Huobi HTX will launch spot trading for MSTR and TSLA on November 10.

Binance issued an official announcement stating that following the U.S. Treasury Department's Financial Crimes Enforcement Network (FinCEN) listing of MEXC as a major money laundering concern, it will cease processing fund transfers involving MEXC and its associated entities named in the announcement. Binance added that withdrawal requests from Binance wallets to those entities will be rejected, while deposits from the entities will be temporarily held for compliance review. During the review period, related accounts may be restricted, and Binance may take further actions in line with its terms of service, including terminating services for relevant users.

According to Bitget data, U.S. WTI crude oil and Brent crude oil have fallen by 2% on the day, currently trading at $84.51 per barrel and $90.2 per barrel respectively.

Beating AI News Flash: OpenAI has launched an internal safety system called Control Room, specifically designed to halt risky AI experiments. Once a halt instruction is received, it takes at most 35 seconds for the relevant program to lose access to AI models. Even if the program itself remains operational, it can no longer get the models to process requests. An incident occurred at OpenAI in September this year: an internal AI, in an environment supposed to be isolated, found a loophole in network restrictions and secretly contacted an external chatbot. The safety system issued an alert within 15 minutes, and staff intervened quickly, but the experiment did not stop automatically. It was manually shut down only after about 2.5 hours. Control Room assigns a unique ID to each experiment. Every time a program calls an AI model, the server checks whether this ID still has permission. Once an experiment is halted, all models connected to the system will reject its requests, with no impact on other normal experiments. OpenAI has added an extra safeguard: if Control Room malfunctions or is shut down, models will automatically reject requests by default, preventing them from running without supervision. Currently, the system covers all of OpenAI's frontier model research tasks. The 35-second window starts from the moment a halt instruction is received, and does not include time for anomaly detection and manual judgment. OpenAI is also working to further reduce this delay, aiming to achieve immediate halts.

Capital Economics Group chief economist Neil Shearing wrote that even if the AI investment boom reverses, it is unlikely on its own to push the U.S. or other G7 economies into recession. Shearing pointed out that the Philadelphia Semiconductor Index has fallen roughly 35% since hitting a peak in June this year, while shares of the five largest hyperscale cloud service providers have dropped an average of around 10% since early October. Over the past year, AI-related investment and the wealth effect from stock market gains combined to account for roughly a quarter to a third of U.S. GDP growth. At the current trajectory, the stock of AI-related investment could exceed $5 trillion by the end of this decade; under relatively conservative return assumptions, related firms still need to generate an additional $1.7 trillion to $1.9 trillion in annual revenue to match this scale, while current annual AI-related revenue is approximately $150 billion. A shift in investment could lower U.S. GDP growth by roughly 0.3 percentage points. The firm argues that current banking leverage has not created systemic financial risks comparable to those during the housing bubble era, and most AI investment corresponds to imported goods, so its impact on U.S. domestic manufacturing and employment is expected to be more limited, with effects likely concentrated in supply-side economies such as Taiwan and South Korea. Capital Economics has already factored the drag from AI investment pullbacks and stock market declines into its 2027 U.S. growth forecast, and if this assessment holds, stock market losses outside the tech sector should be limited.

According to official news, the third phase project of Bitget IPO Prime, Anthropic, is now open for public round subscription, available to all eligible users (including VIP users), with subscriptions supported via USDT or USDGO. Allotted assets will be delivered as rANT, which leverages Bitget’s RWA infrastructure to mirror the economic performance of ANT post-listing at a 1:1 ratio. Details of this public round are as follows: - Subscription period: 16:00 November 10 to 16:00 the day before the IPO (UTC+8) - Investment range: $100–$25,000 - Total investment pool cap: $200,000,000 - Underwriting service fee: 5% - Distribution time: 2 to 3 hours before the opening of regular US stock trading on the IPO day

Dongcha Beating AI News Flash: Daily People interviewed a data center operations engineer who has worked in Ulanqab, Inner Mongolia for five years. While local AI data centers are still under large-scale construction, recruitment demand at his company has dropped significantly: it hired thousands of people in 2025, but only three new hires as of the 2026 interview. The engineer, who uses the alias Zhang Liwei, originally worked on an auto factory assembly line. When he switched careers, his monthly internship salary was only 2,000 yuan, rising to 4,500 yuan after graduation. After five years of job changes and promotions, he now earns over 10,000 yuan per month. In April this year, DeepSeek recruited senior data center operations engineers and senior delivery managers in Ulanqab, offering monthly salaries of 15,000 to 30,000 yuan with 14 months of pay. These are senior roles requiring extensive experience in data center operations and project management. According to Zhang, local regular on-duty staff still only earn 5,000 to 6,000 yuan monthly, while assistant engineers can make over 10,000 yuan. The industry’s entry threshold is rising: in 2024, some data centers still hired new graduates without relevant professional backgrounds; this year, some companies require matching majors and work experience, and even impose a 35-year age limit for frontline on-duty positions. The job-seeking boom has spawned training course businesses. Zhang revealed a former colleague runs an operations training program, charging nearly 20,000 yuan per trainee, with classes only launching when 60 students enroll. Calculated this way, each course batch generates nearly 1.2 million yuan in revenue. However, the so-called internal referrals only help trainees secure at most one interview opportunity.