BITCOIN: THE SETUP BEHIND THE CURRENT RECOVERY $BTC has recovered from the mid-$70Ks to around $84K–$85K, supported by ETF demand, short covering and institutional buying. Here’s what matters: ✧ $14.9B BTC options expire today Deribit shows $14.87B in BTC options expiring, including $8.33B calls and $6.54B puts. This is contract value, not $14.9B of BTC being bought or sold. The expiry resets a large amount of derivatives positioning, making the positioning that follows important. ✧ Spot demand has returned U.S. spot Bitcoin ETFs recorded: Sept. 21: +$999M Sept. 22: +$714.7M Sept. 23: +$346.9M That’s roughly $2.06B of net inflows in three sessions, reversing the heavy outflows seen earlier in September. ✧ Institutional accumulation continues Strategy bought another 950 BTC for ~$75.7M, taking its holdings to roughly 846,000 BTC. ✧ Short covering helped the move BTC’s move above $86K coincided with roughly $919M in crypto short liquidations, including about $557M in BTC shorts. Part of the rally therefore came from forced buying, rather than entirely from fresh long-term demand. ✧ Miner economics are improving JPMorgan estimates BTC’s production cost around $85K. Holding above that level could reduce selling pressure from higher-cost miners, although $85K is not a fixed price floor. ✧ Mining is competing with AI Bitcoin hashrate remains about 19% below last October’s peak, while some miners are shifting power and infrastructure toward AI computing. ✧ Technical structure $84.1K → support $85.3K → first resistance $86.1K–$86.3K → next resistance $82K → deeper support BTC remains above the 200-day moving average, with the 20-day EMA rising. AFTER TODAY’S OPTIONS EXPIRY This is the part I’m mainly watching for positioning. The $84K–$86K range becomes important once the current derivatives positioning clears. A clean hold above $84K keeps the recovery structure intact. A sustained move through $86K would show stronger acceptance at higher prices. Losing $84K puts $82K back into focus. ETF flows and new derivatives positioning will help separate genuine spot demand from another leverage-driven move. For now, I’d rather watch how BTC behaves around these levels than chase the move in the middle of the range.

USDT on @BNBCHAIN is now the most-held stablecoin across all chains.

Crypto hacks in 2026 have now passed $2.3B As of September 25, around $2.34B has been involved in 86 major crypto security incidents. That’s the gross figure. Some of the money was recovered, frozen, or returned. THE BIGGEST HITS January — individual wallet theft, ~$282M Attackers posed as hardware-wallet support and convinced victims to give up their recovery phrases. April 1 — @DriftProtocol, ~$285M Attackers gained control of critical security systems, manipulated collateral values, and drained the protocol. April 18 — @KelpDAO, ~$292M Attackers compromised cross-chain infrastructure, forged messages, minted unbacked rsETH, and converted it into real assets. July — @COLDCARDwallet, ~$114M A weakness in recovery-phrase generation allowed attackers to reconstruct private keys from affected wallets. September 6 — @Liquid_BTC, ~$319M Attackers exploited Elements’ validation process to create unbacked L-BTC and convert it into BTC. About 85% was later returned, with roughly $47M remaining according to @trmlabs. September 24 — @bitget, ~$351.6M affected Bitget detected unauthorized transfers from part of its hot and warm wallet infrastructure. Bitget said cold wallets remained secure and its $464M+ protection fund covered the affected amount. The exchange later said a compromised backend wallet-service system generated false transfer information that reached the signing process. Bitget CEO Gracy Chen also said some IPs matched VPN infrastructure previously linked to a North Korean hacking group. Bitget suspects Lazarus, but the attribution has not been confirmed. These six incidents account for roughly $1.64B. The other 80 make up the remaining ~$700M. HOW THEY GOT IN The weak point was not always the smart contract. Private keys → governance → bridges → wallets → infrastructure → exchanges. Attackers went after whatever had enough access to approve, validate, create, or sign transactions. That includes: → Private keys → Multisig signers → Price feeds → Bridge verifiers → RPC systems → Wallet software → Backend services A protocol can have audited contracts and still lose money because something around those contracts was compromised. Bitget is a good example. The transfers happened on-chain, but the reported problem was higher up the chain of operations, inside a backend system connected to the wallet-signing process. HOW TO LIMIT THE DAMAGE Keep less money in hot wallets. Separate transaction creation from approval. Require another check for large or unusual transfers. Set limits on amounts, frequency, and destinations. Pause activity that suddenly looks abnormal. Avoid single points of failure in bridges and critical security systems. Monitor the systems behind the wallets, not just the blockchain. WHERE SECURITY NEEDS TO IMPROVE The $2.34B figure needs context. Stolen, recovered, frozen, returned, and permanently lost funds are different numbers. But 2026 has made one thing clear: securing the smart contract is not enough. Keys, wallets, people, backends, bridges, and signing systems can all become the point of failure. One compromised part should not be enough to move hundreds of millions. Reduce access. Reduce exposure. Limit the damage.

