在 Nvidia 执行长 Jensen Huang 的领导下，该公司正经历著非凡的财务成长，预计 2027 年营收将达到 5000 亿美元。这令人震惊的数字，接近 2023 年全球半导体产业的总营收。Nvidia 高达 75% 的利润率产生了庞大的现金流，Huang 并未将这些资金用于股票回购，而是积极投资于各项交易与合作伙伴关系，旨在消除扩展人工智慧（AI）的任何潜在障碍。
为了解决阻碍资料中心发展的土地、电力和基础设施等瓶颈，Nvidia 正积极推动资本市场的参与。一项重要举措是与华尔街大型投资公司达成协议，提供 5000 亿美元的融资以加速 AI 基础设施的建设。此外，Nvidia 也参与了在俄亥俄州建设大型资料中心园区的计划，该园区将由 OpenAI 租用并由 SoftBank Group 附属公司营运。虽然这些大胆的财务承诺旨在推动 AI 成长，但 Bank of America 的分析师警告，如果 AI 需求减缓，Nvidia 的成长率和资产负债表可能会面临压力。
目前基础设施的高估值部分源于由制造瓶颈造成的关键半导体稀缺。尽管 Nvidia 投入了大量资金，但它仍依赖 Taiwan Semiconductor Manufacturing Co. (TSMC)、Samsung 等代工厂的产能。一个关键的未知数是，如果这些制造商大幅增加产量，缓解供应链限制，将会发生什么事。这引发了关于当前极高的零件需求有多少是受实际 AI 终端使用量驱动，又有多少只是由于担心错失良机（FOMO）而产生的恐慌性购买，进而对长期投资价值的稳定性提出质疑。
Under the leadership of CEO Jensen Huang, Nvidia is experiencing extraordinary financial growth, with projected revenues reaching $500 billion by 2027. This staggering figure is comparable to the total revenue of the entire global semiconductor industry in 2023. Driven by a massive 75% profit margin, this generated cash is not being used for stock buybacks, but rather aggressively invested by Huang into a continuous stream of deals and partnerships aimed at eliminating any barriers to the expansion of Artificial Intelligence (AI).
To address bottlenecks in land, power, and facilities hindering data center development, Nvidia is actively engaging capital markets. A key initiative involves an agreement with major Wall Street investment firms to provide $500 billion in financing to accelerate AI infrastructure construction. Furthermore, Nvidia is involved in backing a massive data center complex in Ohio, slated to be leased by OpenAI and operated by a SoftBank Group affiliate. While these bold financial commitments aim to catalyze AI growth, Bank of America analysts warn that a potential slowdown in AI demand could put pressure on Nvidia's growth rate and balance sheet. (Key numbers: 5000)
The current high valuation of infrastructure is partly driven by the scarcity of critical semiconductors, a result of manufacturing bottlenecks. Despite its massive investments, Nvidia remains reliant on the production capacity of foundries like Taiwan Semiconductor Manufacturing Co. (TSMC) and Samsung. A critical unknown is what happens if these manufacturers significantly increase output, easing supply constraints. This raises questions about how much of the current sky-high demand for components is driven by actual end-use of AI versus panic buying fueled by fear of missing out (FOMO), casting doubt on the stability of long-term investment values.