Risk 4/5 · provisional ⓘ Medium term · matures 12 Mar 2027 · by @mara.sol Head Chef · Automation on: the burner spends the pot's fuel to rebalance when holdings drift past their bands.
Sovereign Compute
“SI demand outruns chip supply through 2027; own the picks and shovels.”
The thesis
Frontier SI training and inference demand grows faster than fab, power and GPU capacity can follow. Owning the bottlenecks (foundry, accelerators, decentralised compute and the chains that settle it) captures that spread.
What would prove it wrong
- Hyperscaler capex guidance cut two quarters running
- GPU rental prices fall more than 40% year on year
Holdings
Live weights drift between rebalances; the rules pull them back inside the bands.
| Asset | Class | Target | Live | Weight (bar = live, ▼ = target) | Oracle | Why this weight |
|---|---|---|---|---|---|---|
| NVDAx | xStock | 14.0% | 13.1% | Raydium CLMM TWAP | Core compute leader; capped at 15% single-xStock limit | |
| TSMx | xStock | 12.0% | 12.4% | Raydium CLMM TWAP | Foundry bottleneck | |
| GOOGLx | xStock | 10.0% | 9.6% | Raydium CLMM TWAP | TPU + cloud capex | |
| SOL | Crypto | 20.0% | 21.3% | Pyth | Settlement layer for agent payments | |
| RENDER | Crypto | 14.0% | 14.9% | Pyth | Decentralised GPU supply | |
| ETH (Wormhole) | Crypto | 10.0% | 10.2% | Pyth | Compute marketplaces | |
| BONK | Meme | 3.0% | 1.8% | Raydium CPMM TWAP | Small cultural sleeve; hard cap 5% | |
| WIF | Meme | 2.0% | 2.1% | Raydium CPMM TWAP | Small cultural sleeve | |
| USDC | Cash | 15.0% | 14.6% | Fixed | Stable sleeve ≥5%; absorbs rounding |
Rules fixed at creation
Drift & bands
Preset “Steady”: rebalance at 4 pp absolute or 25% relative drift, cooldown 7 days, max slippage 1%.
Part 2 triggers
| Asset | Trigger | Action | Status |
|---|---|---|---|
| Part 2 trigger: when BONK hits 3×, offer holders BONK at 1.5%. | Part 2 offer: target 1.5% | Trigger | |
| Part 2 trigger: when RENDER hits 2×, offer holders RENDER at 7%. | Part 2 offer: target 7% | Trigger | |
| Part 2 trigger: when WIF hits 10×, offer holders WIF at 0%. | Part 2 offer: target 0% | Trigger |
Triggers don't change this pot. They open a Part 2 offer that holders can choose to take. When RENDER hits 2×, holders get an offer to move into Part 2 with RENDER at 7%. Nothing changes in this pot.
Rotation plan (via Part 2)
Rotation happens by migration. At the trigger, holders choose whether to move to Part 2. This pot never changes.
“In December, if the pot is up ≥25%, offer holders a Part 2 with no memes and 5% more USDC.”
Structured: time 1–31 Dec 2026 · condition NAV ≥ +25% · target memes 0% → USDC +5%
Term
Medium · matures 12 Mar 2027 · at maturity rotates to the stable sleeve (or into Part 2 if published).
Fuel · burner keeper bot active
Fuel: 0.087 SOL · Fuel 0.12 SOL · planned for the full term.
Add fuel Minimum 0.01 SOL. Any wallet except Fondue platform wallets; the SOL is spent, not refundable.If the pot runs out, automatic rebalancing stops; deposits and withdrawals keep working.
Rule hash 9f2c…a71e on-chain
SI vs Creator
Deviation from SI suggestion: 12 pp total.
| Asset | SI | Creator | Δ |
|---|---|---|---|
| NVDAx | 12% | 14% | +2 pp |
| SOL | 16% | 20% | +4 pp |
| BONK | 0% | 3% | +3 pp |
| USDC | 18% | 15% | −3 pp |
Activity
- BONK hit 3×: Part 2 offer open
2 Oct · - Pot locked: deposits on, creator role → platform fee wallet, managers renounced, 0 open intents
12 Jul · nobody can change fees or rules on-chain
Fees
| Who earns | In | Out |
|---|---|---|
| Platform (fixed at launch) | 0.75% · 75 bps | 0.75% · 75 bps |
| Creator @mara.sol: 100% of this tier, paid monthly in USDC | 0.50% · 50 bps | 0.50% · 50 bps |
| Pot (stays with holders) | 0% · 0 bps | 0% · 0 bps |
| Total | 1.25% | 1.25% |
No management or performance fee. Fees are taken in pot tokens at mint/burn. Fees are locked at launch: every later fee edit fails on-chain.
Worked example: $1,000
Buy $1,000 → $12.50 fees → $987.50 invested.
Sell later at the same NAV → $12.34 fees → $975.16 back.
Round trip: $24.84 (≈2.5%) + keeper ≈ $0.03 + slippage.