I kept seeing $ZEC on my timeline, so I finally stopped scrolling and looked into it properly. At first, I thought it was just another coin getting a lot of attention. Then I checked the numbers. Price: ~$1,485 Market cap: ~$25.2B FDV: ~$25.2B Circulating: ~16.95M ZEC Max supply: 21M 24H volume: ~$1.5B ZEC was around $400 in July. That’s a roughly 3.5x move, and I wanted to understand what was actually behind it. I found out that there’s more going on than people simply trading the privacy narrative. Shielded activity has picked up significantly Zcash recorded 62K+ shielded transactions last week, its highest weekly figure since 2022. Roughly 4.91M ZEC is also sitting in shielded pools. The money coming into the asset caught my attention too. @Grayscale ZCSH ETF has brought in $306M+ in net inflows since launching in August and has grown to around $977M in assets. @21Shares has also launched a physically backed Zcash product in Europe. The network itself is moving forward as well. The @Zcash NU7 upgrade is currently targeting November 5 for mainnet. Block times are expected to drop from 75 seconds to 25 seconds, alongside other protocol changes. Then there’s @paradigm, which has disclosed that it owns ZEC and has invested in Zcash development. That got a lot of attention, but I don’t think it explains the whole move. ZEC had already run hard before the disclosure. After going through everything, what stood out to me was how many things are lining up at once. Price is moving. Shielded usage is rising. ETF demand is coming in. The protocol is being upgraded. Institutional access is expanding. That’s what makes me bullish on it. I’m still not treating the current price like an easy entry, though. At ~$25.2B market cap, ZEC has already gone through a massive repricing. From here, I want to see whether the ETF keeps absorbing supply, shielded activity continues to grow and NU7 delivers. If those things continue, I’ll look for an entry. If they don’t, I won’t force it just because my timeline is full of $ZEC posts. For now, I’m keeping it firmly on my watchlist. This is my opinion based on the data I’ve reviewed, not financial advice.

AI COMPUTE IS BECOMING A FINANCIAL MARKET The AI race is moving beyond models. Underneath every model, agent and inference application sits the same physical bottleneck: AI compute. GPUs need power. Power needs data centers. Data centers need capital. As AI usage grows, that infrastructure is becoming a market of its own. The first shift is already happening. GPU capacity is being measured. Compute prices are being benchmarked. Hardware is being financed. Future capacity is being priced. CME is preparing H100 and B200 compute futures based on GPU rental benchmarks. ICE and NATIVX are developing energy-adjusted compute futures across workloads including training and inference. Ornn is building around compute pricing, spot markets and derivatives. The important change is simple: AI compute is moving from something companies simply buy into something the market can price and hedge. The financing layer is growing with it. NVIDIA has partnered with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR on platforms designed to mobilize more than $500B in third-party capital for AI infrastructure over time. Crypto is building its own financial rails around the same infrastructure. ✧ @akashnet turns distributed GPU capacity into an open marketplace. ✧ @bittensor uses token incentives to coordinate specialized compute, inference and digital work. ✧ @USDai_Official brings stablecoin liquidity and GPU-backed lending into the infrastructure layer. ✧ Venice uses DIEM to turn recurring inference access into a transferable onchain asset. ✧ x402 enables software and AI agents to pay for digital services programmatically. These are not the same products. They are different pieces of an emerging AI compute economy. And compute has an important difference from most financial assets: unused capacity cannot be saved. If a GPU is available for one hour and nobody uses it, that hour is gone. It cannot be carried forward and sold as another GPU-hour. That makes utilization, pricing and financing central to the business. Hardware depreciates. Electricity costs move. New chips arrive. Demand shifts between training and inference. Capital therefore needs better ways to price future compute, finance hardware and manage the risk around changing costs. That creates a natural progression: GPU supply → compute pricing → credit → futures → collateral → tokenization → liquidity This is where crypto’s role becomes much larger than decentralized cloud infrastructure. DePIN can coordinate the supply. Stablecoins can move the capital. DeFi can finance the infrastructure. Tokenization can package future access. Derivatives can manage price risk. Onchain settlement can connect the market. The endgame is not simply putting GPUs onchain. It is making AI compute a programmable financial resource: measured → priced → financed → hedged → tokenized → traded → settled AI creates the demand. GPUs provide the capacity. Capital finances the infrastructure. And crypto is beginning to build the financial market around the compute powering AI.

TOKENIZED STOCKS ARE BECOMING A NEW MARKET LAYER Tokenized equities are moving past the simple idea of putting stocks onchain. The market is starting to form around three layers: STRUCTURE → DISTRIBUTION → UTILIZATION STRUCTURE A token that tracks a stock is not automatically the same as a tokenized security carrying the rights of that stock. That distinction matters as equities move further onchain. On Sept. 17, the @SECGov introduced a temporary five-year Innovation Exemption for qualifying Tokenized Securities Venues to facilitate trading of tokenized NMS stocks through permissioned AMMs and liquidity pools. Qualifying tokenized NMS stock must represent the same interest and rights in the underlying security, including dividends and voting rights. This gives qualifying venues a regulatory path for secondary trading of certain tokenized stocks onchain. DISTRIBUTION The tracked tokenized stocks and ETFs market reached about $3.18B as of Sept. 23, across 19 chains, 3,530 assets and 5,912 token products. Most of the value sits on a few networks: ✧ @BNBCHAIN → $1.13B ✧ @ethereum → $833.2M ✧ @solana → $742.9M ✧ @arbitrum → $183.1M ✧ @RobinhoodApp Chain → $177.9M ✧ @XLayerOfficial → $173.3M ✧ Others → $67.3M BNB, Ethereum and Solana hold roughly 82% of tracked value. The market is already multi-chain, but supply is still concentrated across a few major distribution networks. ARC IS PART OF THE INFRASTRUCTURE Arc is not a major tokenized-stock market today. The interesting part is what is being built around it. @circle launched Arc mainnet on Sept. 16. @DinariGlobal brought its dShares tokenized U.S. stocks and ETFs to Arc, while Circle is working with @The_DTCC on bringing DTC-custodied assets onto the network from the second half of 2027. That puts Arc closer to the settlement and financial-infrastructure side of the market than the current tokenized-equity market. UTILIZATION Once an equity is tokenized, it can move through trading, liquidity pools, lending markets, collateral, structured products and DeFi. That creates a much bigger use case than simply holding a stock token. But the infrastructure still has to handle dividends, corporate actions, voting, redemptions, custody, trading halts, oracles and legal ownership. Two tokens tracking the same company can therefore have very different mechanics. Same stock. Different issuer, custody, claim and redemption path. THE TOKENIZED STOCKS ECOSYSTEM IS FORMING ✧ @OndoFinance → tokenized securities and distribution ✧ @xStocksFi → multi-chain equity distribution ✧ @binance → exchange-native distribution ✧ @RobinhoodApp → brokerage + blockchain distribution ✧ @DinariGlobal → tokenized U.S. equities and ETFs ✧ @arc → financial-market infrastructure and settlement ✧ @Uniswap → onchain liquidity ✧ @Morpho → credit infrastructure ✧ @kamino → lending and liquidity ✧ @JupiterExchange → Solana trading and liquidity The interesting part is no longer how many stock tickers can be tokenized. It is what gets built around them. Nasdaq is already moving in the same direction. The exchange agreed to invest $100M in @Payward, @krakenfx’s parent company, while working on Nasdaq Equity Tokens, with a planned Q2 2027 launch subject to regulatory approval. Crypto is bringing equities onchain. Traditional finance is starting to build around the same infrastructure. markets → wallets → liquidity → lending → collateral → settlement The token is the representation. The infrastructure is where the market gets built. If you find this useful, leave a like and share your thoughts below. See you on the next one.

