$869k already deposited into @Infinit_Labs apex vault 2. deposits close nov 11. what stuck with me from vault 1 wasn't the return. it was how much of it you can check yourself. the headline is +8.59% gross in 90 days, ~39% annualized. that's the number marketing leads with. the more useful question is how $506,142 turned into $43,464 of trading profit: + 1 venue: @HyperliquidX. part of the book ran on @tradexyz, but those are hip-3 markets on hyperliquid + 346 orders, $12.31m volume ➜ capital turned over 24.3x, ~$35.6k per order + ~0.35% profit per $1 traded + crypto was the largest exposure and the largest profit source + dollar-neutral book: longs and shorts sized to offset, so returns came from picking markets, not market direction + max drawdown: 3.7% of capital (vs -10.5% in the backtest) net to depositors after fees: 6.9%. anyone can trace this on the shared trading wallet, or pull it with a few lines of code from the hyperliquid api. ➜ a vault you can audit beats a vault that only shows you an apy.

name one narrative that keeps growing no matter what crypto price does. i'll go first: market data. @PythNetwork closed october at $13.67m arr, up 18.9% from $11.49m in september. +82% since july. 44.6% of gross new arr came from existing accounts expanding. ➜ customers who already pay are paying more the product carrying it is pyth indices: + $2.88m ARR across 21 clients, ~21% of total + added $1.12m in october, about half of all new arr + catalog went from 50 to 77 stable feeds in one month most of pyth's 3,927 feeds aren't crypto. they are: equities, commodities, and fx. and buyers come through the front door: + 4,451 terminal sign-ups in october, +37% mom + self-serve checkout brought in $725k, 29% of gross new arr what's next: + 31 more indices feeds queued + corporate-action data for equities + korean post-market feeds live (samsung, sk hynix, hyundai) + an ai-agent server, so agents can buy data directly since oct 8, 100% of the dao's share of product revenue goes to pyth buybacks. ➜ prices swing. pyth's revenue has climbed every month since july.

trading cards are moving onchain, and the numbers back it up. + global card market: ~75b in 2026, upfrom ~2b in 2016 (TCGplayer) + web3 alone now does ~$300m in monthly card sales (Surgence) ➜ collectibles aren’t a niche anymore. this is a real market with real volume. hat’s the setup for @collector1agent, an ai agent built to snipe and trade cards. how it works: + scans the market in real time for collectible cards like Pokémon + funds itself from its own trading fees + the agent owns the cards, while $COLLECTOR holders own the token behind it; it's also one of the first agent launches on @districtxyz new launchpad $COLLECTOR holders earn a share of the vault and get a say in how the agent hunts it’s also the first successful launch on @districtxyz, solana. ➜ in short: collector finds cards, decides to hold or flip, and acts on its own reasoning. take a look: https://t.co/82Z0JJxqeL

apex vault 1 just matured. vault 2 deposits open today. quick recap before you roll over. APEX multi-strategy, run by Apex Quantitative Capital on @Infinit_Labs, closed its first 90-day period (jul 30 to oct 28) at +6.9%. same period: + vault 1: +6.9% + s&p 500: +4.9% + gold: +0.35% (mostly unchanged) vs. on-chain yield in the same tvl range (~$500m): + tokemak arbUSD: 9.56% APY (30d avg, ~half paid in TOKE) ➜ ~2.3% over 90 days, less if you held since jul 30 + gauntlet gtUSDCp: 6.02% APY ➜ ~1.4% over 90 days + neverland nWETH: 5.99% APY ➜ ~1.4% over 90 days ➜ roughly 3–5x what passive stable yield paid over the same window the trade-off is that this is an active long/short strategy, not a lending pool. returns aren’t fixed, and the 6.9% came in under the manager’s own target range. rollover is not automatic: + redeem from vault 1 + deposit into vault 2 before nov 11 i’ll be tracking whether vault 2 can beat round one.

$aero dropped 24% since the offical aero launch, from $0.89 to $0.68. what does the drop tell us about what’s next? fyi, @aeroxyz officially launched their new dex site on oct 21 with a new amm, velo to aero conversion, and new networks added. on that day, aero hit $0.89, the highest mark of the week. the 2026 high was $0.99, 10 days before the launch. meanwhile, btc fell only 3.5% over the same week. aero fell ~5x more. ➜ the dip looks like supply/migration risk, not fundamentals. + max supply of aero rose from 2.01b to 2.127b (+5.8%) + $saero positions are now transferable + custodial conversion from coinbase is still ahead on nov 2 no on-chain proof of dumping yet. the market is pricing in dilution before it happens. meanwhile, the product is working. slipstream did $442.9m of aero’s $581m daily volume, suggesting the new build is operating well. what i’m watching: + nov 2 coinbase conversion and exchange inflows after it + post-launch fees and revenue on aero + aero’s share on base, currently 54.4%, holding above 50% through next quarter base is still aero’s largest market. it’s also the biggest risk: base is exploring its own token and already has a dex aggregator inside its app. ➜ the merger hedges that. aero is no longer only a base bet.

