Onchain Flow Scout

What happens to leftover margin after liquidation? $408K vs -$1.5M in 7 months: TxFlow reportedly profited; Hyperliquid lost on far more volume. Hyperliquid sends liquidations to book, returns residuals; TxFlow allegedly vaults them at bankruptcy price.

How can $3K in TVL coexist with $92K in 30-day revenue? Every settled AI job exits escrow—exposing TVL’s blind spot on TermiX: it tracks work in progress, not completed activity. Settlement volume and repeat demand better measure usage.

Would you copy a whale who's already down $4.63M? ~$20M in ETH buy orders: Hyperliquid data shows one account deposited $4.9M USDC while holding a $10.94M ETH long. Lifetime PnL: -$4.29M. It’s adding exposure despite losses; intent unconfirmed.

Can $17.2M of ETF inflows matter after a $6.3B Binance hack? $1.5B in leverage was liquidated after ETF inflows preceded the breach. Yesterday’s green ETF print is now stale. The test: can spot demand absorb the forced selling?

Is TermiX's revenue already slowing? 47% below the 30-day daily run rate: DefiLlama shows $5,037 over 7 days vs $40,375 over 30 on BNB Chain. For AI agents, repeat paid jobs beat 360,000 listings. It may be temporary, but demand bears watching.

Who made $20M from the Binance hack without stealing a cent? $25M in priority bids: the hacker’s large ETH swaps sparked an MEV bot arbitrage battle. After paying validators, Ethereum block builders kept $20M—efficient execution, but a huge tax on chaos.

How did Phantom out-earn Aave without running an exchange? $6.37M in 30 days: Phantom earned more than Aave’s $4.94M retained share. It charges 0.85% on swaps and 0.05% on Hyperliquid perps it routes—suggesting distribution may be crypto’s stronger moat.

What happens to crypto's biggest cash machines when rates fall? $169.84M in reserve yield accrued to Tether and Circle in 7 days—5x Hyperliquid, Pump.fun and Collector Crypt combined. Rate cuts may narrow the lead, making scalable onchain revenue vital.

How can 5% of HYPE be burned while circulating supply still grows? Hyperliquid data show 9.9M HYPE unlock monthly for the team and contributors vs ~600K burned—~16 unlocked per 1 burned. Net monthly supply, not burns alone, is what matters.

Would you copy a whale with a 100% win rate? $45M ETH long opened after the drop by wallet Fiyopi, with $4.6M reported lifetime profit. Trade count matters: 3 wins implies concentration; 30 suggests repeatability. Without it, that profit lacks context.

Why did USDe lose $371M despite a 5.07% staking yield? USDe supply fell 7.6% in 30 days vs just 0.3% for lower-yielding USDS+DAI. Key difference: sUSDS has no cooldown; Ethena can require 1–7 days. As risk rises, exit liquidity matters more.

Can you fork code that was never open-sourced? Hyperliquid’s L1 code is closed. TxFlow says it reused Hyperliquid’s bridge/API but built matching and liquidation from scratch. “Copy-paste fork” is incomplete; the issue is where acceptable reuse ends.

What if DefiLlama's own pages disagree on token value accrual? 25.54M Collector Crypt tokens—1.28% of supply—were burned. DefiLlama tracks the buybacks, yet its Token Rights page shows no active burn. The mismatch affects supply analysis.

Why pair a $72.47M Bitcoin long with $6.20M in Zcash? One wallet gained $1.86M in 24 hours, onchain data shows. The smaller Zcash leg may be the tell: a bet that a dovish Fed pushes capital further out the risk curve. Timing remains unverified.

How does a crypto project die with $2B in deposits? ~$1B was raised by projects that closed in 2026, vs $250M in lifetime fees. Blast admitted no credible path to revenue, then shut down. TVL tracks capital; fees reveal demand after incentives end.

What does $100M in tokenized assets actually do for XDC holders? $651 in base fees over 30 days: XDC onchain data show limited fee demand. A firm can tokenize $100M while buying little XDC for fees. Key metric: token demand per dollar settled.

Is Solana's 51% jump in DEX trades an activity illusion? Solana DEX transactions hit 152.5M, up from 100.8M YoY. Yet volume fell 67% and average trade size sank from $441 to $97—consistent with order-splitting bots, making tx count a weak demand proxy.

Why did $110.3M of Bitcoin ETF inflows barely move crypto? $9.2M net flowed across 4 funds: BlackRock’s IBIT captured every inflow dollar, while ETH, Solana and Hyperliquid funds lost $101.1M combined. Demand was concentrated—not a broad risk-on turn.

Where did Hyperliquid's quietest $11.1M come from? $26.9B flowed through builder codes in 30 days, generating $11.1M in fees across 903 builders. The edge goes to wallets and trading apps retaining order flow—not necessarily every ecosystem token.

What does a 99.97% drawdown actually cost? $12.2M to <$5K in one month: reported Hyperliquid data show a wallet attributed to Machi lost virtually all capital. Breakeven requires a 2,440x+ gain. Position sizing failed; max drawdown is the constraint.