MarsBit has published its Weekly Editor’s Picks for Oct. 10-16, bringing together longer reads across macro markets, investing and startups, AI and storage, CeFi and DeFi, Bitcoin, Ethereum scaling, security and the week’s major headlines.

Macro outlook
In the macro section, MarsBit highlighted an article titled Today’s S&P 500: Historical Highs, Bottom of the Bull Market. The piece argues that the S&P 500 can be close to all-time highs while still sitting at the “bottom” of a broader bull market. Inflation, war and rate hikes, it says, have failed to stop U.S. equities from pushing higher, showing unusual resilience.
The article also says that most S&P 500 constituents have not performed well. In that reading, the breadth weakness can be seen both as evidence of a bear market and as a signal for a rebound. From a macro perspective, demand for holding assets is also described as being near historical highs. If conditions line up, the piece says, the S&P 500 could enter a sharp rally, while 2027 may become a year full of major variables.
Investing and startups
One featured article, Real Vision Founder: Why Most Public Chains Are Destined to Die, argues that major platform markets such as cloud computing, operating systems, smartphones and public blockchains tend to end with only a handful of survivors. The piece defines the actual product of a public chain as blockspace and says everything else is mostly marketing. In its framework, judging an economy, including a blockchain economy, comes down to how much real economic activity is produced on-chain.
The article lists four survival paths: becoming an independent economy, with Ethereum and Solana cited as examples; taking a vertically specialized route, with Hyperliquid named here; operating as a layer-2 network; and BNB, which the piece describes as a centralized platform token wearing an L1 wrapper. Outside those categories, it says, most public chains have no moat and their value eventually goes to zero. On the practical side, the article says the mechanism for token listing and issuance is what actually kills them.
It offers three main indicators to watch: whether economic density per user holds up when token prices collapse; whether stablecoin balances continue to grow during a token drawdown; and whether developers keep arriving.
Another article, More Weekly Revenue Than Hyperliquid, $1 Million in Daily Buybacks: Is Pump.fun Undervalued?, focuses on Pump.fun. It says the platform remains highly profitable even after the meme-trading wave cooled. Over the past 30 days, Pump.fun generated $65.63 million in protocol revenue, close to Hyperliquid’s $71.05 million.
Yet PUMP’s valuation is described as much lower than HYPE’s relative to revenue. Their circulating market capitalizations are given as about $2.325 billion and $18.198 billion, a gap of nearly eight times. Using a simple annualization of the most recent 30-day protocol revenue, the article says the ratios of circulating market cap to annualized revenue are roughly 3x for PUMP and 21x for HYPE.
It also says Pump.fun’s monthly buyback spending is about 1.27% of its market capitalization. After the first large unlock, there was no concentrated wave of selling, and later sell pressure weakened noticeably. Following changes to Callout Rewards and an expansion in trading categories, traders are again treating Pump.fun’s profit engine seriously. The piece says Pump.fun could develop into a super app in this cycle.
Even Losing Money Requires Waiting in Line: How Far Can PaperTrade’s 1,000x Leverage Experiment Go? examines PaperTrade, which supports leverage of up to 1,000x. Users who lose money or get liquidated can receive the native token PAPER. The structure described in the article has users trading directly against a shared USDC liquidity pool: when users lose, funds go into the pool; when users win, the pool pays out the profit.
While the rules are public and there was no pre-allocation, controversy followed. The article says smaller users face longer waits, and because PaperTrade relies on Hyperliquid pricing, it has run into price-manipulation issues. After the frenzy of the first day, open interest on the platform has already pulled back sharply. The lost USDC is certain, while whether PAPER can offset that cost depends on future platform distributions and market pricing.
Another featured piece, Not Even 237,000 Users Could Save StandX: Perp DEXs Enter the Elimination Round, says decentralized perpetuals platform StandX is shutting down, canceling its planned platform token launch and preparing a one-time airdrop in DUSD for users. The article argues that perp DEXs remain one of the strongest demand segments in crypto trading, but growth in the segment no longer means every participant will get a share. Subsidies have raised user acquisition costs, and airdrops can manufacture volume but not real trading demand. In that framing, perp DEXs have entered an elimination phase.
