Unitree Robotics rose 8.01% intraday, with a 24-hour trading volume of $21 million on Trade.xyz.

Recently, CZ gave an interview to PIVOT, a leading Japanese business media outlet, where he discussed his voluntary decision to plead guilty in the U.S. and pay a substantial fine. He stressed that he refused to be a fugitive or evade accountability. Binance only faced registration-related violations, with no fraud involved and no victims harmed, and users were consistently well protected. Serving a prison sentence was an unpleasant experience and a valuable lesson. CZ also noted that Bank of America (BoA) committed more serious violations recently but only received a $1.2 billion fine and no criminal prosecution. "The world is inherently unfair, but I’m no longer fixated on that; I’m focusing on moving forward and doing more positive work by learning from this experience."

According to CNN, two U.S. officials said U.S. troops in the Middle East are ready to launch a new, large-scale military operation against Iran within 96 hours if ordered by President Donald Trump. Sources noted that the U.S. military assets currently deployed in the region are sufficient to carry out potential strike missions. The new U.S. operations may no longer be limited to targeting Iranian military assets that can be quickly restored, such as radar systems, missile launchers, and other similar facilities. Such actions are often referred to as "lawn mowing"—a tactic designed to regularly weaken an adversary’s military capabilities. Since Trump halted full-scale combat operations earlier this year, the U.S. military has employed similar measures. One U.S. official stated, "This will be a large-scale operation." Current and former officials familiar with U.S. military internal combat planning discussions echoed this view. Some argue the question is not whether Trump will order an attack, but when. They believe the order could come at any time before the end of this year, but is more likely to be issued after the U.S. midterm elections. (Jin10)

Crypto influencer Kimchi’s X account bio was changed back to @pumpfun this morning. According to GMGN data, the previously fast-launched token Kimchi surged over 100% from a $6 million market cap, hitting $12 million before pulling back, and is now trading at $9.23 million with a 24-hour trading volume of $4.1 million. The collaboration between crypto influencer Kimchi and pump.fun has been a farcical mess. On October 20, the fast-launched token Kimchi quickly reached a $47 million market cap amid a series of positive updates, before plummeting. Market rumors claimed pump.fun paid 1.35 billion PUMP tokens (worth around $6.25 million) to secure Kimchi as its spokesperson. However, on the 23rd, Kimchi posted a QR code linking to another Kimchi token, which spiked to a $10 million market cap briefly before crashing to zero. During this period, multiple wallets bought KIMCHI at low prices in advance and sold after the price rose, reaping over $385,000 in total profits. Kimchi later removed the @pumpfun tag from their bio, only to add it back again this morning. BlockBeats reminds users that celebrity meme coins are heavily influenced by the actions of their promoters, exhibit extreme price volatility, and are prone to insider trading and front-running; users should exercise caution when investing.

AllianceDAO co-founder Wang Qiao stated he is not concerned about the crypto community’s ability to address future AI and quantum threats. He even asserted that within the next two years, the world will recognize that cryptocurrencies are far better prepared for such threats than other segments of the internet and traditional finance. For a decade, the crypto market has functioned as a trillion-dollar, irreversible "honeypot" (a highly attractive attack target), fostering a system far more battle-tested than almost any other on Earth, as well as an adversarial security culture.

Beating AI News (Dongcha): Nanbeige Lab, a subsidiary of BOSS Zhipin, has released the open-source model Nanbeige4.3-33B, a model focused on programming and office agents, supporting a 512K context window and licensed under Apache 2.0. The team compared it with models including Qwen3.8-27B, Meta Muse Glimmer 30B, and Gemma 4 31B. In the 20 tests released, Nanbeige ranked first in 17. It scored 80.9 on Terminal-Bench 2.1, outperforming Qwen’s 73.0; its DeepSWE score reached 65.5, while the competitor’s was only 42.5. It also led Qwen by 54.2 to 38.3 in the office automation test Automation Bench. However, Qwen still holds the lead in SWE-Bench Pro and GDPval. Notably, Nanbeige Lab previously focused on developing 3B-scale small models; its prior generation model Nanbeige4.2-3B improved the model’s effective computational depth without increasing parameters by reusing Transformer layers. This time, the team instead first trained a 33B large model, planning to use it as a "teacher" to distill capabilities such as programming and tool calling into the upcoming Nanbeige4.3-3B.

According to monitoring by Onchain Lens, the Ethereum Foundation has just sold ETH worth $6.42 million. The institution converted 2,785 ETH (approximately $6.42 million) to USDS via CoWSwap.