Summary of Key Points from the Capex Tracker Industry Overview - The Capex Tracker indicates a trend of capital expenditure (Capex) cuts across various industries, with notable exceptions in Technology and Utilities [3][4]. Core Observations - General Industrial Capex is projected to have a compound annual growth rate (CAGR) of 5.5% for the period 2024-2028, which is a slight decrease of 0.4 percentage points compared to the previous update in July [3][4]. - Positive growth in Capex is observed in the following sectors: - **Datacenters**: 26.5% CAGR, an increase of 3.5 percentage points from July [4]. - **Pulp & Paper**: Improvement noted, but specific growth figures not provided [3]. - **Conventional Power Generation**: Positive outlook with companies like Wartsila and Accelleron showing growth [3]. - **Mining**: Companies such as Epiroc and FLSmidth are expected to benefit [3]. - Conversely, significant declines are noted in: - **Vehicles/Autos**: Negative growth, with a decrease of 1.8 percentage points to 2.0% CAGR [4]. - **Pharma and Biotech**: Both sectors are experiencing negative trends, with Biotech showing a decline of 8.6% [4]. Detailed Capex Growth by Sector - **Datacenters**: - 2025 Capex growth projected at 51.7%, a significant increase of 15.2 percentage points [4]. - **Renewables and T&D**: - 2025 Capex growth at 17.5%, down by 8.0 percentage points [4]. - **Semiconductors**: - 2025 Capex growth at 15.7%, a decrease of 1.2 percentage points [4]. - **Healthcare**: - 2025 Capex growth projected at 0.0%, indicating stagnation [4]. - **Consumer Sector**: - 2025 Capex growth at 0.7%, reflecting a decline of 1.1 percentage points [4]. Additional Insights - The Capex Tracker highlights a robust growth trajectory in Datacenters, Renewables, and Mining, while traditional sectors like Vehicles and Pharma are facing headwinds [4]. - The report emphasizes the importance of monitoring these trends for potential investment opportunities and risks in the respective sectors [3][4]. Conclusion - The Capex Tracker serves as a critical tool for understanding industry trends and making informed investment decisions, particularly in identifying sectors poised for growth versus those facing challenges [3][4].

Summary of Key Points from the Conference Call Industry and Companies Involved - **Industry**: Data Center and Cloud Services - **Companies**: Alibaba, GDS Holdings, VNET Group Core Insights and Arguments 1. **Alibaba's AI Investment Commitment**: Alibaba's CEO, Eddie Wu, emphasized a commitment to AI investments, with a target of Rmb380 billion over three years, and plans to increase global data center capacity by 10 times by 2032 to support anticipated global AI investments of US$4 trillion in the next five years [1][3] 2. **Expansion Plans**: Alibaba plans to build data centers in Brazil, France, and the Netherlands, while expanding capacity in Mexico, Japan, Korea, Malaysia, and Dubai [1] 3. **Beneficiaries of Alibaba's Investments**: GDS Holdings is identified as a key beneficiary of Alibaba's increasing AI investments, particularly in overseas and domestic data center build-outs. VNET is also expected to benefit from domestic orders [1] 4. **Potential Catalysts for Growth**: The anticipated B30 server updates and shipments in Q4 2025 and Q1 2026, along with new orders in Q4 2025, are seen as positive catalysts for GDS and VNET [1] 5. **Valuation Metrics for Alibaba**: The target price for Alibaba's H-shares is set at HK$183, based on a 12x P/E on FY2026E Ecommerce Group net profit and various P/S ratios for other segments [3] 6. **Risks for Alibaba**: Key risks include failure in executing retail