In 2024 an Arup employee wired $25.6M after a video call full of deepfakes. Now World ID runs inside Zoom. AI made faces cheap to copy. So apps now ask: is a real person behind this screen, wallet or agent? Each team asks for one piece of you, then proves only the fact an app needs: > Your eyes: @worldnetwork verifies people with its Orb. World ID now covers Zoom calls, banking and AI agents, with Pantera and VanEck as design partners. > Your passport chip: @selfxyz reads the chip in a passport or national ID and proves age or nationality with zero-knowledge proofs. Self says it serves 20M+ users. > Your records: @Humanityprot proves facts like age, income or identity while the data stays with you. Its site counts 8M+ Human IDs. @zkPass turns data from any website into a proof. > Your wallet history: @HumnPassport, formerly Gitcoin Passport, gives each wallet a humanity score. Airdrops use it to filter bots. It lists 2M+ users. > Your agents: @billions_ntwk ties each AI agent to a verified human. It counts 2.5M+ verified users and 25K+ verified agents. One important distinction: proof of human ≠ proof of identity. The strongest designs tell an app "real and unique" and keep your name private. My filter for any proof-of-human project: what it collects, where that data lives, and which apps already call the check. Proof of human is becoming a normal login step for calls, banks and agents. The networks already inside those apps will set the standard 🫡

A memecoin now holds 30.5% of all tokenized AMC on @RobinhoodApp Chain. The frenzy behind it hit $440M a day, then fell 96%. Each buy of a meme paired with a Stock Token pulls that stock into the meme's pool. Traders moved on, and the pools still hold the stock. How to read the bars: each █ is 5% of a stock token's onchain supply, held in its biggest meme pool (Oct 19). MEME/AMC 30.5% ██████░░░░ BONER/HIMS 20.7% ████░░░░░░ AI/NVDA 6.6% █░░░░░░░░░ INU/AAPL 5.1% █░░░░░░░░░ Three launchpads mint these pairs: - @longdotxyz: tokens on top of Stock Tokens, 28.6K launches and $909M traded in 30 days - @ponsdotfamily: "home of memestocks", 155.7K launches and $3.1B traded - @AlfaClubdotapp: social launchpad since Sep 28, keyholders split 50% of fees The pool sets its own price: tokenized HIMS hit 4x the stock in August Open launches cut both ways too. Sleuth Wazz links one crew to 53 launches and $18.43M extracted (his claim). Memes rent the stock as a trading pair today. Tomorrow the deepest meme pools become the stock's main onchain market

Arbitrum launched a growth program for USDG. The first 10 teams holding $10M of USDG each earn 200K ARB. The bar is a 30-day average, and Arbitrum has yet to name accepted teams. So I ranked the apps holding USDG on Arbitrum by distance to $10M, using each project's own latest numbers (Oct 18). Who looks close depends on whose counter you read. How to read the bars: each █ is $1M of USDG, and 10 full blocks reach the $10M bar. Over the line @gauntlet_xyz USDG Premium: $27.9M deposited, per Morpho ██████████ @GMX_IO: $12.6M USDG per Arbitrum's dashboard, $11.4M of it in the Dollar Vault per GMX ██████████ @SteakhouseFi USDG Pro: $10.2M deposited, per Morpho ██████████ Those vault figures count deposits. Arbitrum's USDG dashboard counts the dollars that stay put, and there Morpho holds $4.2M. The gap is a loan book $33.8M of the vault money is lent out, per Morpho. The borrowers are holders of @maplefinance syrupUSDG, with $38.0M posted as collateral. Still climbing Outside lending, the pools sit further back, per Arbitrum's dashboard: @0xfluid USDG/USDC pool: $2.4M ██░░░░░░░░ @Uniswap pools: $1.3M █░░░░░░░░░ @mayan and @RIFTHQ route money into USDG, so balances pass straight through ARB rewards on both Morpho vaults end Oct 20, per the DRIP app. The first 30-day average will separate borrowed demand from reward farming. The 2M ARB goes to the teams whose dollars stay after the rewards stop.