10 CRYPTO INFRASTRUCTURE PROJECTS I’M WATCHING BANK DEPOSITS ✧ Cari Network Tokenized bank deposits → $32.5M funded entirely by banks. 30+ institutions involved. Pilot mint → transfer → burn is live with design-partner banks. Building regulated infrastructure for bank deposits to become programmable onchain money. CLEARING ✧ @GlacisLabs — ZeroDelta Institutional clearing → $6.8M seed. $1B+ lifetime volume and $1.5B annualized run rate reported by the company across 40+ chains. Matches opposing flows, nets them and settles the residual instead of moving the full amount across chains. STABLECOIN PAYMENTS ✧ @rtp — Latitude Stablecoin → local payouts → $35M Series A, $43M total. Stablecoins settle into local bank accounts and mobile wallets without requiring recipients to hold crypto. Building the last-mile layer between stablecoin liquidity and local payment rails. ✧ Fin.com Stablecoins + local rails → $20M seed. Infrastructure spanning local payment rails, USD accounts, SWIFT and stablecoins. Focused on moving stablecoin liquidity through fragmented payment corridors. ONCHAIN FX ✧ @KiiChainio Stablecoin FX + settlement → Live 24/7 FX and cross-border settlement focused on emerging-market currencies. Combining stablecoin liquidity, FX and settlement in one onchain stack. PRIVATE CREDIT ✧ @tarecredit — Tare Onchain private credit → $13.25M seed led by Blockchain Capital. Founding team from BlockTower Credit, JPMorgan Kinexys and Centrifuge. Building infrastructure for private-credit origination, servicing and distribution. INSTITUTIONAL NETTING ✧ @cyclesmoney Obligation netting → $8.7M raised. Multilateral netting without a CCP, collateral or asset movement on Cycles Prime. Institutional product is in beta. Reducing the capital and liquidity required for institutional settlement. PAYMENT COORDINATION ✧ @atumhq — Atum Open payment network → $13.5M from Variant, PayPal Ventures and others. No token, no chain, no custody. Building coordination infrastructure between stablecoins, chains, cards and payment providers. RWA PERPS ✧ @Ostium Onchain RWA derivatives → Live perps on gold, FX, indices and equities with public market activity. Bringing traditional market exposure onchain through derivatives before full asset tokenization. THE STACK Bank money → Stablecoins → Payments → FX → Clearing → Credit → RWA markets Research watchlist. Not financial advice. Funding and volume figures are largely company-reported. Early-stage projects carry execution, regulatory, liquidity, security and adoption risk, always DYOR. NFA

MIDWEEK ALPHA · v.04 BUILD A BETTER RESEARCH STACK One of the biggest upgrades you can make as a crypto researcher is changing where you get your information. I used to spend too much time consuming what CT had already interpreted. Now I try to get closer to the evidence. Here are some of the tools and research sources I’d keep close. PRIMARY SOURCES Protocol docs. Governance forums. GitHub. Token contracts. Official announcements. Audits. Research papers. This is where you get closest to what a project actually built, changed or reported. Everything else should help you verify it. @Dune — dune.com Useful for onchain research. You can query blockchain data, inspect dashboards and queries, and build your own analysis with DuneSQL. That lets you go beyond someone else’s chart and investigate the data yourself. @DefiLlama — defillama.com A broad source for DeFi data across protocols and chains, including TVL, fees, revenue, DEX volume and stablecoins. The key is knowing what each metric actually measures. A number is evidence, not automatically a conclusion. @tokenterminal — tokenterminal.com Useful for analysing blockchains and applications through financial and usage metrics. Fees, revenue, earnings and users can help you look beyond headline growth and examine the economics behind it. @nansen_ai — nansen.ai Useful when wallet activity matters. Wallet labels, Smart Money signals and wallet activity can help you investigate onchain participants and capital movements. Then there’s the research layer: @MessariCrypto — messari.io @Delphi_Digital — delphidigital.io @galaxyhq — galaxy.com Don’t read these firms just to borrow their conclusions. Study how they build a thesis, choose evidence, compare markets and identify risks. The same applies to the researchers you follow. Look for people who show their sources, explain their reasoning, use data and change their views when the evidence changes. Study the work, not the reputation. Over time, your process should look something like: Primary sources → onchain data → market & financial data → independent research → your own analysis. When two sources disagree, check the definitions, timeframe, dataset and methodology before deciding which one makes sense. Sometimes that disagreement is where the useful research starts. You don’t need every tool in crypto. You need to know where to look, how to verify what you find, and when the evidence is strong enough to support a conclusion. I’d genuinely like to see people try this. Don’t just bookmark the tools. Use them. Pick a project, go back to the primary sources, pull the data yourself, test the narrative and show what you find. I’d like to see the results. If you found this useful, show some love with a like, drop your thoughts in the comments, and follow for the next research piece. See you on the next one.

