Smarter than the average bear market vault. The vault that survives when others bleed.
Yogi is a production-grade USDC vault on Solana that runs dynamic tilted delta-neutral positions on Drift — buying spot and shorting perps to eliminate price risk while collecting funding. A 5-dimensional anomaly detection engine (including real-time cross-venue funding comparison against Binance and Bybit) dynamically adjusts tilt, deployment, and leverage. While BTC dropped 7.4% in one week, Yogi delivered +$8.08 (+1.61%) with zero drawdown. It's not the highest yield — it's the one that survives.
Live on Solana mainnet since March 20, 2026. $500 deposited. Zero drawdown. Keeper running 24/7 on AWS EC2.
We evaluated every major perp DEX for delta-neutral feasibility. DN requires both spot and perp markets on the same venue. Most don't qualify:
| Venue | Chain | Spot + Perp | DN Feasible | Status |
|---|---|---|---|---|
| Drift | Solana (L1) | SOL, BTC, ETH | Yes — multi-asset | $1.1B TVL, growing |
| Hyperliquid | Own L1 | HYPE only | Yes — single-asset | $4.5B TVL, dominant |
| Vertex | Arbitrum (L2) | Was yes | No — shutting down | Migrating to Ink (Kraken L2) |
| GMX v2 | Arbitrum (L2) | Pool-based | Possible but hard | AMM slippage, not order book |
| dYdX v4 | Cosmos | Perp only | No | Moved off EVM |
| GRVT | ZKsync (L2) | Perp only | No | No spot markets |
| Gains/gTrade | Arbitrum (L2) | Perp only | No | |
| Kwenta | Base/Optimism (L2) | Perp only | No |
No viable EVM venue exists for delta-neutral today. Vertex was the best EVM candidate but shut down on Arbitrum. With Ethereum refocusing on L1 scaling (Vitalik's recent direction), L2 platforms face existential uncertainty — building long-term yield infrastructure on an L2 is risky.
Drift and Hyperliquid are both independent L1s — no Ethereum dependency, no L2 migration risk. This wasn't accidental. It was a deliberate platform selection.
| Drift (Yogi) | Hyperliquid (Kodiak) | |
|---|---|---|
| DN-eligible assets | SOL, BTC, ETH, POPCAT, DRIFT (5 markets) | HYPE only (1 market) |
| Lending | Native — Drift auto-lends idle collateral | Bridge required to HyperEVM (HyperLend) |
| Funding settlement | Continuous | Hourly |
| Maker fees | -0.2 bps (rebate — you earn) | 1.5 bps (you pay) |
| Cross-margin | Spot + perp in one account | Separate accounts |
| TVL | $1.1B | $4.5B |
| Daily volume | $118M | $6.7B |
Drift's composability — spot + perp + lending in one account, no bridges — makes it the natural platform for institutional multi-asset DN. Hyperliquid's HYPE funding is structurally positive due to ecosystem conviction, providing stable single-asset yield.
Neither platform "won." They serve different purposes and complement each other. But if institutions are choosing one foundation for yield infrastructure, Drift's composability, diversification, and independence make it the stronger long-term choice.
To understand something properly, you need to see outside of it. We built on both. PerpU — our perp DEX learning platform — shares what we learned.
During a week where BTC dropped 7.4% ($73,872 → $68,402), Yogi:
| Metric | Value |
|---|---|
| P&L | +$8.08 (+1.61%) |
| Max Drawdown | $0.00 |
| Account Health | 98% |
| Net Funding | $0.07 |
| Positions | 4 (SUI, BTC, AVAX, DOGE — rotating) |
| Taker Volume | $232.45 |
DOGE funding collapse: Drift DOGE funding spiked to -3,271% APY. Yogi's cross-venue detector showed Binance at -6.4% and Bybit at -6.9% — confirming the distortion was Drift-specific. The keeper exited the DOGE SHORT and rotated to SUI (+1,034% APY).
Cross-venue divergence: Drift BTC funding at +1,445% while Binance at -7.5% and Bybit at -9.6%. A +1,454% spread. Yogi sees this and flags convergence risk on every BTC entry decision.
