Stablecoins are a $305.6B market. Four projects offer very different ways to capture that growth. Here is how I would map Sky, Ethena, Aave and Pendle as of September 7, 2026. 1. @SkyEcosystem: earn the spread Sky earns interest and income from the assets and credit supporting its dollar system. The economic opportunity is to grow USDS while keeping asset income ahead of savings payouts and other costs. DefiLlama tracks $13.51M in protocol revenue and $2.91M in holder revenue over 30 days. Protocol surplus already supports SKY buybacks and staking rewards. Watch USDS growth alongside the surplus retained per dollar. Expensive deposits can grow a stablecoin without improving its economics. 2. @ethena : scale the yield engine Ethena generates income through USDe's backing assets and hedging strategy. That income also funds sUSDe returns and distribution, so the amount available for ENA depends on how revenue is allocated. USDe currently has roughly $4.37B in circulation. The August 27 fee-switch proposal places the first buyback threshold at $7.5B, with 95% of net revenue paid to the Foundation from its core business lines earmarked for buybacks once activated. The opportunity is conditional: grow USDe, preserve competitive yields and turn the fee framework into executed token purchases. 3. @Aave: monetize borrowing demand Aave earns a share of the interest borrowers pay. More stablecoins become economically useful when borrowers actually draw on that liquidity. It recorded $4.57M in 30-day protocol revenue. AAVE buybacks were paused in April, and DefiLlama records zero holder revenue for the current 30-day window. The business can keep earning while token purchases pause. Watch stablecoin borrowing, utilization and confirmed buyback resumption. 4. @pendle_fi : build markets around yield Pendle lets holders separate principal from future yield, creating markets for fixed returns and changing yield expectations across different issuers. Pendle V2 allocates 80% of protocol fees after the LP allocation to PENDLE buybacks, with repurchased tokens supporting eligible sPENDLE holders. DefiLlama tracks approximately $521K in holder revenue over 30 days. Its opportunity grows when more yield-bearing assets find active markets on Pendle. The relevant signals are yield fees, trading activity and recurring demand across maturities. The financial figures above cover whole protocols under DefiLlama's definitions. They are not a measure of stablecoin-only profit. My research priority would be SKY and PENDLE for cash flow already reaching tokens, AAVE for borrowing demand, and ENA for conditional fee activation. A growing market creates opportunity. The token thesis depends on how much of that opportunity it can actually retain.

Most prediction apps give you a question and ask you to pick yes or no. @DeepBookonSui Predict lets you choose a BTC price or price range yourself. DeepBook App launched in Alpha on Sui on September 24, with Spot trading and Predict in one interface. For Predict, you can choose: - A specific BTC price or range - A timeframe as short as one minute The app requests a quote for that position. A volatility oracle helps price it, while a shared liquidity pool takes the other side of the trade. That means DeepBook does not need to create a separate market for every price users might choose. The useful shift is flexibility: traders can express a more specific view than a preset yes/no question. The part worth watching is whether the shared pool can keep offering useful quotes as more people choose different prices and timeframes.

“Private transfer” does not always mean the same thing. eERC: hide the numbers @avax eERC encrypts token balances and transfer amounts. The sender, recipient, and transaction time remain public. Imagine a payroll transfer: an observer can see who paid whom, but not the amount. An authorized auditor can be given access to the encrypted figures. @RAILGUN_Project : hide the transfer details You first move tokens from a public wallet into a RAILGUN private balance. This deposit reveals the token and amount. Once the funds are inside, a transfer to another RAILGUN wallet hides the sender, recipient, token type, and amount. A Broadcaster can submit the transaction without using your public address as its sender. Withdrawing to a public wallet is visible again. The question to ask Do you need to hide how much was sent, or also who paid whom? eERC addresses the first case. RAILGUN’s private transfers address both, but funds cross public boundaries when they enter and leave its private system

A claw machine earns every time someone plays. How does that become onchain income? PLAY from @DualMintRWA connects arcade revenue to a machine vault on Solana, built around a 200-machine fleet. The useful detail is how the money moves. Players pay to use the machines. The operator runs the business and pays a fixed monthly lease. Those lease payments are the intended source of distributions to PLAY holders. According to DualMint’s product page, the lease payment comes before the operator takes their share, and DualMint charges 10% of that payment. So the underlying business earns from gameplay, while the holder’s income follows the lease arrangement. More plays do not automatically mean a bigger monthly payout. PLAY targets 12-15% annual yield with monthly distributions. That is a target, not a guaranteed return: the operator still needs to meet its obligations. This gives readers a concrete question to investigate: can the machines’ operating income consistently support the lease payments?

