投资者对"旧世界"资本密集型资产的重新定价，正从数据中心蔓延至更广泛的实体经济供应链。 华尔街见闻提及，2月24日，高盛研报提出AI时代资金正在涌向"重资产、低淘汰"（HALO）实体资产，比如 电网、管道、公用事业、交通基建及关键工业产能 。这些资产难以复制、物理壁垒高且不易过时。 （2月下旬以来HALO一篮子股票持续上涨） 与此同时，美国关税政策的变动也为实体资产交易提供了额外支撑。 高盛初步预测，最高法院裁决触发的关税政策调整，将使美国实际关税税率下降约100个 基点。 尽管关税政策走向仍存在不确定性，但对依赖实物进口的基础设施及工业企业而言，成本的缓释将进一步强化市场向实体资产倾斜的短期逻辑。 近期高盛策略师Chris Hussey强调，投资者正在给予传统资本密集型企业更高的估值，且向供应链上下游及更广泛的经济领域延伸。 高盛的Oppenheimer阐述称，这是互联网商业化约四分之一世纪以来，技术增长前景首次对数据中心、能源供应等实体资产形成高度依赖。 实体资产重估，从数据中心到整条基础设施链 这一轮资本密集型企业的重估，高盛认为背后逻辑是：淘金热中，最稳定的获益者往往是铲子的制造商。 一是资产重建成本 ...

Core Viewpoint - Investors are re-evaluating capital-intensive assets in the "old world," with a shift from data centers to a broader range of physical economy supply chains, driven by the AI era's demand for "hard assets" [1][4]. Group 1: Investment Trends - Goldman Sachs reports that capital is flowing towards "hard assets" such as power grids, pipelines, utilities, transportation infrastructure, and critical industrial capacity, which are difficult to replicate and have high physical barriers [1][4]. - Since late February, a basket of HALO stocks has been on the rise, supported by changes in U.S. tariff policies that may lower actual tariff rates by approximately 100 basis points [3]. - The market is increasingly favoring traditional capital-intensive companies, extending this trend to supply chain and broader economic sectors [3][5]. Group 2: Characteristics of HALO Assets - Goldman Sachs defines optimal HALO assets as those with high reconstruction costs, deep regulatory barriers, and long construction cycles, making them hard to disrupt or replace [5]. - The focus areas for these assets include power grids, pipelines, utilities, transportation infrastructure, critical machinery, and long-cycle industrial capacity [5]. Group 3: Impact on Soft Assets - While hard assets are gaining popularity, soft assets such as software, media, consulting, and certain financial sectors are under pressure, with significant declines in software and IT service stocks observed in February [6]. - Concerns about AI disrupting these sectors and enabling low-cost competitors are leading to a fundamental reassessment of their business models [6]. - However, not all software companies are equally affected, and those demonstrating a history of delivering high-quality AI outcomes may still be viable investments [6].

Core Insights - Meta has offered over $100 million in compensation packages to AI model developers, which has caused a significant stir in the tech industry. Although these salaries are typically paid out over several years, they are still notable enough to make headlines [2][6] - This move is part of a larger strategy, as Meta plans to invest between $66 billion and $72 billion in capital expenditures this year, with a substantial portion allocated to building AI infrastructure [2][6] - The compensation structure in AI companies contrasts sharply with traditional software startups, where 70% to 80% of budgets are typically spent on employee salaries. In AI model companies, salaries represent a smaller portion of overall costs [3][8] Investment Strategy - High salaries are not merely emotional expressions but are strategic investments. Companies building AI foundational models tend to have fewer employees but require high capital investment, creating a natural environment for offering substantial salaries [3][4] - Meta's approach is not unique; Netflix also follows a similar model, planning to invest $18 billion in content while maintaining a relatively low employee count of 14,000, allowing for higher-than-market salaries [3][9] Industry Dynamics - The compensation logic for AI companies has diverged from traditional businesses. Capital-intensive firms can leverage capital to attract talent, while labor-intensive companies must rely on a larger workforce to maintain operations [4][9] - Meta's AI training system is becoming a critical and costly part of its business, which includes various platforms like Facebook, Instagram, and WhatsApp. The rise of AI-generated content (AIGC) poses both a threat and an opportunity to the existing user-generated content (UGC) model [4][8] Competitive Landscape - Other platforms like TikTok and YouTube are also recognizing the potential of AIGC to reshape the social media landscape, leading them to invest heavily in AI strategies and high salaries for talent acquisition [4][9] - Hiring key talent not only secures future contributions but may also provide insights into competitors' technological strategies, making high salaries a rational business choice as long as they do not harm company culture [5][9] Historical Context - The trend of offering high salaries in capital-intensive industries is not new. For instance, Netflix's approach has fostered a unique corporate culture that emphasizes teamwork over familial bonds [3][9] - A decade ago, models were created to optimize the balance between hiring employees and purchasing GPUs for AI expansion, indicating a shift towards hardware-centric spending in the industry [5][9] Conclusion - The current landscape reflects a significant opportunity for those involved in AI development, as they are positioned at a pivotal moment in technological evolution, deserving of fair compensation for their contributions [10]