Regime discipline: Vol stuck at 57-61% (high regime). Yogi automatically limited deployment to 50% at 0.5x leverage. Many vaults would be fully deployed — Yogi chose safety, and the bear market proved it right.
Zero drawdown through a -7.4% BTC week. That's the story.
Yogi's primary mode is delta-neutral: buy spot + short perps on the same asset. Price movement cancels out — profit comes purely from funding rate collection. A dynamic tilt adds slight short bias (0-10%) in calm markets for extra yield, automatically reducing to 0% (pure DN) when signals detect stress.
User deposits USDC --> Voltr Vault
|
+-- Idle USDC --> Drift auto-lends to borrowers (1-5% APY)
|
+-- Deployed Capital --> Delta-Neutral Positions
|
+-- 70% --> Spot BUY (SOL, BTC, ETH on Drift)
+-- 30% --> Perp margin
|
+-- Perp SHORT = spot size × (1 + tilt%)
| Price cancels out. Funding collected.
|
+-- Dynamic Tilt (0-10%)
| +-- CLEAR signals + low vol → 10% (max short bias)
| +-- LOW signal → 3% short bias
| +-- HIGH/CRITICAL → 0% (pure DN, protect capital)
| +-- Negative funding → reduce tilt
|
+-- Signal Detector (every 5 min)
| +-- OI imbalance shift
| +-- Liquidation cascade (OI drop proxy)
| +-- Funding rate volatility
| +-- Spread blow-out (mark/oracle)
| +-- Cross-venue funding (Drift vs Binance/Bybit)
| --> Severity: CLEAR / LOW / HIGH / CRITICAL
|
+-- Regime Engine (vol × signal → deployment)
+-- Adjusts deployment %, leverage, tilt
+-- Emergency rebalance on regime shifts
| Step | Action | Purpose |
|---|---|---|
| 1 | Buy spot on Drift (e.g., 0.0036 BTC) | The hedge |
| 2 | Short perp on Drift (e.g., 0.0037 BTC) | Funding collection + tilt |
| 3 | Price goes up | Spot +$X, Perp -$X → net $0 |
| 4 | Price goes down | Spot -$X, Perp +$X → net $0 |
| 5 | Funding accrues | Short collects funding hourly |
| Result | Zero price risk, pure funding yield | + tilt bonus in calm markets |
Unlike static delta-neutral vaults, Yogi adjusts its short bias in real time:
| Condition | Tilt | Effect |
|---|---|---|
| Calm market (low vol, CLEAR signals) | 10% | Perp short 10% larger than spot → extra yield from short bias |
| Normal (LOW signal) | 3% | Mild short bias, conservative |
| Stress (HIGH/CRITICAL) | 0% | Pure DN — zero price exposure |
| Negative funding | 0-2% | Reduce tilt when shorts pay instead of collect |
The tilt is computed every rebalance cycle using signal severity, vol regime, and current funding direction. This is not available on any other Drift vault.
-
Delta-Neutral on Drift — buys spot + shorts perps within Drift's own markets. Zero price risk. Most Drift vaults are directional and bleed during drawdowns.
-
Dynamic Tilt — adjustable short bias (0-10%) responds to market conditions in real time. Pure DN in stress, tilted for extra yield in calm markets. No other vault does this.
-
5D Signal Detection — OI shifts, liquidation cascades, funding instability, spread blow-outs, AND cross-venue funding divergence. Five early warning systems, checked every 5 minutes.
-
Cross-Venue Intelligence — the only Drift vault that compares funding rates against Binance and Bybit in real time. When Drift DOGE funding is +2,000% but Binance is -6%, Yogi knows that's a distortion, not an opportunity.
-
CEX Open Interest Tracking — monitors $9.3B of BTC open interest and $1.3B of SOL open interest across Binance and Bybit. OI surges signal incoming volatility before it hits Drift.
-
Regime Discipline — vol at 57%? Yogi goes cautious (55% deployed, 0.5x leverage, 0% tilt). Not because it's told to, but because the deployment matrix says high vol + signals = protect capital.
-
Multi-Asset DN — SOL, BTC, ETH, POPCAT, and DRIFT all have spot+perp markets on Drift. Yogi can run DN on up to 3 simultaneously, dynamically selecting the highest-funding markets. Drift has 20+ DN-eligible pairs — a long tail of Solana-native assets unavailable on any other venue. Kodiak (Hyperliquid) is limited to single-asset DN on HYPE.