MIDWEEK ALPHA · v.03 READ THE DATA, NOT THE NARRATIVE One of the easiest ways to get crypto research wrong is to find a true number and build the wrong conclusion around it. TVL is up 80%. Volume doubled. Revenue hit $2M. Users are growing. The numbers can all be real. The mistake is assuming they automatically mean what the narrative says they mean. ➤ UNDERSTAND THE METRIC TVL measures the value of assets held in a protocol’s smart contracts. But dollar-denominated TVL can rise because the assets themselves appreciated, even if no new capital entered. So higher TVL ≠ automatically more capital. Fees and revenue are different too. Fees are what users pay. Revenue is the portion the protocol retains. Mixing those numbers can completely change how you understand a protocol’s economics. ➤ UNDERSTAND THE METHODOLOGY An onchain metric isn’t simply pulled from the blockchain and dropped onto a dashboard. There are definitions, contract mappings, events, pricing, aggregation and methodology behind the final number. Before using a metric, understand where it came from and how it was calculated. ➤ PUT THE NUMBER IN CONTEXT “Revenue grew 300%” sounds impressive. But 300% growth from $10K is very different from 300% growth from $10M. The starting point, timeframe and comparison matter. A percentage without its baseline is incomplete information. ➤ LOOK BEYOND THE NUMBER Higher TVL doesn’t explain why it increased. Higher volume doesn’t prove the activity is sustainable. More fees don’t mean the protocol keeps all of them. Strong protocol revenue doesn’t automatically make the token undervalued. One metric can tell you something important without telling you everything. I’ve had to become more careful about this in my own research too. It’s easy to find a number that fits your thesis and build the entire narrative around it. The better habit is to slow down and ask what the number actually proves. What caused the change. What is it being compared with. How was it calculated. What does it leave out. And what conclusion can it actually support? That is where research gets better. You stop collecting numbers because they look impressive and start using them because they mean something. Read the data. Understand the metric. Check the methodology. Put it in context. Then build the narrative. Not the other way around. If you found this useful, like and share your thought on the comment, do well to follow for the next piece. See you on the next one.

MIDWEEK ALPHA · v.02 HOW TO BECOME A BETTER RESEARCHER A lot of people on CT don’t have a research problem. They have a discipline problem. You see a project you like, read a thread about it, find another account saying the same thing, save a few charts, and suddenly you feel like you’ve done your homework. You haven’t. You’ve just seen the same story from a few different angles. I’ve been working on this in my own research too, and I can already see the difference. I question claims more. I go back to the source more. I spend less time asking, “What is everyone saying?” and more time asking, “What can I actually verify?” I still catch myself slipping sometimes. That’s exactly why I’m writing this. If you want to become a better researcher, start by becoming harder to convince. Someone makes a claim about a protocol. Don’t believe it because the account has 200k followers. Don’t believe it because the chart looks clean. Don’t believe it because three respected researchers posted it. Instead, make these part of your normal research process: ➤ FIND THE SOURCE Don’t stop at someone else’s explanation. Go to the announcement, documentation, dataset or original report and see what was actually said. ➤ CHECK THE CONTEXT Look at the date, methodology and what the number actually measures. A number can be completely real and still be used to tell you the wrong story. ➤ VERIFY IT INDEPENDENTLY Five accounts repeating the same statistic are not five confirmations. If they all got it from the same source, you still have one piece of evidence. ➤ TRY TO PROVE YOURSELF WRONG Think a protocol is growing? Go find the numbers that could challenge that conclusion. Think a token is undervalued? Find the strongest argument for why the market might be right. Love the project? Good. Now look for the reasons you might be wrong. This is where research can go bad. Once you get attached to a thesis, you can stop investigating and start defending it. You become the lawyer, the judge and the person deciding what evidence gets admitted. I’ve caught myself doing this too. So before I get too comfortable with a thesis, I ask: “What would make me change my mind?” Then I go looking for it. And don’t blur the line between what happened and what you think it means. ➤ SEPARATE FACTS FROM INTERPRETATION “Protocol generated $2M in fees” is something you can verify. “Therefore the token is undervalued” is an interpretation. Know where the evidence ends and your conclusion begins. Sometimes, after doing all of this, you end up with: “I don’t know yet.” That’s fine. You don’t need to force an answer just because you’ve already spent three hours researching. None of this guarantees that you’ll become a great researcher. There’s no magic formula. But these habits make it harder to fool yourself, easier to catch weak assumptions, and much easier to tell the difference between something you know and something you simply believe. That’s what I’m working on too. So when you research something, don’t focus on how much information you collected. Focus on whether you actually tested your own belief. Find the source. Check the context. Verify it independently. Try to prove yourself wrong. Separate facts from interpretation. And if the evidence gives you a different answer than the one you wanted, take the different answer. If you found this useful, show some love with a like, drop your thoughts in the comments, and follow for the next research piece. See you on the next one.