Three ways to use AI in crypto research beyond asking which token will pump Kaito Pro: find the context Use it when information about a project is scattered across sources Its search and tracking tools help you investigate a ticker, topic or narrative without relying entirely on what appears in your feed The useful output is a set of sources worth reading, not just another summary Dune: turn a claim into a chart A project says activity is growing Ask what that actually means: more transactions, more wallets, or more returning users? Dune's MCP integration lets a connected AI agent find datasets, write queries and build visualizations You still need to check the query's definitions Unique wallets aren't automatically unique people Nansen: monitor what happens next Once you've found wallets worth following, its AI can help configure Smart Alerts Supported alerts include token transfers, aggregated Smart Money flows and changes to Hyperliquid perp positions You review the setup and create the alert, instead of repeatedly refreshing the same wallets A practical workflow: • Find the claim • Check it against data • Set an alert for what would change your view These tools help with different parts of the job The value is in connecting those steps, rather than collecting three AI-generated opinions

The same 5% borrowing rate can mean different trade-offs on @LiquityProtocol, @Morpho and @eulerfinance Liquity V2: you choose the rate Borrowers set their own interest rate or delegate its management Choosing a lower rate puts your loan closer to the front of the redemption queue within its collateral branch A redemption exchanges some of your collateral for BOLD to reduce your debt, even when your loan is healthy Morpho: the rate adapts Markets using AdaptiveCurveIRM adjust rates based on how much supplied capital is borrowed, targeting 90% utilization The model also adjusts its rate level over time Two markets with the same utilization can therefore charge different rates because their histories differ Euler adds another layer: the individual vault Euler supports multiple interest-rate models, including utilization curves and adaptive models You need to check which model a vault uses, its parameters, and who can change it Imagine comparing loans that all display 5% That's an example, not a quote of current rates The useful questions are: • What makes that rate change? • Who controls the rules? • Does paying less introduce another trade-off? On Liquity, a lower rate can increase redemption exposure On Morpho, borrowing demand and its history shape the rate On Euler, the answer depends on the vault's configuration The APR is the starting point, the rules explain what happens next.

Your BTC stays on @Bitcoin, your USDC loan lives on @ethereum. How? @ZestProtocol opened a Bitcoin Collateral Vaults mainnet demo on September 23. The interesting part is how the collateral connects to the loan. Two chains, one position BTC is locked in an individual Taproot vault on Bitcoin. Ethereum holds a record tied to that vault, allowing the owner to borrow USDC against it. The BTC isn't pooled with other users or converted into a wrapped token. What happens if you get liquidated? According to Zest, the permitted paths for moving BTC are signed at deposit. An operator cannot add a new destination afterward. If the position breaches its liquidation threshold, the design sends only the required portion to a registered liquidator, rather than liquidating the entire vault. That distinction matters: keeping BTC on Bitcoin doesn't mean keeping it beyond a lender's reach. Still an early demo Collateral is capped at 0.001 BTC per wallet. Zest describes BitVM-based verification in its architecture, but that alone doesn't establish which verification components are active in this demo. The discovery here is a different way to connect Bitcoin collateral with Ethereum liquidity. Its spending permissions and settlement checks matter more than the “no wrapping” headline.

➥ top protocols building AI analysts for all the first 2 tools are most important to aid my research flow with real-time and precise data for context, AI analysts are becoming one of the most practical applications of AI in finance they will not eliminate the research edge but they will remove much of the operational gap between an independent researcher & a professional research desk these are the 3 i’m equiped: [1] @artemis AI Analyst [2] @DefiLlama LlamaAI [3] @SoSoValueCrypto AI Socrates i’ve personally used Artemis and LlamaAI the most i use LlamaAI at the start of my research flow it helps me screen protocols, compare competitors, and quickly check changes in TVL, rev, vol, fees, stablecoin supply, or market share instead of opening ten dashboards and manually building a table now i can first ask a specific question, identify what changed, and decide whether the signal deserves deeper research i use Artemis when the thesis requires more context for example, onchain activity alone cannot explain how Robinhood Chain may affect Robinhood’s earnings i also need txn rev, company guidance, public-market comparisons, and broader financial conditions Artemis is useful for connecting those layers in one workflow i’m still exploring Socatis, mainly as another source for macro context, market trends, and cross-checking my initial view the benefit is not limited to saving time, these tools can help people: - turn a broad idea into a testable research question - compare protocols using consistent metrics - find changes that are easy to miss across dashboards - connect crypto activity with macro and public markets - build charts, reports, decks, and financial models - challenge a thesis before publishing or allocating capital i still verify important figures against primary sources AI can retrieve and organize the evidence, but it cannot decide which metric matters, whether the data is comparable, or what the market has already priced in that is why i think AI analysts expand access without removing the edge the edge shifts from finding data to asking better questions, checking the evidence, and forming an original view

