Will Morgan Stanley and JPMorgan Chase Go Long Together? AI Narrative Begins to Shift, Will Taiwan Stocks See a V-Shaped Recovery? Which Targets Should Be Noticed?

Morgan Stanley's recent report significantly raised the capital expenditure forecasts for the world's top five hyperscale data center companies, estimating $1.2 trillion and $1.4 trillion for 2027 and 2028, respectively. Morgan Stanley's optimism stems from the belief that hyperscale data centers have entered a phase where capital expenditures yield actual returns and cash flow. Morgan Stanley outlines three different frameworks for return on invested capital (ROIC): 1. Infrastructure Leasing (estimated ROIC 25%-40%): Building AI data centers and leasing computing power, represented by companies like Microsoft, Amazon, Google, and CoreWeave. 2. Proprietary AI Models and Infrastructure (estimated ROIC >40%): Developing large AI models while owning AI data centers, represented by companies like OpenAI, Google, Microsoft, and Meta. 3. AI Models Operating on Third-Party Infrastructure (estimated ROIC >25%): AI model developers leasing AI data centers, represented by companies like Anthropic, Mistral AI, and Moonshot AI. More investment strategies and stock selection information can be found at: LINE@ https://user225916.pse.is/money668 Channel link: https://www.youtube.com/@money66868. Last week, Amazon's CEO, Jassy, expressed a similar view in the earnings report, predicting that AWS server investments would recover costs within two to three years, with a server lifespan of five to six years. This shows that hyperscale data centers can recover capital investments in the short term, which is why they are willing to increase capital expenditures. Meanwhile, JPMorgan Chase's latest report points out that hyperscale data center companies currently have backlog orders totaling $1.7 trillion, with a year-on-year increase of over 150%, far exceeding the 80% year-on-year growth rate of capital expenditures. The scale of backlog orders can support future high capital expenditures, and the performance of hyperscale data center companies relative to the S&P 500 index has fallen to the bottom of the three-year range. After global large-scale deleveraging, the 'AI stock god' Leopold's hedge fund (Situational Awareness) was liquidated, and the credit default swaps (CDS) of hyperscale data center companies hit a new high. The market's investment logic and AI narrative have begun to shift. Morgan Stanley believes that AI infrastructure providers and AI model developers will reap substantial returns. JPMorgan Chase believes that the backlog orders of AI infrastructure providers can support future high growth, and their stock prices are still at relatively low levels. After the global capital cleansing, the large capital expenditures of hyperscale data centers may affect the narrative of capital returns, shifting to the phase where hyperscale data centers enter the capital return period. U.S. stock funds have begun to flow significantly into the AI supply chain. Taiwan's stock market has shifted from a U-shaped bottom to a V-shaped recovery. Although Taiwan's stock market has risen nearly 5,000 points from last week's low, many AI semiconductor and key component stocks recommended to investors have also surged. It is reminded that many AI component leading stocks are still at relatively low levels; memory and CPO supply chains are also still at low levels; July's high-growth revenue stocks and key stocks in earnings calls are all industries and individual stocks to watch. What other targets should be noticed? As follows: Taiwan's stock market V-shaped recovery, the low point is not waiting. Do you want to participate in the subsequent market? Join https://user225916.pse.is/money668 and send a private message [+88] to receive a limited-time project. Advanced Packaging: Lumileds (3711-TW), Force (6239-TW), Greatek (3481-TW), Wanrun (6187-TW), Wanshi (6223-TW), Changxing (1717-TW). AI Outstanding Stocks: Trade Link - KY (3665-TW), Delta Electronics (2308-TW), Lite-On Technology (2301-TW), Chuang Lake (2059-TW), Qixuan (3017-TW), Jiance (3653-TW), Weiying (6669-TW), Hon Hai (2317-TW). PCB: Nanke (8046-TW), Xinxiang (3037-TW), Zhen Ding - KY (4958-TW), Taitien (2383-TW), Taishao (6274-TW), Nan Ya (1303-TW), Jinju (8358-TW). More industry trends preview + intraday information → All in Teacher Ye's Line@: LINE@ID: @money668 (remember to add @) https://user225916.pse.is/money668 After joining, you can leave a message [《鉅亨網》] to let me know where you know me from, and I will prioritize notifying you of preferential activities or gifts! Article source: Moore Securities - Analyst Ye Junmin. The individual securities recommended by our company for analysis have no improper financial interest relationship. Past performance does not guarantee future profits. Investors should make independent judgments, carefully evaluate, and bear investment risks.