MIDWEEK ALPHA · v.01 REPETITION IS NOT RESEARCH Let’s be honest. A lot of what we call “research” on CT is just reading until we feel confident. You see a claim, believe it, then go looking for five more posts saying the same thing. You find them, and suddenly you feel like you’ve verified something. You haven’t. You just found five people agreeing with the first person you believed. One person says a protocol’s revenue is growing. Another repeats it. Someone makes a thread. Another adds a chart. A bigger account quotes it. Now the claim is everywhere. You think, “Everyone is saying it, so it must be true.” No. You might have just watched one claim travel through twenty accounts. That isn’t twenty confirmations. It’s one claim being repeated. And yes, I’ve caught myself doing this too. Read enough posts that agree with you and you start feeling like a researcher when, really, you’ve just gotten very good at scrolling in one direction. 😂 It gets serious when money enters the picture. You buy based on a claim you never checked. You overlook the red flags. You get attached to the thesis. Then someone challenges you and, instead of checking the evidence, you send them another tweet. That isn’t research. You’re defending a belief you never properly investigated. So change the habit. If a claim can influence your money, leave CT and find the source. Check the data, the date and how the number was produced. Understand what it actually measures and what it leaves out. Then do what most people avoid: Try to prove yourself wrong. Look for the strongest argument against your thesis. Check independent sources. Follow the evidence instead of collecting posts that agree with you. If the thesis survives, good. If it falls apart, good. You found out before your money did. Reading 30 tweets isn’t research. Following 50 analysts isn’t research. Knowing every narrative on CT isn’t research. Those things can give you a lead. They cannot verify the claim for you. Use CT to find ideas. Then leave CT and do the work. Because if your only reason for believing something is: “I’ve seen it everywhere.” You haven’t necessarily found more evidence. You may have just heard the same thing enough times to stop questioning it. From now on, don’t publish something just because you saw it everywhere. Trace it. Check it. Challenge it. Then publish it. If the evidence changes your mind, change your mind. Being wrong during research is fine. Being confidently wrong because you never checked is not. Use CT to find the lead and Do the research yourself. If you found this useful, like and drop your thoughts in the comments. See you on the next one.

Crypto Market Review: 2026 — What Actually Moved the Market 2026 has looked like two very different markets. Crypto moved from roughly $3.2T early in the year → $2.4T after Q1 → $2.1T after Q2 → ~$3.0T now. From the June low, total market cap has recovered roughly 43%. The chart shows what happened. The real story is why it happened. ↳ Q1 — Higher rates hit crypto Total crypto market cap fell 20.4% / $622B in Q1, ending at $2.4T. Most of the drop came between mid-January and early February, around Kevin Warsh’s nomination as Fed chair, as markets prepared for a potentially tougher rate environment. Higher rate expectations → less appetite for risky assets → crypto falls. BTC fell 22% during the quarter. ↳ Trading slowed down Top-10 spot CEX volume fell 39.1% to $2.7T. March was even weaker, with just $0.8T in volume. When fewer people are buying and selling, large orders can move prices more easily. Leverage then made the fall faster: prices fall → traders get liquidated → more selling → prices fall further. ↳ Q2 — The market hit its low Another $304.8B left the market in Q2, taking total crypto market cap from $2.4T to $2.1T. June saw the biggest drop, with ETF outflows, a hawkish Fed, changing U.S.–Iran tensions and Strategy’s BTC sale adding pressure. Spot CEX volume fell another 27.9% to $1.95T. Perp volume also fell, but by less: 10% to $12.7T. ↳ Stablecoins stayed relatively strong Stablecoin market cap ended Q2 at $305.1B, down only 1.6%. That was a much smaller decline than the wider crypto market. So while investors moved away from riskier coins → a large amount of dollar liquidity stayed inside crypto. ↳ Q3 — Money started coming back BTC led the recovery. ETF demand improved, companies continued buying BTC, and traders betting against the market were forced to close positions as prices rose. By September 21, BTC had climbed above $86K. ↳ ETFs became an important source of demand August saw roughly $3.52B of U.S. spot BTC ETF inflows, the strongest monthly figure of 2026. Another ~$987M followed the next week. More ETF inflows → more BTC demand. ETF flows became an important signal for how much new money was moving into BTC. ↳ Rates were still high The Fed raised rates another 25 bps to 3.75%–4.00% in September. Its 2026 median PCE inflation forecast was 3.7%. So the recovery happened even though money was still relatively expensive. That made actual buying demand more important. ↳ Regulation also affected the market The Senate failed to advance the CLARITY Act on September 15. That added uncertainty around U.S. crypto regulation, while regulators continued working on digital-asset rules elsewhere. ↳ Hacks continued taking money out CertiK recorded more than $1.3B lost across 344 security incidents in H1 2026. Wallet compromises alone accounted for more than $444M. The KelpDAO RPC compromise and Drift Protocol breach made up roughly 44% of H1 losses. Market losses change prices → security losses remove actual capital. ↳ RWA brought new markets onto crypto Crypto platforms continued adding: Equities → Commodities → Indices → FX This gave traders more ways to use crypto infrastructure beyond trading crypto tokens. ↳ Prediction markets added another source of activity Prediction-market volume reached $113.8B in Q2, up 48.7% from Q1. June alone reached $52.8B. Crypto activity is no longer just about buying and selling tokens. »»»» 2026 has been a cycle of money leaving, then coming back. Higher rates, weaker trading activity and leverage pushed the market down to $2.1T. ETF demand, BTC buying and short covering then helped push it back toward $3T. At the same time, stablecoins, RWA, prediction markets, security events and regulation kept changing where money moved across crypto. That’s the 2026 crypto market so far.