Earlier this week crypto bought the licenses. Now the card networks and core banks are wiring stablecoins into their own back office. Visa, Mastercard, Marqeta, Cross River, Jack Henry and DTCC all moved inside 48 hours. So I sorted all seven by how real they are today. LIVE @Visa Direct: USDC prefunding for cross-border payouts, via Inyo Global. Reach: 180+ countries. @Mastercard x @BVNKFinance: accounts, payouts and cards in one integration. Reach: virtual cards in 130+ countries. ROLLING OUT @StripeCrypto: stablecoin card issuing expands to 65+ countries. @crossriverbank: @Ripple RLUSD on XRPL inside its core banking OS, limited states for now. RLUSD size: $2.5B+. PLANNED Marqeta x @coinbase: stablecoin-backed cards settled on @base. Scale: ~$400B processed in 2025. Jack Henry x @googlecloud: tokenized deposits on Universal Ledger. Reach: 7,200+ banks and credit unions. @The_DTCC: 25 securities at launch on Canton and Besu. Eligible pool: ~$30T, launch due end of October. My 3-word filter for any TradFi stablecoin headline: Layer: which part moves onchain. Owner: who keeps the customer. Status: live, rolling out or planned. Crypto spent a decade asking users to switch apps. The next wave of stablecoin users arrives inside the apps they already pay with Earlier this week crypto bought the licenses. Now the card networks and core banks are wiring stablecoins into their own back office. Visa, Mastercard, Marqeta, Cross River, Jack Henry and DTCC all moved inside 48 hours. Every deal follows one mechanism. The customer stays with the bank or the card brand, and only the plumbing underneath changes: funding, settlement and card spend start running on stablecoins around the clock. So I sorted all seven by how real they are today. LIVE @Visa Direct: USDC prefunding for cross-border payouts, via Inyo Global. Reach: 180+ countries. @Mastercard x @BVNKFinance: accounts, payouts and cards in one integration. Reach: virtual cards in 130+ countries. ROLLING OUT @StripeCrypto: stablecoin card issuing expands to 65+ countries. @crossriverbank: @Ripple RLUSD on XRPL inside its core banking OS, limited states for now. RLUSD size: $2.5B+. PLANNED Marqeta x @coinbase: stablecoin-backed cards settled on @base. Scale: ~$400B processed in 2025. Jack Henry x @googlecloud: tokenized deposits on Universal Ledger. Reach: 7,200+ banks and credit unions. @The_DTCC: 25 securities at launch on Canton and Besu. Eligible pool: ~$30T, launch due end of October. My 3-word filter for any TradFi stablecoin headline: Layer: which part moves onchain. Owner: who keeps the customer. Status: live, rolling out or planned. Crypto spent a decade asking users to switch apps. The next wave of stablecoin users arrives inside the apps they already pay with 💳

Sui wrote a $30M check for AI data. The money landed on provenance. Sui Foundation invested $30M into @datafdn as DATA Network prepares to move its protocol onto @SuiNetwork. 250M+ AI data contributions have already been registered through Trace. Now that provenance layer is moving into a stack Sui has already been building around AI: - Provenance: @datafdn tracks where training data came from and the rights attached to it. - Storage: @WalrusProtocol stores datasets, models and inference logs. - Access: Seal controls who can decrypt and use the data. - Compute: Nautilus runs sensitive workloads inside TEEs with results verifiable on Sui. - Training: @flock_io coordinates privacy-preserving model training. - Agents: @Talus_Labs brings autonomous agents onchain. - Inference: @Atoma_Network provides private AI compute and inference. - Payments: @humntech gives agents programmable wallets, while Mysten Labs is part of Google AP2’s agent-payment ecosystem. The $30M deal makes more sense when you look at the stack. Sui already had storage, access control, compute and agent rails. DATA brings something upstream of all of them: verified training data with provenance. That may be the scarcer part of the AI stack.

Stablecoins went onchain to be seen. The next wave wants them unseen. Three teams are shipping it on three different chains. > @utexocom: USDT on Bitcoin via RGB + Lightning, private transfers, coming to Gate and GateDEX (12/10) > @0xMiden: testnet era done, more "in a few weeks" (12/10). ZK execution with private state > @raycashxyz: a private bank app on @zama FHE, 70K waitlist, points live since 07/10 Privacy is the next stablecoin UX upgrade.