no quantum computer today can break crypto. but the cost of building one that could keeps dropping. in march, google cut its estimate of the resources needed to break bitcoin and ethereum signatures. by september, an open challenge had cut that number by more than half again, and 7 researchers from @Starknet’s side were co-authors. so starknet isn’t waiting. they launched a quantum hub that lays out its research, architecture and migration path. the core idea is crypto agility: when a primitive breaks, how fast can a network swap it out? starknet’s answer has 02 parts: + stark proofs are hash-based, so the proof system itself doesn’t need replacing + native account abstraction keeps signature logic inside each account, so wallets can move to post-quantum signatures without a network-wide fork the remaining work is a 3-phase roadmap to remove leftover elliptic-curve dependencies (blake2 replacing pedersen hashing, falcon-512 for consensus signatures). nobody knows when quantum hardware arrives . but replacing deployed cryptography takes years. ➜ starknet isn’t betting on a date. it’s building so the swap is cheap whenever the date comes.

ethereum's glamsterdam is still on testnet. @solana already shipped its next upgrade. alpenglow hit testnet on sep 24 and is now live on mainnet. solana was built for high-performance trading, but even the fastest chain needs a rework under the hood. what changed: + towerBFT and proof of history timing are out, replaced by votor (voting) and rotor (block propagation) + finality drops from 12.8s to ~150ms + validator votes move off-chain, so reported tps and transaction counts will fall + 20+20 resilience model: the network stays safe with up to 20% malicious stake, plus another 20% offline normal users won't notice much yet. the effect shows up over a longer window: fewer vote transactions means more blockspace for actual trading activity. block count and block time on solana have already shifted since activation. ➜ lower tps on dashboards isn't a downgrade. it's the noise getting stripped out.

rwa perps open interest just hit $5.64b. that's 42.7% of all defi perps oi ($13.2b), on a product category that barely existed a year ago. the trigger was hip-3, which went live on hyperliquid on oct 13, 2025. since then, rwa perps went from near zero to almost half the market. defillama data shows the growth is real, but heavily concentrated: + @tradexyz (the frontier of hip-3) - $3.82b (~68% of rwa oi) + @variational_io - $824.5m + @QFEX - $268.2m + @gmtrade_xyz - $176.19m + @Lighter_xyz - $115.2m + @OndoPerps - $94.56m + @entropyIO - $74.27m + @arcus_xyz - $73.48m + @extendedapp - $43.42m + @risextrade - $37.38m top 10 hold ~$5.53b, or 98% of the segment. one venue holds two thirds of it. the part that stands out: equities are the largest share of oi, ahead of gold and oil. that's odd, because spot tokenized stocks never got much traction with crypto natives. so who's driving this? two likely answers: + crypto traders who want leveraged equity exposure without leaving onchain rails + tradfi traders who want 24/7 access when stock markets are closed ➜ either way, the demand was never for owning tokenized stocks. it was for trading them with leverage, around the clock.

10 days live and @papertrade_xyz is already top 5 by 30d revenue. it's a perp dex on hyperevm that launched oct 10. why it's pulling volume this fast: + distribution: it's built on hyperliquid, where the traders already are. no need to build a user base from zero + leverage: up to 1000x on $BTC and $ETH. for comparison, HL itself caps BTC at 20-40x + execution: there's no order book. every trade is a synthetic swap against a USDC pool, priced at HL's best bid/offer midpoint ➜ zero slippage, no spread, no funding + fees: only winners pay. opening, holding, or losing a position costs nothing + token: $PAPER starts at zero supply, with no team, VC or airdrop allocation. it's only minted when a trader closes at a loss or gets liquidated + staking: staked PAPER earns USDC from winners' fees plus a share of pool revenue. once the pool passes $5m, every extra dollar lost goes to stakers the core design: + winners take profit and pay the fees + losers get PAPER, stake it, and earn from the next round of losers ➜ losing becomes a position instead of a dead end. that's the hook imo. the staked ratio is 99.2% and staker rewards have passed $54.87m. one caveat: PAPER transfers are disabled for now, so staking is basically the only thing you can do with it. the ratio shows the loop works, not that holders have conviction. risks worth knowing: + at 1000x, a 0.1% move against you wipes the position + pricing off HL's bid/offer midpoint is a known manipulation vector + if the pool runs short, winners may have to wait to withdraw + staker yield only holds up if traders keep losing i'm not a trader, this is a read on the design. DYOR before you touch it.

defi trends you may have missed this week: 1. 1000x perps on @papertrade_xyz fyi, Papertrade is the latest perps platform in the frontline on HyperEVM, being the first place to offer 1000x leverage compared to the normal 20-40x on HL. do you really need 1000x? note that a move of only 0.1% in price can wipe you out completely. unlike others, Papertrade has no order book; it routes losses into an LP to cover the winning positions. also, traders who lost may claim $PAPER as future earnings. ➜ now it starts to feel that Papertrade already knew that the amount of losses will exceed the winning trades. some may wonder what if losses won't be able to cover the expenses; current metrics will prove: + lp balance: $4.96m + open interest: $657.4m who loves risk will still love risk. 2. pair tokens on @ClutchMarkets there are originally 4,444 Stonkbrokers NFTs that pay stocks to the holders. the model is pretty simple as selling NFTs until they want to start st fresher - pair of stock tokens. for example: NVDA + TLSA = NVDATLSA. holding NVDATLSA means a share of this LP on Uniswap v4. thus, the real value of holding this pair token lies in how much volume created via the LP. ➜ this model is not new to defi, it's new to tokenized stocks. Rather than putting stocks as collateral, contributing to a LP is preferred on StonkBrokers. this is a wrap-up for your next days on defi.