-
DN on DRIFT token — Yogi is the only vault running delta-neutral on Drift's own token. When DRIFT holders are long-biased, the perp short collects funding. This is only possible on Drift — no CEX or other DEX has DRIFT spot+perp.
Every other vault shows a backtest. Yogi shows live mainnet performance through a -7.4% BTC drawdown with zero loss.
| Source | Mechanism | Est. APY | Status |
|---|---|---|---|
| DN funding collection | Short perps collect positive funding | 8-15% | Live |
| Dynamic tilt bonus | Extra yield from short bias in calm markets | 1-3% | Live |
| Premium convergence | Mark/oracle deviation mean-reverts (via DN tilt) | 1-2% | Live |
| Cross-venue intelligence | Entry/tilt optimization via Binance/Bybit comparison | 0.5-1% | Live |
| Drift auto-lending | Idle collateral lent to borrowers natively | 1-5% | Live (native to Drift) |
| LST collateral | jitoSOL staking + MEV | 1.5-2% | Designed, post-hackathon |
| Live combined | 12-24% |
| Module | File | Purpose |
|---|---|---|
| Delta-Neutral Manager | src/keeper/delta-neutral.ts |
DN position lifecycle: open (spot+perp), close, delta drift check, dynamic tilt |
| Signal Detector | src/keeper/drift-signal-detector.ts |
4-dimension Drift anomaly detection (OI shift, liquidation, funding vol, spread) |
| Cross-Venue Detector | src/keeper/cross-venue-detector.ts |
5th dimension: compares Drift funding vs Binance/Bybit for convergence signals |
| Regime Engine | src/keeper/regime-engine.ts |
Vol x signal severity --> deployment %, leverage cap, tilt |
| Imbalance Detector | src/keeper/imbalance-detector.ts |
Reads OI, mark/oracle spread, funding — computes composite signal and direction |
| Yield Stacker | src/keeper/yield-stacker.ts |
Multi-protocol lending optimization, LST yield (designed, post-hackathon) |
| Funding Scanner | src/keeper/funding-scanner.ts |
Fetches and ranks all Drift perp markets by funding rate |
| Cost Calculator | src/keeper/cost-calculator.ts |
Maker fee model — 1.6 bps round-trip cost |
| Leverage Controller | src/keeper/leverage-controller.ts |
Dynamic leverage scaling by vol regime |
| Health Monitor | src/keeper/health-monitor.ts |
30-second health ratio and drawdown checks |
| Emergency Decisions | src/keeper/emergency-decisions.ts |
Pure decision logic for emergency actions — testable without DriftClient |
| Position Manager | src/keeper/position-manager.ts |
Directional position management (fallback mode) |
| Slippage Guard | src/keeper/slippage-guard.ts |
Checks order book depth before DN entry |
| Keeper Loop | src/keeper/index.ts |
Main event loop — DN rebalance, signals, regime, cross-venue |
| Config | src/config/ |
Strategy parameters, signal thresholds, deployment matrices, DN settings |
| Asset | Spot Market Index | Perp Market Index | DN Eligible |
|---|---|---|---|
| SOL | 1 | 0 | Yes |
| BTC | 3 (wBTC) | 1 | Yes |
| ETH | 4 (wETH) | 2 | Yes |
| POPCAT | 20 | 34 | Yes |
| DRIFT | 15 | 30 | Yes |
Drift has 20+ additional spot+perp pairs (JUP, WIF, PYTH, TRUMP, etc.) that could be added as DN candidates. The keeper dynamically selects the highest-funding markets from the eligible list.
