橡树资本共同创办人霍华德·马克斯警告，人工智慧可能带来可怕的就业前景，数以百万计的劳工被取代，而收益却集中在少数受过高等教育、住在美国沿海地区的多重亿万富豪手中。他担心，这种由 AI 推动的极端财富集中，会加深原有的社会与政治分裂，让社会更容易受到民粹煽动与强烈反扑的冲击。

在企业端，马克斯指出，一场赢者全拿的 AI 军备竞赛正在上演，微软、Alphabet、亚马逊、Meta 和甲骨文等巨头投入巨额资金并承担激进负债，以阻止规模较小的竞争对手追上来。华尔街今年已安排超过 1610 亿美元、与美国资料中心相关的融资交易，放款人则收紧契约结构，以防 AI 泡沫形成或需求成长远弱于预期时自身蒙受损失。

他指出，投资人为了支持 AI 晶片与资料中心，愿意在 30 年期公司债上只多拿约 100 个基点、相对同年期公债的利差，却必须承受长达三十年的技术不确定性，以及资产在债务偿清前就丧失生产力的风险。马克斯建议投资人既不要在 AI 上孤注一掷，也不该完全置身事外，因为这场革命一方面可能弥补到 2035 年前数以百万计婴儿潮世代退休造成的劳动力缺口，另一方面也可能释放出不可逆且难以预测的巨大变化。

Howard Marks of Oaktree warns that artificial intelligence could create a terrifying labor outlook, with millions of workers displaced while a small group of highly educated coastal multibillionaires reap the gains. He fears this extreme concentration of AI-driven wealth will deepen existing social and political divisions, leaving societies vulnerable to populist demagoguery and backlash.

On the corporate side, Marks highlights a winner-take-all AI arms race in which giants like Microsoft, Alphabet, Amazon, Meta and Oracle spend vast sums and load up on aggressive debt to prevent smaller rivals from keeping up. Wall Street has already arranged more than $161 billion of US data-center related credit deals in 2025, and lenders are tightening structures to shield themselves if an AI bubble forms or demand growth proves weaker than expected.

He notes that investors are accepting roughly 100-basis-point spreads over Treasuries on 30-year bonds used to fund AI chips and data centers, despite facing three decades of technological uncertainty and the risk those assets lose productivity before the debt is repaid. Marks advises investors not to go all-in or all-out on AI, arguing that the revolution could both offset labor shortages from millions of retiring baby boomers through 2035 and unleash irreversible, unpredictable changes.