Liquid staking on a Proof-of-Work chain? Powfi just launched on @alephium, the interesting part is where the yield comes from. 1. What you're actually staking Lock ALPH and receive xALPH, a token representing your deposit. When rewards enter the contract, each xALPH becomes backed by more ALPH. Your token balance can stay the same while the amount it represents grows. You can also trade xALPH or provide liquidity in supported pools. 2. Follow the money Powfi's docs describe two reward sources: • A share of DEX fees. • Funded staking campaigns. The documented fee model uses a share of protocol fees to buy ALPH, then splits the purchased tokens: 50% burned, 50% distributed to stakers. That connects the staking model to exchange activity. You're depositing into an application, rather than becoming a network validator. But launch incentives and trading income are different things. A high initial APR alone wouldn't tell us how much yield the DEX can sustain. 3. “Liquid” still has an exit cost Official unstaking releases ALPH gradually over 30 days. Selling xALPH on a DEX offers a faster exit, but the price depends on liquidity and slippage. The interesting experiment: can a native DEX turn trading demand into a lasting reason to hold ALPH?

What if you could borrow against a prediction, buy an option on its price, or get paid before your winnings arrive? Three projects are exploring those possibilities. 1. @PositionsFi: borrow while keeping your bet Eligible Polymarket positions can be pooled as collateral to borrow pUSD. You keep your exposure and access cash without selling. The trade-off: you now have debt and can be liquidated. Status: beta. 2. @convallax: trade how a prediction's price changes Convallax is testing calls and puts on Polymarket YES prices. An option can expire before the event ends. You could trade a rise in the YES price ahead of a debate or announcement, without waiting for the final outcome. Status: Polygon Amoy testnet, with a trusted settlement signer. 3. CUSP: get cash while waiting for a winning claim to settle Its proposed model buys eligible winning claims at a small discount, pays the holder now, and collects the full payout later. The holder gets cash sooner. The buyer earns the difference and takes the risk of delayed or failed payment. Status: documented design; live deployment unverified. Three different needs: - Access cash without selling. - Trade price changes before the result. - Shorten the wait for payment. These early projects suggest a financial ecosystem is forming around prediction-market positions.

You can be bullish on funding without being bullish on Bitcoin. @pendle_fi Boros lets traders take positions on funding rates. A long Yield Unit position receives the underlying funding stream and pays a fixed rate set at entry. Imagine entering at 10% implied APR and holding to maturity. If the realized funding stream averages 14% annualized over that period, the funding leg earns the difference before fees. If it averages 6%, that leg loses. Those are hypothetical rates. The four-percentage-point spread is annualized on the position's notional, not a 4% return on your deposited margin. There is also a second moving part: the market's expectation of future funding. Your position can gain or lose value before maturity as implied APR changes. Because Boros uses margin, an adverse move can trigger liquidation before the funding outcome you expected has time to arrive. That makes Boros useful for expressing a specific view: traders will pay more for leverage than the rate market currently prices. The question becomes how much funding will accumulate over the remaining term, and whether the position has enough margin to survive the path. Being right about the eventual average is only part of the trade.