The regime engine is Yogi's core differentiator. It combines two inputs into a deployment decision:
| CLEAR | LOW | HIGH | CRITICAL | |
|---|---|---|---|---|
| Very Low (< 20% vol) | 100% @ 2.0x | 95% @ 1.5x | 70% @ 1.0x | 40% @ 0.5x |
| Low (20-35%) | 95% @ 1.5x | 85% @ 1.2x | 55% @ 0.8x | 30% @ 0.3x |
| Normal (35-50%) | 85% @ 1.0x | 70% @ 0.8x | 45% @ 0.5x | 20% @ 0.2x |
| High (50-75%) | 75% @ 0.8x | 55% @ 0.5x | 30% @ 0.3x | 15% @ 0.0x |
| Extreme (> 75%) | 0% @ 0.0x | 0% @ 0.0x | 0% @ 0.0x | 0% @ 0.0x |
Matrices are loosened vs pure directional because DN positions have partial price hedging — shorts and longs offset, so higher deployment is safer.
| Condition | Tilt % | Rationale |
|---|---|---|
| CLEAR + veryLow/low vol | 10% | Max short bias — calm markets, funding favorable |
| CLEAR + normal vol | 7% | Moderate short bias |
| CLEAR + high vol | 5% | Reduced tilt, vol uncertainty |
| LOW signal | 3% | Conservative tilt |
| HIGH/CRITICAL signal | 0% | Pure DN — protect capital |
| Negative funding | max 2% | Reduce tilt when shorts pay |
| Dimension | LOW | HIGH | CRITICAL |
|---|---|---|---|
| OI Imbalance Shift | 5% in 1h | 15% | 30% |
| Liquidation Cascade (OI drop) | 5% in 1h | 15% | 30% |
| Funding Rate Volatility | 500 bps annualized | 1500 bps | 3000 bps |
| Spread Blow-out (mark/oracle) | 0.5% | 1.5% | 3.0% |
5th Dimension — Cross-Venue Funding:
| Signal | Condition | DN Adjustment |
|---|---|---|
drift_high |
Drift funding > CEX by 5%+ APY | Confirms DN profitability, may increase tilt |
drift_low |
Drift funding < CEX by 5%+ APY | Flag convergence risk, reduce tilt |
aligned |
Drift ≈ CEX (within 5% APY) | High confidence — normal operation |
| Parameter | Value |
|---|---|
| Max drawdown | 3% reduce / 5% close all (both DN and directional) |
| Max leverage | 2x (regime-adaptive) |
| Health check | Every 30 seconds |
| Signal detection | Every 5 minutes |
| Health critical | Close all at 1.08 |
| Signal CRITICAL | Force reduce + set tilt to 0% |
| Max per market | 40% |
| Max DN markets | 3 simultaneous (from 5 eligible: SOL, BTC, ETH, POPCAT, DRIFT) |
| Min funding APY | 5% to open DN position |
| Delta drift threshold | 5% — rebalance legs if spot/perp diverge |
| Max slippage per leg | 1.5% (raised for small-cap DN markets) |
| DN capital split | 70% spot / 30% perp margin |
| Negative equity | Emergency close all (both legs) |
| Risk | Mitigation |
|---|---|
| Spot/perp size diverge | Delta drift check every rebalance — rebalance legs if >5% |
| Funding turns negative | Exit DN when funding APY drops below 5% threshold |
| One leg fails to fill | Automatic unwind — if perp short fails, sell spot immediately |
| Tilt exposure in crash | Dynamic tilt → 0% on HIGH/CRITICAL signals (pure DN) |
| Startup with stale positions | Position loading reconstructs DN pairs from on-chain state |
| Transition from directional | Auto-closes legacy directional positions on DN mode startup |
$500 is proof-of-concept, not proof-of-scale. We know these risks exist. Here's our plan.