strategy, investment spending pressures, user traffic slowdowns, integration risks from acquisitions, economic slowdowns, and regulatory risks [4] 7. **Valuation for GDS Holdings**: The target price for GDS is set at US$51.2 per share, based on a SoTP valuation of 15x FY26E EV/EBITDA for GDS China and 23x for DayOne [5] 8. **Risks for GDS Holdings**: Risks include revenue fluctuations, potential over-capacity due to government policies, execution risks in delivering data center pipelines, increasing competition, rising utility costs, and interest rate hikes impacting earnings [6] 9. **Valuation for VNET Group**: The target price for VNET is set at US$20.0, based on 16x 2026E adjusted EBITDA, reflecting a premium above the historical average [7] 10. **Risks for VNET Group**: Risks include slower-than-expected demand recovery, intensified competition affecting margins, and lower-than-expected pricing impacting cash flows [8] Other Important but Potentially Overlooked Content - The conference highlighted a broader trend in the industry where peers may follow Alibaba's lead in AI investments, potentially leading to accelerated order growth and stock re-ratings across the sector [1] - The report emphasizes the importance of monitoring macroeconomic conditions and regulatory environments, particularly in the context of US-China trade relations and their impact on the Chinese economy [4]

Summary of Key Points from the Conference Call Industry Overview - The focus of the conference was on the **Technology sector**, particularly the **AI investment cycle** and its implications for economic growth and stock valuations [7][4]. Core Insights and Arguments 1. **Federal Reserve Rate Cuts**: - The recent CPI inflation report supports the Fed's potential to cut rates, with expectations of a 25 basis point cut five times by June 2026 [2][3]. - The labor market's softening is a significant factor driving the Fed's decision to cut rates, with a forecasted GDP growth of only 1.3% for 4Q25/4Q24 [2][3]. 2. **Market Reactions to Fed Actions**: - Historical trends indicate that US stocks typically rise when the Fed cuts rates amid economic growth [3][9]. - However, there are risks associated with a potential economic slowdown, which could counteract the benefits of rate cuts [3][9]. 3. **AI Investment Cycle**: - The conference highlighted a sustained investment cycle in compute power to support AI applications, with major companies like NVDA, GOOGL, and MSFT indicating a strong AI environment [7]. - Despite a slight slowdown in AI adoption in 3Q25, compute investment remains robust, suggesting a positive impact on potential GDP growth in the US [7]. 4. **Market Performance Metrics**: - As of September 12, 2025, the S&P 500 showed a year-to-date gain of 12.0% and a 1.7% weekly change, indicating a generally positive market sentiment [13]. 5. **Concerns and Risks**: - Investors are worried about the return on investment (ROI) from AI, the possibility of not achieving a soft landing for the US economy, and the effects of a steepening yield curve on the economic impact of Fed rate cuts [9][3]. Additional Important Content - The conference served as a platform for public and private investors to discuss the future of technology, emphasizing the importance of AI in driving economic growth [7]. - Upcoming economic data releases, including Retail Sales and Industrial Production, are critical to watch as they may influence market sentiment and Fed policy [11]. This summary encapsulates the key points discussed in the conference call, focusing on the implications of Fed policy, the AI investment cycle, and the overall market performance.