Quantum fear finally has a price chart. The newest PQ coins outran the big chains this month. The CoinGecko quantum-resistant basket, 30d (13/10/2026): Old guard, built PQ from day one @QRLedger +72% @tidecoin +66% @cellframenet +68% @QANplatform -3% New guard, launched with PQ as the pitch @QuantusNetwork +544% @NeptuneCash +213% @Qubic +46% Abelian abelian:native +25% Big chains adding PQ later @Starknet +215% @AlgoFoundation +29% @NervosNetwork +33% @Zcash +9% Upgrades on the roadmap @ethereum: leanXMSS research @SuiNetwork: ML-DSA accounts, Q1 2027 target @projecteleven: migration tooling, $20M Series A (Jan 2026) Size tells the real story. $ZEC sits near $20B. base:0x37696cd8f723ce541ac376f79fc109080e505ba3 sits near $93M. $QTC sits near $35M. The narrative runs on small caps. The users still live on the giants. The winners turn that feature into daily usage first 🫡

GPUs from 2020 are booked through 2029. Compute is becoming a financial asset. The capital markets are forming now One old chip tells the story. @CoreWeave booked A100s until 2029 and borrowed $2.6B against those contracts. Old GPUs now work like real estate: rentable, financeable, bankable. Demand keeps rising Agents burn 5 - 50x more tokens than chatbots The pain is price chaos. The same GPU hour costs $2 in one market and $15 in another. @shayonsengupta at @multicoin's fix fits in one line: Turn compute into a receipt you can trade, pledge and borrow against. Crypto fits that last step: > Collateral anyone can audit > Receipts that plug into lending > Stablecoin capital, open 24/7 Teams building pieces of it: GPU-backed credit @USDai_Official: stablecoin loans backed by GPUs @gaib_ai: financing for AI compute @falconfinance: tokenized GPU forward in structuring Tokenized compute @gpunet: stablecoins into physical GPUs, quarterly yield @AskVenice: DIEM, inference credits as a token Compute markets @OrnnExchange: compute price index from live trades @ComputeDesk: market rails for compute traders @Lighter_xyz: early compute futures Verified inference @prlnet: inference that doubles as mining @ambient_xyz: PoW L1 serving verified inference GPU marketplaces @akashnet: decentralized GPU marketplace, $AKT @ionet: decentralized GPUs for AI workloads @AethirCloud: decentralized cloud compute @nosana_ai: instant GPU rental, solana:nosXBVoaCTtYdLvKY6Csb4AC8JCdQKKAaWYtx2ZMoo7 Oil got futures. Power got forwards. Compute is next. The first standard GPU receipt will become the benchmark for the AI economy. Whoever issues it sets the price of intelligence.

Mining gets a second job. @prlnet is building a Proof-of-Useful-Work (PoUW) model that lets miners secure the blockchain while producing useful AI compute. The idea is to generate two revenue streams from the same compute. Pearl uses matrix multiplication, a core operation behind AI training and inference. Miners can: > Run useful computations for AI customers > Convert the results into valid PoUW > Earn $PRL mining rewards The interesting part is the economies of scale. Pearl's 1+o(1) overhead design means the relative cost of converting useful computation into PoUW approaches zero as workloads grow. That could turn mining pools into dual-purpose infrastructure: Network security: Earn $PRL from mining AI compute: Generate revenue from inference workloads And potentially create a reinforcing economic loop: More miners strengthen network security and expand compute capacity. More inference demand could drive greater $PRL payment activity through Pearl's planned marketplace. Stronger miner economics could then attract additional hashrate. PoW security meets AI compute economics. The real test is whether useful compute demand can sustain that flywheel.