The section also points readers to related reads on PaperTrade and to another recommendation, 2027: Token Economics Enters the Era of Value Capture.
AI and storage
In the AI and storage section, MarsBit included The Market Has Finally Realized AI’s Biggest Bottleneck Is Not Power, Not Chips, but Money. The article says that once capital is no longer unlimited or cheap, AI stops being treated as a technology myth and returns to a basic return-on-investment calculation. Tighter credit and a higher cost of capital are raising the risk of delays in compute infrastructure buildouts, and the semiconductor supply chain may face a cliff in 2027 to 2028 if demand does not convert into actual purchasing orders.
The Trading Logic for AI Is Changing, and Global Chip Stocks Are Slumping argues that sentiment is shifting after nearly four years of AI infrastructure investment mania. Rising interest rates, bigger swings in the bond market, higher AI financing costs and closer scrutiny of whether huge capital expenditures can turn into real returns are all becoming more important. The article says ChangXin’s technological progress may have acted as a direct catalyst for the sell-off in storage names, but the deeper change is that investors are moving from chasing AI growth to examining capital returns.
A third piece, Storage Deleveraging and Korean Retail Investors’ Liquidation Fears, lists three sources of pressure: whether continued AI infrastructure expansion can translate smoothly into actual revenue and profit; whether rising prices for upstream equipment and components are rewriting return expectations across the chip supply chain; and whether the fast revenue growth seen at frontier large-model labs can remain sustainable.
CeFi and DeFi
The CeFi and DeFi section features Behind the Boom in Tokenized Stocks: Who Is Making Money and Who Is Building Value for Others?. According to the article, traders and investors now hold more than $6 billion in on-chain equity exposure through two paths: tokenization and synthetics. It says there are two main reasons to bring equities on-chain: buying, selling and holding can be cheaper, and the assets can be used for other purposes as well. It also points out that while Wall Street closes on weekends, on-chain equities do not.
The article argues that perpetuals are useful as directional signals, but poor weekend prices. The key constraint is correct and continuous price feeds. The better perpetuals can price weekends, the more tokenized stocks can function as collateral; the more they are used as collateral, the more hedging is needed, and that in turn can raise perpetual volume.
As of the end of June 2025, foreign investors held $19.86 trillion in U.S. equities, equal to 18.3% of total U.S. stock value, the article says. Tokenized stocks currently amount to only 0.016% of that foreign-held total. On that basis, the piece argues that future growth has to come from stocks already sitting in brokerage accounts, and compares Robinhood’s role in tokenized stocks to Circle’s role in stablecoins in 2018.
Over the past three months, Binance and Robinhood accounted for 86% of new holders. BNB Chain added 1.72 million addresses, while Robinhood Chain added 1.55 million. BNB Chain also brought in $517 million in value, more than all other chains combined. Solana and Ethereum together added 518,000 holders.
The article says the end state for shares moving on-chain is for a company’s transfer agent to record the shares themselves on a blockchain, with no offshore issuer, no standalone custodian and no claim on someone else’s account. In that model, the token is the share. A small group of leaders is already working on that piece, it says. Across the value chain, money is flowing to front-end applications, while chains and custodians are in the weakest position.
MarsBit also listed related readings on Ripple’s largest DAT listing and Pendle’s new arbitrage product.
Airdrop opportunities and interaction guides
The roundup’s airdrop and user-action section includes Popular Interaction Collection | Catalyst Waitlist Application; RocX Testnet Is Live dated Oct. 10, Popular Interaction Collection | LOL Event Goes Live on predict.fun; RAFA Finance Testnet Is Live dated Oct. 15, and After DRV’s New High, What Early Interaction Opportunities Remain in On-Chain Options and New Perpetual Projects?.