Summary of Key Points from the Multi-Industry Capex Tracker Industry Overview - The report focuses on the multi-industry capital expenditure (capex) trends, analyzing approximately €3.1 trillion of capex across around 4,000 companies in 26 different end markets [1][7][43]. Core Insights - **Medium-term Capex Growth**: The medium-term capex growth has been revised upward slightly, now indicating a 5.9% compound annual growth rate (CAGR) for 2024-2028, which is a 0.4 percentage point increase from the previous update in April [7][13]. - **2025 Capex Downgrade**: For 2025, the capex growth expectation has been downgraded by 0.4 percentage points to 4.9%, attributed to some spending being deferred to 2026 due to tariff and policy uncertainties [7][8][13]. - **Sector Performance**: - **Strong Growth Areas**: Technology (specifically Datacenters) and Utilities (Power grids, Renewables) are expected to see favorable capex environments, with companies like Legrand, Schneider, Siemens, Prysmian, and Nexans recommended for investment [7][26]. - **Weak Growth Areas**: Sectors such as Biotech, Pulp & Paper, and REITs are experiencing the largest cuts in capex growth [7][8]. Additional Insights - **Geographical Drivers**: The US remains the primary driver of strength in the capex tracker, despite uncertainties regarding tariffs and interest rates. Europe is also expected to see strong investments in grid and defense [8][9]. - **Capacity Utilization**: Capacity utilization is below historical median levels in Europe but aligns with historical medians in the US and China. Elevated asset age in sectors like Oil & Gas, Vehicles, Airlines, and Healthcare indicates a need for replacement capex [8][9][21]. - **Macro Risks**: Key risks to the capex estimates for 2025 include tariff-driven inflation, structural overcapacity in China, and competitiveness issues for core capex spenders in Europe [9][10]. Sector-Specific Capex Growth Projections - **Datacenters**: Expected to grow at a CAGR of 23.0% for 2024-2028, with a 36.5% growth forecast for 2025 [20]. - **Renewables and Transmission**: Projected CAGR of 16.3% for 2024-2028, with 25.5% growth in 2025 [20]. - **Semiconductors**: Anticipated CAGR of 12.9% for 2024-2028, with 11.2% growth in 2025 [20]. - **Biotech**: Expected to decline with a CAGR of -8.1% for 2024-2028, and a -10.7% growth forecast for 2025 [20]. Conclusion - The capex tracker indicates a mixed outlook for various sectors, with technology and utilities poised for growth while others like biotech and pulp & paper face challenges. The overall capex environment remains healthy, but macroeconomic uncertainties could impact future investments [7][9][10].

Investment Rating - The industry view for Greater China Telecoms is rated as Attractive [2][78]. Core Insights - The datacenter industry is currently experiencing a multi-year upcycle, driven primarily by inference demand, which is expected to significantly influence future capital expenditures [6][7]. - The top six companies in the datacenter sector are projected to increase their capital expenditures by 62% year-over-year, reaching approximately Rmb373 billion [9]. - Datacenter electricity consumption is anticipated to account for 2.9% of total electricity usage in China [13]. Summary by Sections Datacenter Cycle - The industry is in a multi-year upcycle, with strong demand expected to drive capital expenditures [6][7]. Capital Expenditure Forecast - Significant growth in capital expenditures is expected, with a forecast of Rmb373 billion for the top six companies, marking a 62% increase year-over-year [9]. Datacenter Orders and Capacity - Datacenter orders are forecasted to reach approximately 3.7GW over the next three years, indicating robust demand [11]. - New supply was limited during the previous downcycle, leading to a more rational approach to new third-party wholesale capacity [15][20]. REITs and Valuation - Rapid progress in REITs issuance is noted as a key driver for valuation and funding within the industry [20][23]. - A dividend yield of 5-6% is expected at issuance, with trading multiples potentially exceeding the issuance [25]. Regional Market Insights - Various regional markets, including Hong Kong, Greater Tokyo, and Singapore, show differing dynamics in terms of market size, return profiles, and future supply trends [28]. - Finland is emerging as a key secondary market in Europe, benefiting from favorable energy resources and a mix of green energy capacity [39][41]. Financial Forecasts - GDS Holdings is projected to achieve an EBITDA of approximately Rmb6,740 million by 2027, with a consistent growth trajectory in net revenues [49]. - The utilization rate for datacenter area is expected to increase to 81% by 2027, reflecting improved operational efficiency [50].