RWA perpetuals are now a 24/7 derivatives market for traditional assets. Stocks, ETFs, indices, commodities, FX and even pre-IPO names can trade as perps without owning the underlying asset. USDC/USDT provides margin, while oracle infrastructure keeps the perp tied to the underlying price. $3.16T tracked volume Dec. 29, 2025 → Aug. 31, 2026 August alone → $799.5B, a new monthly record. Pyth’s August dataset: ↳ Equities — $487.3B | 64.8% ↳ Commodities — $152.3B | 20.3% ↳ Indices — $103.9B | 13.8% ↳ FX — $8.3B | 1.1% Stocks went from 9% of January volume → 62.3% in August, while commodities fell from 81% → 18.8%. Venue concentration ▸ @binance — $1.592T | 50.4% ▸ @HyperliquidX — $542.8B | 17.2% ▸ @okx — $345.1B | 10.9% ▸ @Bitget — $238.2B | 7.5% ≈86% of tracked volume sits across these four venues. The remaining ~14% is spread across: ▸ @Gate — $148.5B | 4.7% ▸ @Bybit_Official — $105.2B | 3.3% ▸ @edgeX_exchange — $46.3B | 1.5% ▸ @Lighter_xyz — $33.4B | 1.1% ▸ @OstiumLabs — $23.9B | 0.8% ▸ @ExtendedExchange — $19.6B | 0.6% These figures come from CMC’s 19-venue dataset through Aug. 31. For traders, venue quality comes down to OI, depth, funding, execution and liquidation liquidity. Onchain RWA perps DeFiLlama currently tracks roughly $5.21B OI, $6.17B 24h volume and 1,022 markets across its RWA perp coverage, excluding CEXs. @tradexyz currently leads the onchain snapshot: $3.85B OI | 89 markets | $2.42B 24h volume Other active venues include Variational, QFEX, GMTrade, Lighter, Ondo, @OstiumLabs and Extended. Infrastructure is expanding @Deribit offers RWA perps across equities, ETFs, commodities and pre-IPO contracts, settled in USDC. RWAperp launched on @okx X Layer with SK Hynix, Samsung Electronics, SanDisk, Micron, Intel and a DRAM index, settled in USDG. Hyperliquid HIP-3 lets independent builders launch perp markets while using Hyperliquid’s matching, margin and liquidation infrastructure. HIP-3 volume grew from $12.65B in Q4 2025 → $130.87B in Q1 2026. The xyz deployer accounted for roughly 99.6–99.8% of August HIP-3 volume in CMC’s dataset. Oracle risk is now a core market risk RWA perps keep trading while the underlying market is closed. That creates exposure to pre-market prints, earnings, halts, corporate actions and weekend price gaps. Pyth says its feeds priced 96.27% of the volume in its tracked August RWA perp book. The remaining 3.73% was attributed to other providers or remained unverified within that dataset. That figure is Pyth’s share of its tracked book, not 96.27% of the entire RWA perp market. The SK Hynix incident showed the consequence. A faulty pre-market print triggered roughly $57M in liquidations across 960 accounts. The RWA perp stack: Oracle → Liquidity → Margin → Funding → OI limits → Liquidation The volume is already here. Now the infrastructure has to handle 24/7 trading, market closures, corporate actions and volatility without breaking.

THE TWO CRYPTO INFRASTRUCTURE SHUTTING DOWN Two crypto infrastructure projects announced shutdowns within a day of each other. @linera_io stopped operations on September 18 after it couldn’t secure enough funding to keep working toward mainnet. Linera had raised around $12M across two funding rounds. It was building a Layer 1 around microchains, giving applications their own execution space instead of having everything compete for the same blockspace. Its final $LNRA community sale needed at least $1.5M USDC. It raised $848,271 from 617 participants. That wasn’t enough, so the contributions were refunded. The team then tried to secure emergency funding to keep development going, but that also fell through. Linera eventually shut down before launching its mainnet. @SwitchboardFDN is a different case. It is ending its oracle operations on September 25, affecting 550+ feeds across 53 Solana programs. Among the affected protocols are @Kamino_Finance, @jito_sol, @marginfi and @DriftProtocol. Switchboard provides price data that DeFi protocols use to value collateral and manage lending positions. Take MarginFi. It checks how fresh its oracle prices are, and those prices can be needed when users borrow, withdraw or get liquidated. Now those protocols have to move their feeds to other providers. Switchboard has pointed affected users toward alternatives including @PythNetwork and @redstone_defi. The two shutdowns are happening for very different reasons. Linera couldn’t raise enough money to reach mainnet. Switchboard is stepping away while its infrastructure is already being used by live protocols. That creates two very different problems. Linera needed more capital to keep building. Switchboard’s users now need to migrate critical feeds and make sure borrowing and liquidations still work once those feeds are gone. And that’s what stands out to me. In crypto, building the technology is only part of the job. It also has to survive when funding changes or an important piece of infrastructure disappears.

Aave V4 on Arc: the real test isn’t TVL @aave V4 is already at $556M TVL, with about $127M on Arc. That’s roughly 23% of V4’s TVL, just days after launch. That’s a strong start. But TVL isn’t what I’m watching most closely. @arc has generated just $1.41 in fees so far. I wouldn’t read much into that yet. Arc only went live on September 16. The bigger question is what happens to the liquidity now that it’s there. Aave’s Arc deployment has a $2M annual minimum revenue commitment to the DAO, with ecosystem participants covering any shortfall during the first five years. That’s around $166.7K a month or $5.5K a day. Getting USDC onto Arc is one thing. Getting people to borrow it is another. That’s the part I’m watching: Supply → Borrowing → Utilization → Fees → Revenue If deposits keep climbing while borrowing stays low, much of that capital is just sitting there. If borrowing starts picking up, that changes the picture and if the added activity begins generating meaningful fees, we’ll have a much clearer signal that Arc is becoming a real lending market. TVL tells you capital showed up but Utilization tells you whether anyone is actually using it.