Your tokenized-stock trading history can now influence your access to private-market deals. PancakeSwap's first Pre-Access campaign, featuring $pPOLY from @Paimon_Finance, connects two markets through its allocation rules. 1. Your past activity affects your subscription limit Binance's Pre-Access FAQ lists three inputs: - Alpha Points - Trencher Badge status - bStocks On-Chain Tier For pPOLY, the announced bStocks assessment uses holdings and trading volume from September 6–20, before the campaign opened. Two eligible users can arrive with the same capital and receive different subscription limits because their qualifying activity differs. A higher limit still does not guarantee a full allocation. 2. Tokenized stocks gain another use inside the ecosystem The connection now looks like this: Hold or trade bStocks → build a qualifying activity tier → potentially unlock a larger Pre-Access subscription limit This gives users another reason to participate in tokenized-stock markets beyond the stock exposure itself. If future campaigns repeat this structure, access to new offerings could become an incentive to keep assets and trading activity within the ecosystem. For issuers, it offers an existing audience with measurable participation before a sale opens. 3. This changes how we should interpret activity A rise in bStocks volume around these campaigns could have multiple drivers: - Demand for the underlying stock exposure - Positioning for future subscription limits - A combination of both Volume alone cannot tell us which one dominates. That makes campaign timing and allocation rules relevant when researching tokenized-stock adoption. The product boundaries also matter: Paimon issues pPOLY, PancakeSwap hosts the campaign, and Binance Wallet provides access. Pre-Access does not grant direct ownership of the target company's shares or guarantee an IPO. The development worth following is how onchain activity becomes a criterion for distributing private-market exposure, and whether that creates lasting participation beyond individual campaigns.

.@Ondo just changed how institutional share inventory can reach onchain markets. Before: An institution that already held the underlying shares still needed separate cash to mint Ondo Stocks. That created financing costs and a timing gap between its traditional inventory and tokenized positions. Now: - Approved institutions transfer existing shares from their Alpaca account to Ondo’s Alpaca account. - The corresponding Ondo Stocks tokens are minted on Ethereum or BNB Chain. - Redemption reverses the process and returns the underlying shares. This gives market makers a direct route between traditional shares and onchain inventory, making it easier to replenish supply when demand increases. The expected result is tighter spreads and deeper secondary-market liquidity. The limitation is important: this is not available to retail users. Access is approved case by case and requires accounts with both Ondo and Alpaca. The real upgrade is not another tokenized stock. It is a two-way conversion rail between institutional inventory and onchain liquidity.

.@ZetaChain holders approved more than a migration for $ZETA. If the next proposal passes, @Solana becomes its canonical home and the ZetaChain L1 eventually winds down. 1. What governance approved Proposal 68 passed with 99.4% voting Yes and 58% participation. It approved the direction of the migration: - ZETA becomes a native SPL token - Holders convert 1:1 under the same ticker - Total supply and vesting schedules remain unchanged - The ZetaChain L1 winds down after holders and connected-chain assets have migrated This is not a wrapped token or a temporary bridge. Solana would become the canonical home of ZETA. 2. Where the new utility comes from ZETA already has an application-level role inside @AnumaAI, a private AI app that ZetaChain reports has more than 300,000 users and over one million AI requests. The mechanism is simple: - Lock ZETA - Receive application credits - Spend those credits on AI usage Locked ZETA leaves circulating supply, while Anuma gives each user one encrypted memory that can provide context across 35 AI models. The longer-term plan is to let other AI apps and agents use the same Private Memory Layer, with ZETA acting as their common access token on Solana. 3. What has not happened The migration is not live. There has been no snapshot, chain halt or balance conversion. Exchange support must be confirmed first. A second governance proposal will then define the mechanism, dates, claim process and holder protections. The important shift is not Solana by itself. It is the attempt to turn ZETA from the native asset of a cross-chain L1 into an access layer for private AI applications and agents. The product is live. The migration still depends on execution.

.@onrefinance is turning reinsurance yield into a risk-tranched DeFi position. The structure starts with ONyc, whose yield comes from: - Reinsurance premiums - Returns on stablecoin collateral - Liquidity incentives That exposure is then divided into two risk profiles: - jrONyc provides first-loss capital. - srONyc accepts a lower yield in exchange for protection up to the available coverage ratio. The new isolated USDG market on @loopscale makes the senior position more usable across DeFi: - Users can build leveraged srONyc strategies with variable-rate exposure. - @ExponentFinance provides access to fixed-rate positions. Senior still does not mean risk-free. Returns depend on ONyc performance, junior coverage, borrowing costs and market liquidity. Leverage also introduces liquidation risk. The real product is not another yield vault. OnRe is defining who absorbs insurance losses first, pricing each layer differently and making that risk structure composable across DeFi.

.@verantaxyz just gave traders a $100,000 reason to test its new identity. Following its rebrand from Avantis and the launch of v2, Veranta has started a five-week trading competition powered by @base. Upside Perps now supports limit orders, letting traders define an entry instead of waiting in front of the chart. Daily 2x multiplier codes and weekly PvP livestreams add another layer around rankings and participation. New name, upgraded product and a concrete reason to return. Week one is now live.