经济学家警告：一个AI泡沫已破裂

Core Viewpoint - The article discusses the potential unsustainability of the AI-driven performance surge in the tech sector, highlighting concerns about a possible bubble in AI valuations and the overall economic environment [2][5]. Group 1: AI Valuation and Market Trends - Nvidia reported a fourth-quarter revenue of $68.1 billion, a 73% year-on-year increase, but concerns arise about the sustainability of such growth [2]. - As of fall 2025, there are 498 AI unicorn companies with a total valuation of $2.7 trillion, including 100 companies established in 2023 or later [3]. - Over 1,300 AI startups have valuations exceeding $100 million, with OpenAI's valuation reaching $730 billion, up from $500 billion just six months prior [4]. Group 2: Market Corrections and Sector Challenges - The tech sector is experiencing a "return to reality," particularly in the SaaS industry, where companies like Salesforce and ServiceNow have seen market caps shrink by approximately 30% since the beginning of the year [4]. - The semiconductor industry is also showing signs of slowdown due to high demand leading to chip shortages and geopolitical tensions affecting supply chains [4][6]. Group 3: Economic Risks and Profitability Concerns - Higgins suggests that a rare type of bubble may exist within the fundamentals of profitability itself, questioning how long the current growth can last [5]. - AI demand may fall short of initial expectations, making it difficult for tech companies to absorb heavy investment costs, with Goldman Sachs estimating AI capital expenditures to reach $539 billion by 2026 [5]. - Ongoing geopolitical issues, such as the Iran conflict affecting helium supply, could further destabilize the economy and impact companies reliant on AI for profitability [6].