Physical AI needs a real-time view of the world. Position, vision, machine state, airspace and weather are becoming part of the data stack. That’s where these DePIN networks fit. 1/ Positioning @GEODNET provides centimeter-level GNSS/RTK corrections for drones, robotics, surveying and autonomous machinery. @onocoyRTK builds a decentralized GNSS correction network across 180 countries. 2/ Vision & mapping @NATIXNetwork turns distributed cameras into driving and street-level data for mapping, autonomy and world models. @Hivemapper collects fresh street imagery and converts it into map data like road signs, speed limits and traffic infrastructure. 3/ Machine state @DIMO_Network connects vehicle telemetry across dozens of car brands. Location, battery, mileage and diagnostics become programmable data. 4/ Airspace @Wingbits tracks aircraft through a decentralized ADS-B sensor network. Its data supports aviation analytics, flight tracking and machine-readable airspace intelligence. 5/ Environment @WeatherXM uses distributed weather stations to provide hyperlocal ground-truth weather data. That data can feed agriculture, logistics, energy and autonomous systems. This is the data layer for Physical AI. And these networks are building it from the ground up.

The next AI trade may be financializing compute. AI created the demand for GPUs Now the next layer is forming around how compute gets financed, tokenized, priced, traded, and used as collateral. A full compute capital market could look like: - GPU hardware - compute cash flow - tokenized claims - credit - spot / forward markets - derivatives Crypto is already building across that stack. @gaib_ai focuses on GPU financing + tokenization, bringing AI infrastructure cash flows onchain through products like AID and sAID. @USDai_Official is building the credit layer, where AI companies borrow against GPU infrastructure and that yield flows back into DeFi. @Compute_Labs tokenizes physical GPUs into onchain assets, opening the door to fractional ownership, yield, lending, and structured products. @InferraTrade is pushing toward the market layer with GPU-hour pricing, compute indices, and derivatives around compute demand. The broader thesis: - DePIN built the supply - tokenization makes compute transferable - credit turns compute into collateral - markets create price discovery - derivatives let users hedge capacity and funding risk Compute capital markets could be the next layer of AI x Crypto.

➠ @ether_fi is expanding its business model. The team is launching its own stablecoin through Ethena Whitelabel. EtherFi already has: > $300M+ in stablecoin deposits across the platform > nearly $1B in cumulative Cash card spend > 114K+ active cards They are going all-in on the full-stack neobank x DeFi model for users. Instead of relying only on third-party stablecoins, EtherFi can now issue EtherFi USD, while @ethena handles reserve management, mint/redemption, and compliance infrastructure behind it. EtherFi owns the distribution and user relationship, while Ethena handles more of the backend. EtherFi now joins: - @MapleFinance - @JupiterExchange - @megaeth - @SuiNetwork in building stablecoin products on Ethena’s infrastructure. This could become another meaningful revenue driver for EtherFi as it captures more economics from the stablecoin balances already sitting across its platform. Privacy is also an important moat now I’m especially looking forward to seeing a neobank/card product integrate privacy directly into the payment stack. Share your thoughts

Something’s moving at Turtle 👀 @turtledotxyz has been expanding beyond DeFi liquidity into a broader distribution layer for onchain capital. More integrations. More distribution. More ways for capital to move onchain. Now 7.10.2026 is here. Curious to see if today’s reveal is the next piece of that buildout 🐢

Ethereum is pushing privacy into the agent payment layer. @ethereum introduced zkAPI, a new way to pay for APIs using private usage credits. Users deposit ETH into a vault, generate a ZK proof for each session, and receive a short-lived API key with a defined spending limit. The AI provider handles the prompt and inference. zkAPI verifies the payment right. Ethereum handles settlement. That keeps billing identity and API usage in separate layers. I think this opens a much broader research direction around the Private Agent Economy. x402 already made API payments programmable for machines. Now projects are adding privacy to different parts of that stack. > @NethermindEth is exploring PrivateX402 for private payment channels between agents. > @zama provides confidential money through encrypted token balances and transfers on Ethereum. > @aztecnetwork is building programmable private execution, giving apps a private state and execution layer. Aztec Network > @TACEO_IO is also working on confidential x402 through its private execution infrastructure. The keywords I'm researching next: - Private API Payments - Unlinkable Usage Credits - Private x402 - Agentic Payments - Machine-to-Machine Payments - Confidential Compute/Defi - Private AI Inference Preparing $ETH super cycle