| Risk | Impact | Severity at $1M+ | Mitigation Plan |
|---|---|---|---|
| Drift spot liquidity | Slippage on spot buys/sells erodes DN yield. Drift SOL spot does ~$2-5M daily volume — a $1M vault moving 40% into one market hits real slippage. | High | Split entries across multiple blocks. TWAP execution over 5-10 minutes. Reduce max-per-market from 40% to 25%. Multi-asset DN distributes impact across 3 markets. |
| Tilt exposure in flash pumps | 10% tilt = 10% unhedged short. A 15% pump in <5 min (faster than detection cycle) causes ~1.5% loss before tilt reduces to 0%. | Medium | Cap tilt at 5% above $500K AUM. Add 1-min fast-path signal check for price moves >3%. WebSocket price feed for sub-minute detection (post-hackathon). |
| Single keeper SPOF | One EC2 instance, one process. If it dies during a stress event, positions stay open with stale tilt/leverage. pm2 restarts help but don't cover AZ outages. | High | pm2 auto-restart covers process crashes. Position loader reconstructs state on restart. Multi-region keeper with leader election planned for >$100K AUM. Health endpoint for external monitoring. |
| Oracle/mark divergence at size | Larger DN positions influence Drift's AMM. Own orders can move mark price, creating phantom delta drift and unnecessary rebalances. | Medium | Monitor own-order market impact. Widen delta drift threshold from 5% to 8% at scale. Use limit orders instead of market orders above $100K. |
| Funding rate compression | More DN capital on Drift = more shorts = funding rates compress. The strategy partially erodes its own edge at scale. | Low-Medium | Multi-asset DN distributes pressure. Cross-venue monitoring detects compression early. Hard floor: exit DN when funding < 5% APY regardless. Vault capacity cap at $5M until Drift spot liquidity grows. |
Bottom line: The current architecture works cleanly at $500-$50K. Between $50K-$500K, execution improvements (TWAP, limit orders, tighter tilt caps) are needed. Above $500K, infrastructure changes (multi-region keeper, capacity caps) become essential. We've designed for this progression — the risk management framework scales, the execution layer needs investment.
32-day comparative backtest (Feb 13 – Mar 16, 2026):
| Metric | Baseline (vol-only) | Yogi (intelligent) |
|---|---|---|
| Final equity | $100,619 | $100,516 |
| Total return | +0.62% | +0.52% |
| Annualized APY | 7.06% | 5.89% |
| Max drawdown | 0.05% | 0.04% (-21%) |
| Sharpe ratio | 17.64 | 17.60 |
| Trading costs | $371 | $312 (-16%) |
Note: Backtest was run with directional mode. DN mode reduces drawdown further by eliminating price risk entirely.
| Fee | Amount |
|---|---|
| Management fee | 1% annual |
| Performance fee | 20% of profits |
| Deposit fee | None |
| Withdrawal fee | 0.1% |
| Withdrawal period | 24 hours |
62 assertions across 7 test suites covering strategy modules and emergency logic:
npm test| Suite | Module | Assertions | Coverage |
|---|---|---|---|
| cost-calculator | Cost gating, maker fee model | 10 | Trade economics, break-even, profitability |
| leverage-controller | Vol regime classification | 6 | Regime boundaries, leverage caps |
| funding-scanner | Market ranking, filtering | 4 | Whitelist/blacklist, rate sorting |
| imbalance-detector | Signal scoring, direction | 8 | Composite signals, market filtering |
| regime-engine | Deployment matrix | 6 | Vol×signal matrix, emergency rebalance triggers |
| drift-signal-detector | Anomaly detection | 4 | Severity levels, formatting |
| emergency-decisions | Emergency action logic | 24 | Drawdown/health/signal priority chain, DN close regression, orphaned spot min order size |
The emergency-decisions suite includes regression tests for two production bugs discovered on 2026-03-30:
- Drawdown close_all must close DN positions — the original code only closed directional positions, leaving DN positions stuck in an infinite emergency loop
- Orphaned spot below Drift min order size — 0.0999 SOL (below 0.1 SOL minimum) caused repeated transaction failures on startup
- Pitch presentation:
demo/presentation.html— 8-slide auto-advancing pitch (80 seconds) - Live dashboard:
demo/dashboard.html— real Drift data + signal detection, no server needed - Demo video:
demo/yogi-demo.mp4— generated from presentation slides - Voiceover script:
demo/voiceover-script.md— narration for each slide
open demo/presentation.html # View pitch deck
open demo/dashboard.html # Live monitoring dashboard
node demo/record.js # Generate demo video (requires puppeteer + ffmpeg)Yogi is live on Solana mainnet with an automated keeper running 24/7 on AWS EC2.
