BTC $80,520.45 -1.29%
ETH $2,363.23 -1.56%
BNB $730.94 -2.04%
XRP $1.28 -3.21%
SOL $101.42 -2.92%
TRX $0.3300 +0.01%
DOGE $0.0775 -3.36%
ADA $0.2081 -3.92%
BCH $276.85 -2.38%
LINK $11.61 -3.95%
HYPE $75.54 -3.33%
AAVE $140.28 -5.24%
SUI $0.9458 -4.77%
XLM $0.1697 -3.91%
ZEC $1,154.26 -6.22%
AAPL $346.91 -0.38%
AMZN $248.12 -0.88%
GOOGL $320.93 -0.84%
MSFT $522.62 +0.68%
META $641.91 -0.78%
NVDA $211.43 -0.95%
TSLA $371.12 -1.16%
SNDK $1,236.70 -4.13%
INTC $86.58 +0.13%
SPCX $147.06 -0.71%
MU $882.18 +1.24%
AMD $516.96 +0.77%
BTC $80,520.45 -1.29%
ETH $2,363.23 -1.56%
BNB $730.94 -2.04%
XRP $1.28 -3.21%
SOL $101.42 -2.92%
TRX $0.3300 +0.01%
DOGE $0.0775 -3.36%
ADA $0.2081 -3.92%
BCH $276.85 -2.38%
LINK $11.61 -3.95%
HYPE $75.54 -3.33%
AAVE $140.28 -5.24%
SUI $0.9458 -4.77%
XLM $0.1697 -3.91%
ZEC $1,154.26 -6.22%
AAPL $346.91 -0.38%
AMZN $248.12 -0.88%
GOOGL $320.93 -0.84%
MSFT $522.62 +0.68%
META $641.91 -0.78%
NVDA $211.43 -0.95%
TSLA $371.12 -1.16%
SNDK $1,236.70 -4.13%
INTC $86.58 +0.13%
SPCX $147.06 -0.71%
MU $882.18 +1.24%
AMD $516.96 +0.77%

eva

All
Article
Flash

first_img Evan Conrad: Most AI startups cannot purchase computing power

Evan Conrad published an article on October 22 titled "The end of startups," with the subtitle "The coming credit crunch." Evan Conrad stated that within the next six months, most AI startups will be unable to purchase computing power, as global computing power is sold out. To obtain capacity, new clusters need to be contracted, and it is becoming increasingly difficult for non-investment-grade entities to secure online capacity.Evan Conrad mentioned that startups are not investment-grade, thus they are required to pay larger upfront payments. He has seen cases where 100% of the contract amount is required to be prepaid, as lenders and cloud providers are unwilling to take on additional credit risk. He believes that investment-grade tenants are scarce, and due to leveraging and selling to non-investment-grade token buyers, their ratings are downgraded, which he refers to as a credit bottleneck. He anticipates that investment-grade GPU buyers will increasingly demand that token buyers themselves be investment-grade or pay large upfront deposits.In the article, he provides an example where, at the time of writing, B300 could be purchased for $6 per hour and then sold for approximately $30 per hour for inference. He expects GPU prices to rise, and the gross margin for inference layers to decline, making it harder for clients with poor credit to obtain tokens. He proposed that lenders treat GPU clusters as collateral and stated that the physical delivery forward market SF Compute could allow lenders to use GPU hours on a liquidity book as collateral, rather than relying on a single service provider's cluster.

first_img Mysten Labs appointed Kevin Boon as CEO, and Evan Cheng remains as CTO and leads the AI department

Kevin Boon, President of Mysten Labs, has succeeded Evan Cheng as the CEO of this Sui Layer 1 blockchain developer. Cheng will continue to serve as Chief Technology Officer and personally lead the newly established internal AI business unit. In an open letter released on October 16, Cheng stated that Boon has accepted the position and announced the news simultaneously on X. Boon previously served as Chief Legal Officer at Mysten Labs before being promoted to President. Cheng mentioned, "I planned for this day when I appointed him as President."The new AI department will have independent functions such as engineering and product development to reduce external dependencies, and will focus on trust proxy products and AI-native infrastructure, referencing the model of the on-chain financial business unit established by Mysten this summer. Cheng stated that this transition is intended to free up more time for him to focus on technical leadership while continuing to support the Sui and Walrus ecosystems, "This makes my scope of responsibilities more manageable."The goal of the AI business is to allow proxies to autonomously execute tasks within user-defined limits, with related products and services anchored to Sui and Walrus. Walrus is Mysten's decentralized storage protocol, which is set to launch its mainnet in March 2025. Cheng specifically mentioned the Verifiable Agent Arbiter developed in collaboration with Google Cloud, which was announced on October 6, aimed at allowing enterprises to prove that their AI agents operate within authorized boundaries. As designed, detailed proxy records are stored in the customer-controlled Google Cloud Storage, with cryptographic proofs stored on Walrus and coordinated on Sui, with initial enterprise deployment plans set to launch ahead of broader availability.

