MARA CEO: AI power use delivers far higher returns than Bitcoin mining, with crypto miners’ future core competitiveness shifting to power resources.

Beating AI Express News: P Equity Research analyst Mr. P stated in a September 26 podcast that the widely cited claim that hyperscale cloud providers have "10-year demand visibility" is unconvincing, noting cloud firms can barely predict demand changes two years out, and AI infrastructure investment will ultimately remain constrained by cyclical spending. In the near term, AI inference demand will continue driving growth in demand for storage products including HBM, DRAM, and NAND. Mr. P projects next year’s capital expenditure (capex) by hyperscale cloud providers will reach $1.1 trillion to $1.2 trillion, with storage spending accounting for 50% to 60% of that total—equivalent to roughly $500 billion to $700 billion; UBS even forecasts this figure could hit $900 billion. Mr. P added that the current AI computing power bottleneck has shifted from a simple shortage of GPUs to other areas: power, advanced packaging, storage, and ABF substrates. Older-generation GPUs like the H100 still maintain high secondary market prices, while rental costs for B-series GPUs continue to rise. Supply tightness for ABF substrates is expected to persist until 2028 to 2030 or later, and orders for gas turbines from makers including Mitsubishi, Siemens, and GE Vernova are already booked past 2030. On the data center interconnect front, Mr. P predicts copper cables and optical communications will coexist for years to come. NPO is likely to see expanded adoption starting in 2027, while CPO may begin ramping up in 2028 to 2029, though mainstream adoption will not arrive until after 2030.

Goldman Sachs believes the market may currently be overpricing stagflation and U.S. Treasury yield upside risks. As tariff impacts fade, energy prices potentially decline, and AI technology drives cost reductions, U.S. inflation pressure is expected to ease. Meanwhile, while economic growth may slow, corporate core earnings remain resilient. In this "Goldilocks" scenario, enthusiasm for AI investment could reheat, and the year-end U.S. stock market rally may not need to wait until after the U.S. midterm elections to launch. Goldman Sachs partner Mark Wilson stated that recent market moves have shown related signs: AI-related assets, after months of consolidation, are once again attracting capital. Goldman Sachs economist Jan Hatzius noted that upside risks to U.S. economic growth are easing; as fiscal stimulus effects fade, gasoline prices and mortgage rates rise, economic growth may slow further, which will also limit room for global central banks to continue raising interest rates. Goldman Sachs U.S. Strategy Team Head Ben Snider added that while some sectors have posted temporary "excess profits," corporate core earnings are likely to sustain strong growth at least through the end of 2027. Based on this analysis, Goldman Sachs argues that if inflation continues to cool, economic growth slows moderately, and corporate earnings stay resilient, the market may gradually shift from its previous stagflation trade to the "Goldilocks" scenario.

Ethereum co-founder Vitalik Buterin has released a new article titled "The Cryptographic World Computer", stating that after years of technological evolution, Ethereum will gradually become a system with fundamental architectural differences from traditional blockchains. Vitalik points out that Ethereum already has core capabilities including general-purpose computing, proof-of-stake (PoS), zero-knowledge proofs, and layer-2 (L2) scaling. Going forward, it will further introduce adjustable computing architectures, multi-party block construction, highly optimized PoS, and deep integration of zero-knowledge proofs into its base layer. He predicts that by around 2030, Ethereum’s core mechanisms—transaction validation, consensus, block construction, privacy, and computing—will all undergo changes. These shifts include moving away from re-downloading and executing blocks toward SNARK-based validation and PeerDAS for data availability verification; transitioning from proof-of-work (PoW) to highly optimized PoS; and replacing single-party block construction with multi-participant block building. Vitalik notes that the future Ethereum decentralized network will not only serve security and censorship resistance, but also enhance performance through parallel storage, parallel computing, and privacy protection. The Hegota upgrade is likely Ethereum’s last "traditional" hard fork, after which recursive STARKs, automated formal verification, optimized consensus mechanisms, and anti-quantum technologies will become the main development directions. Ultimately, these advances will drive Ethereum to evolve from a mere ledger to a "cryptographic world computer" that integrates blockchain, cryptographic privacy, verifiable computing, and decentralized off-chain components.

IOSG’s recent statistics show that as of August 11, 2026, Binance has delisted 42 spot tokens and 28 USDT-margined perpetual contracts so far this year. The number of delisted spot tokens exceeds any full year since 2022, with an average delisting announcement batch every 28 days, a significant acceleration from the 52-day interval in 2025. The delisting logic for spot and derivatives segments has diverged: the median lifespan of delisted spot projects rose from 4.1 years in 2022 to 5.1 years in 2026, while that of delisted contracts dropped from 1.3 years to 0.8 years; among this year’s delisted contracts, 23 were launched in 2025. IOSG analysis notes that fully diluted valuation (FDV) and open interest (OI) are better indicators of delisting risk than trading volume. In 2026, the delisting rate for spot tokens with an FDV below $10 million hit 49%, while no tokens with an FDV over $100 million have been delisted; the delisting rate for contracts with OI below $1 million is 31%, and no contracts with OI over $20 million have been delisted. In contrast, trading volume has a weaker correlation with delisting risk, as some contracts with high daily trading volume have still been delisted. Additionally, Binance’s in-house launch channels do not guarantee long-term listing: 63% of this year’s delisted contracts came from Binance Alpha, and 11 of the 42 delisted spot tokens were issued via Launchpool or Launchpad.

According to HTX market data, Bitcoin has rebounded to surpass $85,000, posting a 1.04% gain in the past 24 hours.

Hyperliquid announced in a statement that, based on user feedback, it will lower the funding rate cap from 4% per hour to 0.5% in its upcoming network upgrade. The official noted that this cap was rarely hit in actual operations. Additionally, per developer feedback, Hyperliquid will adjust deployment limits for HIP-4 outcome contracts: the maximum number of active outcome contracts a single deployer can hold simultaneously will be increased from 100 to 200, while the daily deployment cap for outcome contracts will be raised from 500 to 1000.