Bitcoin
In the Bitcoin section, MarsBit highlighted PRL Surges 3x: Can Pearl Become the “AI Version of Bitcoin”?. The article says Pearl, or PRL, has been gaining traction in overseas communities as a GPU mining token. Pearl is described as an L1 protocol that combines AI computation with GPU mining. A user’s GPU can serve AI training and inference workloads while also performing proof-of-work for the blockchain and generating PRL, a model the piece calls proof of useful work, or PoUW.
Over the past few months, the article says, one of Pearl’s most credibility-boosting moves was joining Nvidia’s AI startup ecosystem. Pearl is also building an AI privacy trust layer, and its use of zero-knowledge proofs for AI privacy protection may become one of its central narratives.
On mining economics, the piece gives one data point: a 4090 card currently produces a daily net income of 50 yuan. It also says Pearl’s core idea is “one computation, two outputs,” but that the PRL token still does not capture any of Pearl’s AI commercial value. That missing loop is presented as a problem the project still needs to solve.
The section also recommends a Bitcoin mining profitability calculator article covering hashrate income, electricity costs and shutdown prices.
Ethereum and scaling
Ethereum’s Moat Has Not Disappeared; It’s Just No Longer Priced by Transaction Volume focuses on Ethereum’s competitive position. The article lists several common arguments behind the claim that Ethereum is losing its last moat, liquidity, and then offers counterviews.
One claim is that tokenized stocks on Ethereum have almost no trading volume even though Ethereum holds the second-largest share of tokenized stock supply. In that view, stocks sitting on a blockchain do nothing if they produce no revenue, and only trading volume creates revenue. The author’s response is that Ethereum does not necessarily need to generate revenue because it already has monetary premium.
A second claim is that flagship applications are choosing to grow elsewhere first. The author argues that layer-2 growth is itself a win for Ethereum, because applications are expanding distribution channels rather than leaving Ethereum.
A third claim is that Ethereum’s stablecoin dominance is gradually weakening. The response given is that stablecoins are not a winner-takes-all market.
Multi-ecosystem and security
The multi-ecosystem section includes Uniswap v4 Hook Ecosystem Heats Up: Eight Early Projects Race Ahead, but Which Ones Will Build Real Business?.
The security section features Lamborghini, Non-Disclosure Agreements, a Heilongjiang Boss: Questions Keep Piling Up in the Ledger Theft Case and Hardcore Guide: How to Turn Your iPhone Into a Secure Cold Wallet.
Weekly headline catch-up
In its weekly headline recap, MarsBit grouped several policy and macro developments, including the market reaction to U.S. September CPI, where BTC was described as testing $80,000 and U.S. semiconductor stocks dropped sharply; a court rejection of an application to revoke a $5 million settlement agreement, with a judge saying Gemini and the Commodity Futures Trading Commission lacked sufficient grounds; and upcoming MSCI rule changes that may put Strategy and Metaplanet at risk of removal.
In the commentary and public statements category, the roundup cites Ray Dalio’s warning that the U.S. Treasury bear market is “far from over” and that supply-demand imbalances continue to build; Bank of America’s view that a Federal Reserve rate hike in October could become the biggest positive for risk assets and U.S. equities; Goldman Sachs’ warning that deleveraging may only be beginning after hedge funds sold in one week what amounted to two years of net buying; Microsoft’s CEO saying the age of superintelligence will require a controllable architecture and that responsibility cannot be outsourced to AI; and Jensen Huang backing CoreWeave while calling the idea that AI compute demand comes from only two labs a major misunderstanding.
Under institutions, large companies and top projects, the roundup notes the launch of UMX under Li Lin.
It also updates the Ledger hardware-wallet theft case, saying the affected devices all came from CryptoBilis and that Ledger’s own systems and official direct-sales channels were not affected. In another security-related item, Blur’s front end is set to be formally discontinued, and Yuga Labs’ vice president of blockchain urged users to revoke Blur contract approvals immediately.
The article closes by linking to the Weekly Editor’s Picks archive and signing off until the next issue.


