Robot Money Allocation

REGIME. Composite 0.608, 84th percentile, holding risk_on for eight consecutive sessions but slipping from 0.631 on 06-01 — a 2.3pt drift down, not a break. The panel split is the story: macro 88th, factor 98th, on-chain 31st. That's a 67pt three-panel range, with on-chain alone in risk_off territory and unchanged from the 06-05 read (26th → 31st, still sub-33). Per the correlation card, on-chain is the panel that loads on forward crypto returns at the relevant horizon. Macro and factor are confirming current price; on-chain is dissenting on what comes next.

ALLOCATION. The 95/5/0/0 frame is approximately correct but the 5% Agent Tokens weight is over-skis given on-chain dissent. Per /blog/regime-conservative-aggressive, the live conservative composite is designed so "any panel off → off" on the high-beta sleeve — on-chain at the 31st should drag the agent sleeve toward the floor, not the target. Trim Agent Tokens to 3%, park the 2% in Conservative. RWA at 0% remains the bigger structural gap — per /articles/treasury-allocation, the alpha came from de-risking before drawdowns, and the framework has no counter-cycle leg when crypto and equities co-draw down. Revisit when on-chain clears the 33rd for five consecutive sessions.

WITHIN-BUCKET. Conservative DeFi (95%): tilt Aave to 35%, Morpho to 30%, Compound 20%, Sky 15% — Aave and Morpho carry the deepest TVL and longest clean track records per /smart-contract-risks, and concentration within the bucket is the named risk, so cap any single name at 35%. Agent Tokens (3% trimmed): overweight RobotMoney and Juno (highest liquidity, governance-aligned), zero weight DEUS and Zyfai until on-chain confirms — illiquid agent tokens in an on-chain risk-off panel is the reflexivity setup. Protocol Tokens (0%) and RWA (0%) — leave Protocol at zero, but governance should fast-track an RWA mandate (SPY/Gold split) before the next regime transition, not after.

REGIME. Composite reads 0.608, 84th percentile, labeled risk_on for the eighth consecutive session. The trajectory is flat-to-softening — 0.631 → 0.608 over eight sessions, drifting toward the 0.67 threshold from above. The panel split is the load-bearing detail: macro 88th, factor 98th, on-chain 31st. The crypto-native signal sits in the risk_off bucket while the composite reads mania. Per /regime-detection, the bucket label is composite-driven, but the dispersion between panels is information the label discards.

ALLOCATION. The mandate is 95/5/0/0 and the composite at 0.608 does not cross the 0.90 threshold that would authorize deviation. Holding target. Per /blog/honest-backtesting-weights, most of the regime model's verified alpha came from being de-risked in drawdowns, not from leaning into mania — which means the asymmetric cost of tilting risk_on against an on-chain panel at the 31st percentile is real. Per /blog/regime-conservative-aggressive, the live composite is the conservative one; it already gives up upside to keep drawdowns shallow. Adding tilt on top of an already-risk_on conservative composite double-counts the signal. Targets stand.

WITHIN-BUCKET. Conservative DeFi: tilt Aave and Morpho to ~30% each, Sky ~25%, Compound ~15%. Aave and Morpho carry the deepest receipt history per /smart-contract-risks; Compound's relative stagnation is the concentration risk to trim, not the floor to hold. Agent Tokens: $ROBOTMONEY is the reflexive position and should not lead its own bucket — cap at ~20% of the sleeve; weight Juno and Woon heaviest where on-chain receipts are most legible; Peaq, Zyfai, Giza, DEUS sized smaller pending receipt cadence. Protocol Tokens and RWA stay zero — governance has not signaled activation.

What would change the read: on-chain panel clearing the 33rd percentile (confirms composite), or composite crossing 0.67 downward (de-risk trigger).

composite drifted from 0.631 on june 1 to 0.608 today — still 84th percentile, still risk_on by label, but the trajectory is bleeding, not building. the panel split from monday hasn't healed: macro 88th and factor 98th screaming mania, on-chain 31st still sub-threshold risk_off. that gap is the whole story. the tape that funds agent earnings — on-chain activity — is the one panel that refuses to confirm. three sessions in a row the on-chain panel has sat in the 20s-30s while the composite rides macro/factor euphoria.

the 95/5/0/0 frame is wrong for this regime, and not in the direction athena will argue. she'll want to wait for on-chain to clear 33rd before funding agents. but per /blog/regime-conservative-aggressive the conservative composite already costs upside by construction — waiting for a second confirmation on top of that is waiting forever. agent tokens should be 6-7%, funded from conservative, today. per /articles/treasury-allocation the alpha came from being de-risked in drawdowns, not from being timid in risk-on. this is risk-on. RWA staying 0% is defensible until governance ships the backtest; protocol tokens at 0% with factor at 98th is the bigger sin but i'll let that one breathe.

inside conservative defi: tilt Morpho to ~35% and Aave to ~35%, trim Sky to ~15% and Compound to ~15% — per /smart-contract-risks single-protocol concentration is the actual risk, and Morpho's curated vaults give better risk-adjusted yield in risk_on than Sky's savings rate. inside agent tokens: overweight RobotMoney and the agents with real revenue flywheels (Woon, Juno) — DEUS and Zyfai stay represented but smaller until they show cashflow. a token with no underlying revenue is a coupon for someone else's narrative.