| Component | Address |
|---|---|
| Voltr Vault | BFDTTG8nJF7uLf3wsqFJpYCvC6wA6BahBRKjTgtKPy4n |
| Vault Admin | 58dpPSAM3PuzziKgRkFGTFM3xY4fH4xFgzDBe3nnF6gg |
| Vault Manager | 9Rx3i7GyVFFUYGKCMDZQpqnoJgSzo3R1qwmsAN5aiiSu |
| Drift Strategy PDA | FCxpLFtjChHjJ86V9mYFTN2QGXHPxR5LjLPGVGBA68rJ |
| Vault Drift User | HURzSVDvBBA9VEZ1j1SQhn4iHUhxua3ZNhFejGgfpuq8 |
| Voltr Program | vVoLTRjQmtFpiYoegx285Ze4gsLJ8ZxgFKVcuvmG1a8 |
| Drift Adaptor | EBN93eXs5fHGBABuajQqdsKRkCgaqtJa8vEFD6vKXiP |
| Drift Program | dRiftyHA39MWEi3m9aunc5MzRF1JYuBsbn6VPcn33UH |
User deposits USDC --> Voltr Vault (BFDTTG8n...)
|
+-- manager-deposit-strategy --> Drift Adaptor CPI
| |
| +-- Vault Drift User (HURzSV...)
| |
+-- Keeper (EC2, 24/7) |
| |
+-- Delegate trading authority --+
+-- Delta-neutral: spot buy + perp short
+-- Dynamic tilt (0-10%)
+-- Signal detection (5 min)
+-- Health monitoring (30 sec)
+-- DN rebalance (4 hours)
The keeper operates as a delegate on the vault's Drift user account. Capital flows through the Voltr adaptor (deposit/withdraw), while trading uses Drift SDK's placeSpotOrder and placePerpOrder for DN execution. All trades and PnL are on-chain and verifiable through the vault's Drift account.
# 1. Clone and install
git clone https://github.com/psyto/yogi.git
cd yogi
npm install
# 2. Configure
cp .env.example .env
# Edit .env with RPC URL, keypair paths, vault address
# 3. Initialize vault (one-time)
npm run admin:init-vault # Creates Voltr vault, outputs VAULT_ADDRESS
npm run admin:add-adaptor # Registers Drift adaptor
npx tsx src/scripts/manager-init-strategy.ts # Initializes Drift user for vault
# 4. Deposit USDC
npx tsx src/scripts/user-deposit-vault.ts # User deposits to vault
npx tsx src/scripts/manager-deposit-strategy.ts # Manager moves funds to Drift
# 5. Run keeper
npm run keeper # Local
# Or with pm2 for production:
pm2 start 'npx tsx src/keeper/index.ts' --name yogi-keeper
pm2 save && pm2 startup- Vault infrastructure: Voltr / Ranger Earn — deposits, LP shares, fee collection
- Trading: Drift Protocol v2 — spot + perpetual futures execution via delegate model
- Keeper: TypeScript bot on AWS EC2 with pm2 process management
- Signal detection: 5-dimension anomaly detector + cross-venue (Binance/Bybit) + CEX OI tracking
- Vol computation: Parkinson estimator on SOL-PERP hourly candles
- Data feed: Drift Data API — funding rates, market stats, OHLC candles
- RPC: Helius (websocket subscription mode)
Yogi's intelligence layer is available as a paid API. Traders and vault builders can consume regime decisions without running the full keeper.
curl -H "Authorization: Bearer fbrk_your_key" https://yogi-api.fabrknt.com/v1/regimeReturns: { deployment_pct, max_leverage, rebalance_mode, signal_severity, vol_regime }
| Endpoint | What |
|---|---|
/v1/regime |
Deployment %, leverage, mode — the main decision |
/v1/signals |
5D anomaly detection state |
/v1/funding |
Market rankings by funding rate |
/v1/cross-venue |
Drift vs Binance/Bybit comparison |
/v1/imbalances |
Market imbalance scores |
/v1/health |
Health check (no auth) |
An open-source keeper template consumes this API — users run it with their own wallets. Intelligence from Fabrknt, execution under your control.
Get an API key at flow.fabrknt.com/dashboard ($99/mo Pro tier).