. @monad hit $1B TVL Now the real question is whether that capital stays. Q3 was the breakout quarter. TVL climbed from ~$358M in July to ~$1.02B by the end of September, driven mainly by lending, curated yield and stablecoin liquidity. @aave was the biggest catalyst, while @eulerfinance and @Morpho helped deepen the lending base. At the same time, @pendle_fi, syrupUSDC, sUSDe and curated strategies from K3 Capital and Hyperithm gave capital more places to earn. That pushed Monad into a much stronger position: ~$706M stablecoin supply ~$430M active RWA AUM ~$1B+ DeFi TVL But hitting $1B is only the first half of the story. A lot of that capital arrived through incentives, yield opportunities and new deployments. The harder test is whether Monad can turn it into sticky, productive liquidity once incentives cool down. That's why I'm watching Q4 around: Aave V4, tokenized equities, more RWA collateral, payment rails, institutional privacy and stronger organic borrow/trading demand. With a major $MON unlock also approaching in November, capital quality matters more than another TVL headline. Monad already proved it can attract liquidity. Now it has to prove the liquidity has a reason to stay

Confidential DeFi is gaining traction. $197.6M TVL, up 170% over 18 months, with FHE now capturing 41.7% of the sector (excl. mixers). A recent @DefiLlama report highlights how @zama is turning privacy into a broader financial stack. Study $ZAMA and start building your watchlists The product layer is expanding Lending: 16 confidential vaults on @Morpho, curated by Steakhouse, Flowdesk, Armitage, RockawayX and Bitwise. Trading: Zama Swap enables private swaps between confidential assets. RWAs: Apex Group announced a $100B tokenization initiative using Zama as its confidentiality layer (target, not deployed TVL). Infrastructure: ERC-7984, developed with @OpenZeppelin, aims to standardize confidential tokens. And early traction is showing: $82.7M TVL | $58.5M Confidential Vault TVL | $532.3M Cumulative Shielded Volume The next milestone I'm watching is whether confidential assets can move beyond Morpho into broader lending, trading and payment applications. My watchlist in this narrative ZK (Zero-Knowledge): Zcash, @RAILGUN_Project, @0xprivacypools TEE: @OasisProtocol Sapphire, @SecretNetwork MPC: @Arcium Private Chains: @CantonNetwork FHE: @zama Privacy is starting to look more like a financial primitive than a standalone DeFi niche.

DePIN is generating real revenue, and $GRASS is a good example. Multicoin announced its investment in Grass, and you can do too Impressive growth here > 2025 revenue: $17M > H1 2026 revenue: $17M > 2026 revenue guidance: $75M (training data alone) > 6M+ network contributors With AI demand expanding from training data to real-time inference, DePIN deserves another look. Some tokens to keep on the radar: AI Compute: $TAO solana:rndrizKT3MK1iimdxRdWabcF7Zg7AR5T4nud4EkHBof $ATH $AKT solana:BZLbGTNCSFfoth2GYDtwr7e4imWzpR5jqcUuGEwr646K Data & Bandwidth: $GRASS $TRAC theta-token:native Storage: filecoin:native $AR $AIOZ Wireless & Connectivity: solana:hntyVP6YFm1Hg25TN9WGLqM12b8TQmcknKrdu1oxWux solana:7JA5eZdCzztSfQbJvS8aVVxMFfd81Rs9VvwnocV1mKHu The next DePIN narrative could be driven by commercial demand rather than token incentives alone.