The RWA market has a utilization problem. A lot of real-world assets are moving onchain. But what matters more is what actually happens to them once they get there. DeFiLlama currently tracks $33.15B in RWA onchain AUM, but only $3.8B in DeFi active TVL. So roughly 12.7% of that value is being actively used across DeFi. And when you look at individual platforms, the gap becomes even clearer. ↳ @Securitize ~$5.18B onchain AUM ~$27.8M DeFi TVL ↳ @OndoFinance ~$3.36B onchain AUM ~$60.5M DeFi TVL ↳ Spiko ~$2.63B onchain AUM ~$40M DeFi TVL Now compare that with: ↳ @centrifuge ~$1.67B onchain AUM ~$441M DeFi TVL ↳ @maplefinance ~$1.90B onchain AUM ~$1.12B DeFi TVL That’s a pretty big difference. Putting an asset onchain is one thing. Finding ways to actually use it in DeFi is another. And that’s where things are starting to get interesting. ↳ @aave — Horizon Aave Horizon lets qualified institutions use tokenized RWAs as collateral to borrow stablecoins. Instead of selling the asset to get liquidity, the asset itself can help unlock it. ↳ @MorphoLabs × @OndoFinance Tokenized equities like SPYon and QQQon are moving into lending markets. Morpho also lets each market set its own risk parameters, which is important because not every RWA carries the same risk. ↳ @BitwiseInvest — PAPY PAPY uses Morpho to lend against approved RWA collateral. So the asset isn’t just sitting onchain anymore. It’s being used as part of a lending strategy. That’s the shift worth watching. The RWA story isn’t just about putting more assets onchain. It’s about making those assets useful once they’re there. Tokenization puts the asset onchain. Credit puts it to work.

THIS WEEK’S TOP 10 TOKEN UNLOCKS More than $349M worth of tokens are scheduled to be released across the crypto market this week, according to DeFiLlama’s unlock tracker. THE 10 TO WATCH: $ORDER — Sep 21 $18.6K unlock → 0.12% of circulating supply $15.4M market cap · $0.037 $RIVER — Sep 22 $165.6K weekly → 0.71% of circulating supply $23.5M market cap · $1.20 $H — Sep 23 $22.16M unlock → 14.58% of circulating supply $152M market cap · $0.076 $THL — Sep 24 $472/week → 0.097% of circulating supply $486.6K market cap · $0.0066 $RESOLV — Sep 25 $33.8K unlock → 0.41% of circulating supply $8.26M market cap · $0.017 $XPL — Sep 25 $7.99M unlock → 3.20% of circulating supply $249.9M market cap · $0.09 $ALT — Sep 25 ~$1.49M unlock → 3.37% of circulating supply $44.3M market cap · $0.0062 $WAL — Sep 25 $1.04M unlock → 1.28% of circulating supply $81.7M market cap · $0.032 $SCA — Sep 25 $1.89K weekly → 0.25% of circulating supply $772K market cap · $0.0047 $GRASS — Sep 27 ~$3.6M unlock → 1.55% of circulating supply $232.5M market cap · $0.34 That’s it for this week’s unlocks. See you next week with another update.

Weekly Crypto Recap SEPTEMBER 14–20, 2026 20 of the biggest crypto events from September 14–20: ⋮➤ SEC gives tokenized U.S. stocks a five-year regulatory window ⋮➤ CLARITY Act fails to advance in the Senate ⋮➤ CFTC sends crypto market rules to the White House ⋮➤ Fed raises rates 25 bps to 3.75%–4.00% ⋮➤ Circle launches Arc mainnet with 100+ applications and builders ⋮➤ Aave V4 deposits reach ~$709M, with Arc at ~$58M ⋮➤ Hyperliquid adds HYPE/BTC-backed borrowing ⋮➤ Bitcoin moves back above $80K ⋮➤ Bitcoin ETFs finish slightly positive while Ethereum ETFs lose $140M ⋮➤ Zcash ETFs record $98.2M in weekly inflows ⋮➤ Solana moves to 250ms slots ⋮➤ Ethereum sets October 6 for the Glamsterdam Sepolia fork ⋮➤ Symbiosis attacker creates 46.1B fake BTC tokens ⋮➤ Two former Robinhood engineers charged over alleged insider trading ⋮➤ Deutsche Bank moves toward institutional crypto custody ⋮➤ Bastion gets conditional approval for a U.S. national trust bank ⋮➤ Coinbase moves to bring stock perpetuals to U.S. markets ⋮➤ World launches World Money across 150+ countries ⋮➤ S&P Global agrees to acquire OpenZeppelin ⋮➤ Zama expands private DeFi to 16 vaults ⸻ THE DETAILS ⥤ Tokenized stocks: The SEC approved temporary rules allowing certain platforms to trade tokenized U.S. stocks for five years, with applicable shareholder rights preserved. ⥤ CLARITY Act: The Senate procedural vote was 50–49, short of the 60 votes needed to advance the bill. ⥤ @CFTC : Its crypto market-structure proposal entered White House review through OIRA. ⥤ Fed: Rates increased 25 bps to 3.75%–4.00%, the first hike since 2023. ⥤ @arc : Circle launched its mainnet with 100+ applications and builders. USDC is the native currency and pays network fees. ⥤ @aave V4: Deposits reached about $708.6M, with Arc at roughly $57.6M. ⥤ @HyperliquidX : HYPE and BTC can now be used as collateral to borrow USDC or USDT. About $269M was reportedly borrowed on day one. ⥤ Bitcoin: BTC moved back above $80K, with leveraged positions liquidated during the move. ⥤ ETF Flows: Bitcoin ETFs ended the week around +$6.2M, while Ethereum ETFs saw about $140M in outflows. ⥤ @Zcash : Zcash-related ETFs recorded roughly $98.2M in weekly inflows. ⥤ @solana : Mainnet slot time was cut to 250ms, making block production faster. ⥤ Ethereum: The Glamsterdam Sepolia fork is set for October 6, with mainnet still targeted for Q4. ⥤ @symbiosis_fi : An attacker created 46.1B syBTC after starting with roughly $0.25. The estimated initial loss was around 10 BTC. ⥤ Robinhood: Two former engineers were charged over alleged insider trading involving confidential listing information and Hyperliquid perpetuals. ⥤ @DeutscheBank: The bank plans institutional custody for assets including BTC, ETH, USDC and EURC. ⥤ @Bastion: The OCC gave Bastion conditional approval to establish a national trust bank focused on services including stablecoin custody and payments. ⥤ @coinbase : Its derivatives arm proposed perpetual futures tied to individual U.S. stocks and ETFs, subject to approval. ⥤ @worldmoney : World Money launched in 150+ countries, combining stablecoins, payments, trading and virtual accounts. ⥤ @OpenZeppelin : S&P Global agreed to acquire the crypto security and development company. ⥤ @zama : Its private DeFi system reached 16 vaults, with the first private Morpho vault passing $40M in shielded TVL. That was the week in crypto September 14–20. Hope you found this week’s recap useful. See you next week.

