Bank of America: Three key thresholds for US stocks to trigger a "full risk-off" this summer, with signals for these conditions already building up.

According to official announcements, ForeGate Supercomputing Database, in collaboration with Michael Owen, OKX, WEEX, and OnebullEX, has officially released the *ForeGate 2026 World Cup Winning Guide* mid-tournament research report. The report integrates ForeGate’s AI prediction model and OKX.ai’s data path simulation to systematically analyze the 48 participating teams in the 2026 FIFA World Cup (co-hosted by the U.S., Canada, and Mexico), covering their qualification odds, advancement paths, championship win probabilities, win/draw/loss trends, and handicap betting tendencies. The model currently boasts an overall prediction accuracy rate of 93.8%, demonstrating strong advantages in match outcome responsiveness, points calibration, and potential matchup path judgment. ForeGate stated that during the World Cup, it will continuously update its predictions based on match results, team form, and model simulation outcomes to help users understand the probabilistic logic behind schedule changes. Additionally, ForeGate’s million-dollar World Cup prediction campaign is ongoing, where users can participate in match predictions to split the $1 million prize pool.

During the weekend when traditional stock markets are closed, Trade.xyz—dubbed the "on-chain Nasdaq"—enables continuous trading and real-time price discovery via perpetual contracts, capabilities traditional finance cannot provide, pricing in advance the upcoming Monday’s U.S. and South Korean stock market moves. Popular U.S. stock assets on Trade.xyz have posted mixed performance compared to their Friday after-hours levels, and are expected to consolidate with minor fluctuations ahead of Monday’s pre-market. Their weekend performance is as follows: SpaceX is trading at $154.3, versus $152.769 in U.S. Friday after-hours trading; Micron Technology (MU) is at $1138.2, versus $1133.5 in Friday after-hours; SanDisk (SNDK) is at $2102.7, versus $2109 in Friday after-hours; NVIDIA is at $193.63, versus $192.71 in Friday after-hours; Marvell Technology (MRVL) is at $270.28, versus $265.248 in Friday after-hours; Intel is at $128.13, versus $127.62 in Friday after-hours; Alphabet (Google) is at $340.48, versus $336.15 in Friday after-hours; AMD is at $519.36, versus $518.7 in Friday after-hours. Popular South Korean stock assets on Trade.xyz saw slight gains on Saturday, and are expected to rise around 2% at Monday’s opening. Their weekend performance is as follows: Samsung Electronics is at $224.11, up from $221.17 in Friday’s closing price; SK Hynix is at $1786.1, up from $1741.37 in Friday’s closing price.

According to monitoring by Beating, a16z co-founder Marc Andreessen noted that many AI practitioners and industry insiders consider Zhipu GLM-5.2 to be the first Chinese AI model that can match or even outperform open models from leading U.S. labs on most tasks, while also being balanced across multiple capability dimensions. This development carries "extremely critical timing significance" amid accelerating global AI competition, as large model capabilities are gradually shifting from being dominated by a small number of U.S. labs to a multipolar competitive landscape. Click the original link below to join Beating’s Feishu AI News Channel, which provides 24/7 uninterrupted monitoring of global AI hotspots and news.

Michael Saylor, founder of MicroStrategy, has once again shared updates on Bitcoin Tracker, remarking, “We are gonna need more charts.” Per historical trends, MicroStrategy typically announces increases to its Bitcoin holdings the day after such statements.

Bank of America Securities chief strategist Hartnett outlined three triggers for a "full risk-off" environment this summer in his latest Fund Flow Report: Mag7 ETF falling below $60, USD/JPY dropping below 110, and the yield curve inverting again. None of these conditions have been met yet, but signals are building. U.S. equity funds have posted a net outflow of $8.5 billion, the first such outflow since March, following a historic $119.2 billion net inflow. The divergence where hyperscale cloud providers continue to underperform chip stocks has put the sustainability of AI capital expenditure at the center of market debate: Apple’s MacBook price hikes and Microsoft’s Xbox price increases are both directly tied to rising memory costs. Vera Rubin rack memory prices have surged 435% cumulatively, and Goldman Sachs projects AI capital expenditure could reach as high as $1.4 trillion by 2027. Hartnett’s core ongoing question is: How much further do cloud providers need to fall before the market begins pricing in capital expenditure cuts? U.S. equity funds have already shifted ahead of time; liquidity flowing out of tech giants is pouring into cyclical assets like semiconductors, small-caps, housing, and REITs, a move the market interprets as a front-run bet on policy shifting toward "affordability." At the asset class level, Hartnett believes gold still holds strong allocation value below $4,000, and going long on long-dated U.S. Treasuries is the most contrarian long-term trade right now. The U.S. dollar is only suitable for short-term holding rather than long-term allocation, and his strategic view is to go long on emerging markets over the long term. Since Fed Chair Waller took office on May 22, U.S. Treasuries have risen 3.2% cumulatively, while stocks have fallen 1.6%, with bonds significantly outperforming.

Samsung Electro-Mechanics is in final negotiations with a major U.S. tech firm over a supply contract for MLCCs (multi-layer ceramic capacitors) for AI servers. Worth around 500 billion won, the deal equals 10% of its component division’s annual revenue last year, making it an extremely large-scale order. Industry sources speculate the client is a leading player with significant clout in the data center sector. Each AI server requires 15,000 to 25,000 MLCCs—over 10 times the number used in a smartphone—with a unit price more than triple that of smartphone-grade MLCCs, earning the component the nickname "golden rice of the AI industry". Samsung Electro-Mechanics currently holds the No.2 position in the global MLCC market with a 20%+ share. The contract is expected to significantly boost its standing in the AI server MLCC market. Industry observers note the deal essentially secures quality validation from a global "big buyer", which will prompt other major tech firms to follow suit. Turning to its glass substrate business, Samsung Electro-Mechanics will formally sign a joint venture agreement with Japan’s Sumitomo Chemical this week. The two firms will invest a combined 500 billion won to set up a glass substrate JV, with Samsung holding a majority stake and contributing around 300 billion won. The JV will be based at the Pyeongtaek plant of Sumitomo Chemical’s South Korean subsidiary Dongwoo Fine-Chem, with production scheduled to launch in early next year. Glass substrates, which offer superior heat resistance to traditional plastic substrates and can accommodate more HBM (high-bandwidth memory) and GPUs, are hailed as "rule changers" for AI semiconductor packaging. Samsung Electro-Mechanics aims to lock in a high-quality supply chain in advance for mass production, while Sumitomo Chemical is using this partnership to enter the next-generation packaging materials market.