priors to decay; so should the reader.
4. Specific position risks, named: HYPE — float 0.233 with $51.7B of FDV yet to circulate; the buyback is strong but the vesting is stronger; one bad unlock cycle can dominate the cash-flow story. TIA — sV 26.2, near-zero fees; if modular DA loses to integrated L1s, there is no valuation floor. SKY — illiquid (turnover 0.005) and its seigniorage is a rates bet in disguise. GRASS/AKT — venture assets; failure is a normal outcome. SPX/BONK — expected long-run value zero; they are momentum instruments with mechanical exits.
5. Regulatory tail risk is structural and unmodelable. We sized it (0% exchange tokens, sector caps) rather than forecast it. A US securities-status shock to DeFi tokens would hit Sleeves C–G simultaneously.
6. Data is adversarial. Fees, volumes and adoption series can be wash-traded, Sybil-ed and incentive-inflated; the framework prefers burn/fee-paying series for exactly this reason, but no screen is clean. The composite can be right and the P&L wrong for quarters; calibration reporting (framework §4.2) is the discipline that keeps the scorecard honest either way.

Disclaimer: This document is research, not investment advice, and not a solicitation. Digital assets are extremely volatile and can go to zero. Positions, weights and playbooks describe a model portfolio constructed from the Synthos scoring framework as of 2026-07-04; they do not account for any individual's circumstances, jurisdiction, or risk tolerance. Do your own research.

APPENDIX — PORTFOLIO AT A GLANCE

20 positions. Top 5 = 58%. Convexity sleeve = 10%, clearly labeled.

Synthos Research — honesty is the product. The composite ranks tokens; the regime sizes crypto.