HYPE just hit a new ATH at $94.47. The price is getting all the attention, but the numbers behind it are more interesting. At that price, HYPE was around a $20.6B circulating market cap. So what’s actually behind the move? @HyperliquidX is way past being just another perp DEX. By June 30, it was doing roughly 9.4% of global perp volume, including CEXs. By Aug. 23, it had about 63% of decentralized perp OI. That’s a serious amount of the market. Hyperliquid isn’t just taking volume from other DeFi venues anymore. It’s starting to capture derivatives flow from CEXs as well. (Nasdaq) Then there’s the HYPE tokenomics. 99% of protocol fees go to the Assistance Fund, which uses those fees to buy HYPE from the market. By Aug. 23, it had bought 46.7M HYPE, permanently removing roughly 4.7% of the initial supply from circulation. The loop is simple: Trading → fees → HYPE buys → supply removed. That doesn’t mean HYPE has to keep going up. But it does give the token a clear link to activity on the protocol. Then there’s RWA perps. RWA markets made up more than 50% of Hyperliquid’s weekly volume for two straight weeks in July. HIP-3 OI also crossed $4B in August. (Nasdaq) That’s where the Hyperliquid story starts getting bigger. It’s no longer just about crypto perps. The platform is adding markets around RWAs, outcome markets and other products on the same trading and margin infrastructure. (Nasdaq) Now look at @Robinhood Chain. It launched on July 1, and the early numbers have moved quickly. CoinDesk had TVL around $757M on Sept. 4. Later reports put it near $1B, while Bernstein cited roughly $1.5B TVL and $50B+ in cumulative DEX volume around Sept. 8. Those are different dates and sources, so I wouldn’t treat them as one current figure. Still, the amount of activity this early is worth watching. And there’s another number that stands out: Robinhood’s own app users were estimated to account for only ~1–2% of Chain transactions in early September. So far, a lot of the activity has been coming from crypto-native users. That matters because Robinhood already has something most new chains have to spend years building: an existing user base with assets already sitting on the platform. At the end of August, Robinhood reported 28.6M funded customers and $384B in platform assets. (Robinhood Markets, Inc.) The question is how much of that activity eventually makes its way onchain. Then there’s tokenized equities. Tokenized-equity value on Robinhood Chain was estimated to grow from around $10–12M to $140–150M in roughly two months. And on Sept. 17, the SEC announced a five-year exemption framework for certain tokenized-stock trading, with conditions around investor rights and issuer notification. That doesn’t mean every tokenized-stock model works. But it does make the tokenization side of Robinhood worth watching more closely. Hyperliquid → trading infrastructure. Robinhood → users, assets and distribution. Hyperliquid is proving that serious derivatives activity can happen onchain. Robinhood already has the customers and financial assets. The interesting question is how much of that existing ecosystem can actually be brought onchain. If tokenized equities become collateral, RWA perps keep growing, stablecoins become the settlement layer, and more Robinhood activity moves onchain, these two models could eventually start overlapping. That’s a much bigger story than another perp DEX or another L2. It’s about where capital, assets and trading activity end up living.

Most people see confidential DeFi as a privacy upgrade. I think there’s more to it. What happens when privacy doesn’t just protect your balance, but your whole financial activity deposits, positions, swaps and rewards? DeFi’s transparency is great for verification, but it comes with a cost. You can track wallets, see positions, follow large transfers and watch capital move in real time. For someone moving serious money, that can become a problem. A $20M position isn’t just a balance. Other traders can study it, trade around it and potentially copy the strategy. That’s where @zama gets interesting. Instead of building another isolated privacy ecosystem, Zama is adding confidentiality to existing DeFi infrastructure, including @Morpho. ERC-7984 is a key part of this. It allows existing ERC-20 assets to have confidential versions, so assets like $USDC, $USDT and $WBTC can gain a privacy layer without moving liquidity into a completely separate ecosystem. Zama’s first confidential Morpho vault reportedly crossed $40M in shielded TVL within seven weeks. It has since expanded to 16 vaults covering $USDC, $USDT, $WBTC, $AUSD and $TGBP, with curators including @SteakhouseFi, @flowdesk_co, @RockawayX and @Bitwise. But one detail really stood out to me: The rewards need to stay private too. Hiding someone’s $USDC position means less if their public rewards can still reveal its size. That’s where @merkl_xyz comes in. Zama and Merkl are working on calculating incentives from encrypted balances and distributing rewards privately. So privacy doesn’t stop at the deposit. It follows the activity. But there’s another problem: A private balance doesn’t automatically mean private activity. Someone can deposit $USDC privately, then unshield to trade on a public DEX. The balance may be hidden, but the transaction trail can still reveal what they’re doing. That’s why the next challenge is private execution. Zama’s Confidential Swap Protocol is designed to let confidential assets and vault positions trade without exposing the size or intent of the trade. The technology behind this is FHE — Fully Homomorphic Encryption. Simply put, it allows systems to work with encrypted data without exposing the underlying information. You can verify what happened without seeing every financial detail. There are still trade-offs. Private positions can make liquidation, risk monitoring and liquidity discovery harder. And privacy needs to remain connected to the wider DeFi ecosystem. But that’s what makes this worth watching. The goal isn’t to hide everything. It’s to keep what needs to be verified public, while protecting what doesn’t need to be exposed. Once confidentiality covers deposits, positions, swaps and rewards, it becomes much more than a privacy feature. It starts looking like a new layer for onchain finance.