5 projects unlocking Bitcoin's next DeFi layer. From native BTC lending to programmable Bitcoin, each is building a different piece of BTCFi. A new wave of BTCFi builders is approaching the problem from different directions. Here's how @Stacks, @AlpenLabs, @ZestProtocol, @Arch_Prime, and @citrea_xyz are putting BTC to work. 1. @Stacks | Making Bitcoin Programmable Stacks is an established smart-contract layer connected to Bitcoin, with sBTC serving as the bridge between native BTC and DeFi applications. Flagship: sBTC, a 1:1 BTC-backed asset using a dynamic threshold-signer model. Solution: Gives Bitcoin holders access to lending, trading, and yield applications through programmable assets on Stacks. Funding: Blockstack raised $23M through regulated token offerings in 2019. Current phase: Production. Nakamoto and sBTC are live, with an established DeFi ecosystem and liquid $STX token. 2. @AlpenLabs | Building Bitcoin-native Financial Infrastructure Alpen combines EVM-compatible execution with Strata, an orchestration layer connecting the network to Bitcoin. Flagship: Alpen Rollup + Strata Bridge, powered by the Glock/Mosaic verification architecture. Solution: Targets a 1-of-N security model for BTC deposits and withdrawals, allowing applications to build Bitcoin-native lending, trading, and stablecoin markets. Funding: $8.5M strategic round co-led by DBA and Cyber Fund. Current phase: Public Testnet III, launched in July 2026 with 21+ launch partners. USDC and CCTP integration was announced in September. 3. @ZestProtocol | Turning Native BTC into Productive Collateral Zest is expanding from its established lending market on Stacks into native Bitcoin collateral infrastructure. Flagship products: - Stacks Market V2: BTC lending and borrowing. - Bitcoin Collateral Vaults: Borrow stablecoins on EVM chains against BTC held on Bitcoin L1. Solution: BTC stays in a self-custodial Bitcoin vault while users access liquidity on another chain, avoiding the need to wrap or bridge their collateral. Funding: $3.5M Seed led by Draper Associates, with Binance Labs (now YZi Labs) and Flow Traders participating. Current phase: Market V2 is live, with $100M+ in historical peak deposits reported by the team. Bitcoin Collateral Vaults entered a capped mainnet demo on September 23, allowing native BTC collateral to back real USDC loans on Ethereum. 4. @Arch_Prime | Bridgeless Bitcoin Programmability Arch takes a different architectural route from conventional Bitcoin rollups. Flagship: ArchVM, an eBPF-based execution environment combined with FROST/ROAST threshold signatures. Solution: Enables programmable financial applications around Bitcoin UTXOs, aiming to keep settlement tied to native Bitcoin transactions without relying on a conventional wrapped-asset bridge. Funding: $20M disclosed across two rounds: - $7M Seed led by Multicoin Capital. - $13M Series A led by Pantera Capital. Current phase: Arch previously reported 50M+ testnet transactions. Comparable public production TVL remains unverified. 5. @citrea_xyz | Bitcoin as DA + Settlement Citrea uses ZK-rollup technology to expand Bitcoin's financial utility, with Bitcoin serving as both its data availability and settlement layer. Flagship: Citrea zkEVM + Clementine Bridge + ctUSD. Solution: Brings EVM-compatible lending, trading, and USD liquidity into a Bitcoin-oriented execution environment, using a BitVM-based bridge for BTC interoperability. Funding: $16.7M total: - $2.7M Seed led by Galaxy Ventures. - $14M Series A led by Founders Fund. Current phase: Mainnet launched in January 2026 alongside ctUSD, issued by MoonPay and powered by M0. Its base:0x11030f79109269d796fd0fb956d6244e502757f7 token subsequently launched in May. Different approaches, same goal: making BTC more productive. The next thing to watch is real BTC utilization, lending demand and production adoption.

RWA perps are scaling fast. @DefiLlama × @RWAFoundation_ now track 1,086 active markets across 19 venues, with: - $117.3B 30D volume - $5.15B average OI Key venue data: - @tradexyz $54.8B, 46.7% market share - @Variational_io $26.9B, including $14.2B from its new Nasdaq-100 perp ( TGE coming) - @gmtrade_xyz $13.8B, up 28.9% MoM - @HyperliquidX HIP-3 remains a major infrastructure layer behind RWA perp deployment. Asset mix is expanding beyond single stocks: Public equities $38.5B Equity indices $31.4B Precious metals $19.0B Oil $14.9B FX $8.9B RWA perps are moving from single-stock trading into indices, commodities, FX and broader macro exposure. RWA trading onchain is becoming a full derivatives market