Summary of Key Points from the Conference Call Industry Overview - The discussion centers around the impact of large-scale AI training workloads on the power grid, particularly focusing on the challenges faced by data centers and the potential risks of power outages due to rapid fluctuations in power demand [3][4][5][6][9][32]. Core Insights and Arguments 1. **Power Grid Stress**: The increasing demand from multi-gigawatt-scale data centers is stressing the century-old power grid, which was not designed to handle the unique load profiles of AI training workloads [3][4][5]. 2. **Load Fluctuations**: AI training workloads can cause instantaneous power consumption fluctuations of tens of megawatts, which can lead to significant challenges for grid stability [4][5][20]. 3. **Risk of Blackouts**: The worst-case scenario involves potential blackouts affecting millions of Americans if the power grid cannot cope with the rapid load changes from AI data centers [3][4][5]. 4. **Engineering Solutions**: Engineers have created temporary solutions like dummy workloads to smooth out power draw, but these can lead to annual energy expenses in the tens of millions [5][6]. 5. **Battery Energy Storage Systems (BESS)**: Tesa's Megapack system is highlighted as a leading solution for managing power quality issues in data centers, capable of rapid charging and discharging to respond to load fluctuations [6][67][69]. 6. **Demand Response Programs**: Participation in demand response programs can help data centers manage peak loads, but challenges remain in implementation and utility-side management [78][81][86]. 7. **Cascading Failures**: The risk of cascading blackouts is significant if large amounts of load disconnect from the grid simultaneously, as seen in previous incidents [38][56][65]. Additional Important Content 1. **Grid Design Considerations**: The discussion includes insights into the fragility of voltage and frequency in electric systems, emphasizing the need for a stable balance between supply and demand [10][13][15]. 2. **Historical Context**: The Texas winter freeze of 2021 is cited as an example of how extreme conditions can lead to significant grid failures [14][15]. 3. **Future Projections**: There is a forecast of over 108GW of large loads, primarily from data centers, looking to connect to the ERCOT grid, which exceeds the US's peak load of 75GW [28][31]. 4. **Technological Innovations**: The rise of new technologies, such as the 800V DC architecture, is expected to impact the supply chain and improve the management of power fluctuations in data centers [107]. This summary encapsulates the critical points discussed in the conference call, focusing on the implications for the power grid due to the demands of AI training workloads and the potential solutions being explored.

Investment Rating - The report maintains a constructive outlook on datacenter operators, indicating they can sustain profitability levels above historical norms, despite tempered AI demand expectations [5]. Core Insights - The global datacenter supply/demand model has been updated, indicating a loosening of supply constraints earlier than previously expected, with peak occupancy now forecasted for 2025 instead of 2026 [3][13]. - Occupancy rates are projected to remain above historical levels, with a gradual loosening expected from 2026 through 2027, stabilizing around average levels seen over the past 18 months [3][45]. - Demand growth forecasts have been adjusted, particularly for AI workloads, reflecting a more gradual pace of AI training demand and a net decrease in incremental demand over the next 18 months [12][15]. Supply and Demand Overview - The current global datacenter market capacity is estimated at approximately 63 GW, with a significant portion owned by hyperscaler and cloud providers [25]. - By 2030, the total datacenter capacity is expected to reach approximately 131 GW, translating to a 6-year CAGR of ~14% [35]. - The report highlights that the mix of datacenter workloads will shift further towards cloud workloads, with hyperscale and wholesale datacenters expected to comprise about 75% of the total by 2030 [35]. Demand Forecast - The global datacenter market demand is estimated to grow by ~50% to 86 GW by 2027, with AI workloads increasing to 27% of the overall market [15]. - AI workload demand is projected to grow at a 38% CAGR, while traditional corporate workloads are expected to grow at a modest 3% [20]. Supply Forecast - The report indicates that supply sufficiency is expected to decrease by an average of 3% in 2025, 7% in 2026, and 6% in 2027, with a long-term forecasted supply sufficiency of 86% by 2030 [13]. - The datacenter supply model reflects a historical increase of ~2 GW due to actual capacity increases and adjustments to historical supply tracking [12]. Power Demand and Sustainability - Datacenter power demand is projected to increase by ~160% by 2030 compared to 2023 levels, contributing approximately 1% CAGR to overall US power demand [75]. - The report anticipates that 40% of the datacenter power increase will be met with renewables, with the remaining 60% expected to be driven by natural gas generation [76]. Grid Investments - The report estimates that approximately $720 billion will be required for grid investments through 2030, primarily focused on distribution and transmission upgrades to support datacenter growth [68].