Last week, funds have flowed into #Bitcoin, #Ethereum, and #Hyperliquid.

According to Hong Kong Exchanges and Clearing (HKEX) filings, Alibaba Group chairman Joseph Tsai purchased an additional 720,000 Alibaba shares on Tuesday, spending HK$82 million, following his initial purchase of the same number of shares. After Alibaba announced its HK$80 billion new share placement plan, Tsai and CEO Eddie Wu made combined share purchases totaling approximately HK$120 million yesterday, signaling confidence in the company’s AI strategy. Wu bought 350,000 Alibaba Hong Kong-listed shares at an average price of around HK$111.6, costing roughly HK$40 million. Together, the two acquired 1.07 million shares for a total of about HK$120 million. All proceeds from Alibaba’s HK$80 billion new share placement will be fully invested in full-stack AI capabilities and AI infrastructure construction. The offering was actively subscribed by long-term investors including global sovereign wealth funds, ultimately achieving nearly 3 times oversubscription.

BitMEX founder Arthur Hayes is once again bullish on ENA: Over-the-counter brokers have started reaching out to us to inquire about borrowing U.S. dollars. Interest rates are still too low, but this is a good sign that basis trading is making a comeback, and ENA will benefit significantly with massive upside potential. Earlier reports noted that Hayes accumulated 22.64 million ENA tokens at the start of the month, totaling approximately $2 million. He also stated earlier this month that if increased U.S. dollar liquidity drives Bitcoin higher, a rebound in Bitcoin basis returns could draw funds back into USDe, adding that ENA has the potential to rally fivefold in the coming months. According to HTX market data, ENA is down 7.1% in the past 24 hours, trading at $0.15.

Beating AI Express News: Alibaba Cloud’s Token Plan is now officially integrated with Qianwen App and its PC version. After users bind the Token Plan’s dedicated API Key in the "Work Assistant", they can directly use their subscription quota for complex tasks, including operating computers and browsers, calling Skills, and creating Office documents, applications, and websites. This effectively extends the same model quota from development tools to Qianwen. The Token Plan was already compatible with AI programming and Agent tools such as Codex, Claude Code, Cursor, Qoder, and OpenClaw; now, it is also accessible for use with Qianwen’s Work Assistant. Supported models include Qwen3.8-Max, among others, with all models and tools charged uniformly via Credits.

Per Lookonchain’s monitoring, a crypto whale who had remained dormant for four years has sold 1,400 Bitcoin, valued at $111.61 million. The whale purchased 2,200 Bitcoin four years ago at an average price of $45,024, totaling $99.05 million. Notably, the whale did not sell even when their unrealized profits peaked at $178 million, but has recently begun offloading Bitcoin, and has now realized a profit of $76.5 million.

CryptoQuant analyst Axel Adler Jr. stated that after a pullback, Bitcoin has reclaimed its short-term holder cost base of roughly $68,200. Meanwhile, with rising trading volume, the price has broken through the SMA200 (200-day simple moving average) level of around $69,000 and SMA111 of approximately $67,500. A large number of holders who bought recently are back in profit, easing the selling pressure from short-term holders who were in the red. However, Bitcoin’s moving average alignment still maintains a bearish structure: SMA111 sits below SMA200, and both are lower than SMA365’s roughly $83,200. This indicates the current move is merely a strong short-cycle recovery, not a confirmed reversal of the broader trend. The most critical upside resistance is the SMA365 level at $83,200. If the price stalls and pulls back at this level, the existing bearish structure will hold; if it can effectively reclaim it, this could mark the first serious evidence of a trend reversal. Key downside support lies in the $67,000 to $69,000 range, which encompasses both the short-term holder cost base and the dense zone of SMA111 and SMA200. As long as the pullback holds this range, short-term holders remain profitable, and the price stays above the moving averages, the market structure still holds upside potential; a break below would resume downside risks. The current market is leaning positive but has not yet fully entered a high risk appetite mode.

Data from Bank of America (BofA) shows that active long-term funds significantly reduced their holdings of global semiconductor stocks last month, with sell-offs totaling around $44.4 billion, indicating institutional capital is exiting the most crowded AI trades. Capital flows have instead shifted to sectors including telecom, energy, materials, and grid modernization, reflecting a more pronounced reallocation within the AI theme. This data explains some of the recent pressure in the market. Ahead of NVIDIA’s earnings report, the market still holds high expectations for AI demand, but chip stocks have rallied sharply and become heavily concentrated in positions. Should long-term interest rates rise, AI revenue expectations cool, or returns on cloud providers’ capital expenditures come under question, the semiconductor sector will be the first to face selling pressure. BofA also noted that the top themes funds have sold over the past year include AI computing and quantum computing, indicating capital has not fully exited AI but is reducing exposure to highly crowded segments. BofA forecasts that chip stocks will remain driven by NVIDIA’s earnings, cloud providers’ guidance, and interest rate movements in the short term; over the medium term, capital may be more willing to allocate to sub-sectors that can benefit from AI infrastructure spending, such as power, equipment, networking, and storage.