Yogi is one piece of a larger build toward Syntx — a composable perp DEX on Solana. What Yogi proves (regime engine, anomaly detection, multi-asset DN) becomes Syntx's built-in vault intelligence.
| Product | Role | Status |
|---|---|---|
| Yogi | Proves DN + regime engine on Drift | Mainnet, keeper live |
| Kodiak | Proves 5-layer yield on Hyperliquid | Mainnet, keeper live |
| Flow | Capital flow intelligence | Live at flow.fabrknt.com |
| PerpU | Cross-protocol education | Live, 182+ lessons |
| Syntx | Composable perp DEX (endgame) | Development |
29 @fabrknt/* infrastructure packages (margin, fees, privacy, security, state) + 5 Percolator matchers form Syntx's technology stack. The vaults prove the intelligence works before it ships as a platform.
Yogi pairs with Kodiak (Hyperliquid) to form a diversified DN funding portfolio. They harvest funding from uncorrelated sources:
| Yogi (Drift) | Kodiak (Hyperliquid) | |
|---|---|---|
| DN assets | SOL, BTC, ETH, POPCAT, DRIFT (5 markets) | HYPE |
| Funding driver | Broad crypto + Solana ecosystem conviction | Hyperliquid ecosystem conviction |
| Funding profile | Mixed: blue chips (volatile) + Solana natives (stable 80-100% APY) | Stable, moderate (5-11% APY) |
| Strength | Multi-asset diversification + Solana long tail | Structural long bias = persistent funding |
Why they complement:
- BTC/SOL/ETH funding is cyclical — spikes in bull markets, can flip negative in bear.
- POPCAT/DRIFT funding is structurally positive — community conviction keeps holders long, similar to HYPE on Hyperliquid. When blue-chip funding dips, Solana-native assets often still pay.
- HYPE funding is structurally positive — HYPE holders are ecosystem believers who stay long. Kodiak earns steadily regardless of broader market.
- When one source dips, the other typically holds. Natural diversification without correlation engineering.
Combined portfolio estimate:
| Condition | Yogi (60%) | Kodiak (40%) | Blended APY |
|---|---|---|---|
| Bull | ~20% | ~8% | ~15% |
| Neutral | ~12% | ~6% | ~10% |
| Bear stress | ~5% | ~5% | ~5% |
Both vaults share the same strategy brain (regime engine, signal detection, dynamic tilt) but operate on different chains, different assets, and different funding dynamics.
Built for the Ranger Build-A-Bear Hackathon (Mar 9 – Apr 6, 2026).
- Track: Main + Drift Side Track
- Base asset: USDC
- Lock period: 3-month rolling
- Vault on-chain:
BFDTTG8nJF7uLf3wsqFJpYCvC6wA6BahBRKjTgtKPy4n
1. We proved Drift is for institutional yield, not just HFT — we built the same DN strategy on Drift and Hyperliquid. Drift's composability (spot + perp + lending, no bridges) makes multi-asset DN easy. This is the story the Drift ecosystem needs to attract institutional capital.
2. Delta-neutral eliminates price risk — the hackathon is called "Build-A-Bear." Yogi doesn't just survive bear markets — it's structurally immune to price direction. Spot + perp cancel out. Yield comes from funding, not price bets.
3. Dynamic tilt is novel technology — no other Drift vault adjusts its hedge ratio in real time based on signals. Pure DN in stress, tilted for extra yield when safe.
4. Multi-asset DN on Drift — SOL, BTC, ETH, POPCAT, and DRIFT simultaneously. Five DN-eligible markets from Drift's long tail of 20+ spot+perp pairs. The keeper dynamically rotates to the highest-funding markets. No other submission uses Drift's spot markets for DN.
5. DN on DRIFT token — Yogi runs delta-neutral on Drift's own token. When DRIFT holders are long-biased, the perp short collects funding (~106% APY). This is only possible on Drift and demonstrates the protocol's unique composability.
6. 5D cross-venue intelligence — comparing Drift funding against Binance and Bybit in real time. No other Drift vault does this.
7. Live mainnet performance, not backtests — deployed March 20, running continuously. On-chain verifiable.
8. Ecosystem contribution — Yogi isn't just a vault. PerpU is a learning platform that teaches why Drift is the right platform for institutional yield. Vault + education = ecosystem growth.
9. Production-ready — Voltr integration, delegate trading, position loading on restart, 30-second health monitoring, 62 unit tests with regression coverage. Ready for $500K seeding on day one.
MIT