电力与算力成为新的硬通货，中国将迎来电力超级周期

Core Insights - Wall Street's valuation model for tech companies has shifted from focusing on user numbers and growth rates to assessing the amount of H100 GPUs and stable clean power supply [1][8] - Amazon announced a significant layoff of 30,000 employees, the largest since late 2022, not due to performance issues but to reshape its business structure [2][8] - Amazon's CEO Andy Jassy stated that the company plans to invest approximately $125 billion in capital expenditures by 2025, primarily directed towards AI-related data centers, power, and chips [2][8] Group 1: Market Trends - The trend of prioritizing "power + computing" as a competitive advantage is becoming evident in the capital markets, with significant stock price increases for companies like Nvidia and energy firms [5][8] - The global AI competition has shifted from a "chip shortage" to a "power shortage," highlighting the physical energy requirements of AI technologies [7][8] Group 2: Energy Demand Projections - According to UBS, China's electricity demand is expected to grow at an annual rate of 8% from 2028 to 2030, driven primarily by the explosion of AI data centers [12][13] - The report predicts that AI data centers will contribute 2.3 percentage points to this growth, making it the largest single driver [13] - The demand for electricity from electric vehicles (EVs) is also projected to grow significantly, with a compound annual growth rate (CAGR) of 33%-39% from 2025 to 2030 [14][15] Group 3: Structural Drivers of Growth - Three structural drivers are identified for the increased electricity demand: the growth of AI data centers, export-driven demand, and accelerated electrification [12][14] - The report suggests that the previous reliance on traditional industries for electricity demand is shifting towards technology innovation and industrial upgrades [15] - UBS has raised its forecast for China's new installed capacity during the 14th Five-Year Plan by 14% to 438 GW, indicating a potential "super cycle" in electricity demand [15]