Trump plans to transfer senior officials from the AI testing agency to the White House, with a 40-person team evaluating top models

The Trump administration plans to transfer three senior officials from the U.S. artificial intelligence model testing agency CAISSI to the White House, reporting directly to White House technology policy director Michael Kratsios and national cyber director Sean Cairncross. The government had previously considered moving the entire agency to the White House but ultimately only adjusted part of the senior leadership; the related transfers have not yet been completed, and the identities of the three officials have not been disclosed. Since its establishment by the Biden administration, the agency has used three different names in less than two years, and Chris Fall, the head, resigned just three months after taking office in July of this year, with a successor yet to be determined.CAISSI is affiliated with the National Institute of Standards and Technology (NIST) under the U.S. Department of Commerce, has about 40 employees, and an annual budget of approximately $15 million, responsible for model safety assessments. According to informed sources, the agency is collaborating with the National Security Agency to test OpenAI's Astra and Anthropic's Mythos before public release; it has previously assessed DeepSeek V4 Pro and Zhiyuan GLM-5.3. Google confirmed that Gemini 4 Argon has participated in the voluntary testing process before the U.S. government release. The White House hopes that this adjustment will allow model testing experts to be more directly involved in policy decision-making.

WSJ: Iran's sanctions evasion network transfers funds through Binance, with a total transaction amount involved reaching 850 million USD

According to a report by the WSJ, a previously undisclosed internal investigation report from Binance shows that Iranian financier Babak Zanjani used affiliated companies, shell companies, and multiple trading accounts to transfer funds through Binance, establishing a funding channel for the Iranian Islamic Revolutionary Guard Corps (IRGC) to evade Western sanctions. The related network involves 21 individual and corporate trading accounts, with a total trading volume of approximately $850 million, of which at least $67 million flowed from two accounts to wallets associated with the IRGC.The report indicates that Tajik businessman Sukhrob Oimakhmadov, classified as a "high-value customer" by Binance, received over $10 million through his account, with sources including digital wallets later identified by the Israeli Ministry of Defense as belonging to the IRGC. The cryptocurrency platform Zedcex, associated with Zanjani, had a total trading volume close to $830 million, with approximately $56 million withdrawn to related wallets. The network also involves the Turkish payment company ZedPay, as well as the cryptocurrency token USDZ used for trading.Binance stated that it has identified the related activities and taken measures, with all accounts associated with sanction violations being restricted from trading and gradually closed. The company also emphasized that the total trading volume of accounts does not equal the actual funds flowing into the IRGC and acknowledged past procedural issues in internal approvals and execution timeliness. The U.S. Department of the Treasury had previously imposed sanctions on Zanjani and related companies and individuals.

first_img Robinhood Chain evaluates the paid trading ranking system, with technical support from Arbitrum

According to CoinDesk, the Layer 2 blockchain Robinhood Chain, designed for tokenizing real-world financial assets, is evaluating a transaction sorting system that allows users to pay fees for prioritizing specific transactions. This technology is developed by Arbitrum, the infrastructure provider for Robinhood Chain. Insiders say that Arbitrum's previously launched paid transaction sorting technology, Time Boost, could provide users with a 200-millisecond prioritization advantage, but it was replaced on September 24 by the new Priority Gas Auctions model, which allows traders to pay higher fees to prioritize individual transactions. This new version is what Robinhood Chain is currently evaluating.Currently, Robinhood Chain employs a first-come, first-served transaction sorting strategy and has not yet utilized Arbitrum's original Time Boost application. Since launching its Ethereum-compatible network in July, Robinhood Chain has rapidly gained popularity and has entered the top ten blockchains ranked by total locked value. Robinhood's traditional brokerage business has long profited through "payment for order flow," where market makers pay brokers for the routing rights of customer orders.The mechanism of on-chain paid transaction prioritization operates differently; it provides advantages to traders by accelerating transaction processing speeds, helping professional institutions compete for fleeting trading opportunities.
app_icon
ChainCatcher Building the Web3 world with innovations.