← 返回 Avalaches

随著人工智慧热潮席卷全球市场,华尔街大型退休基金与机构投资者面临前所未有的系统性风险与多元化难题。以纽约市退休体系(NYCRS)为例,投资长 Monte Tarbox 近期拒绝了多档重度押注 AI 的基金,原因在于科技巨头不仅在公开股票市场占据庞大市值,更广泛渗透至私募股权、企业债与基础设施等各类资产,导致传统依赖资产类别进行分散投资的策略逐渐失灵。

数据显示,AI 基础设施与相关企业已占标准普尔 500 指数市值的近 40%,并占据了今年大量投资级债券发行与创投资金,甚至出现企业间互相资助的循环融资现象。摩根士丹利与摩根大通的策略模型均指出,多资产投资组合与 AI 风险因子呈现高度正相关,一旦当前狂热的资本支出周期见顶或无法变现,市场将面临估值回调的巨大隐忧,7 月晶片股抛售与避险基金的亏损即为警讯。

为了应对缺乏标准定义的 AI 曝险,包括 CalPERS、LACERA、澳洲 Aware Super 以及 Marsh 等大型机构开始转向「整体投资组合法」(Total Portfolio Approach),利用量化工具与情境分析全面监控跨资产的相关性与 AI 依赖度。尽管许多基金主管担忧过度限制 AI 部位可能错失高额报酬,但为了保障数百万人的退休安全,机构内部正加速提高监控频率与设定曝险上限,力求在成长动能与结构性风险之间取得平衡。




As the artificial intelligence boom sweeps global financial markets, major pension funds and institutional investors on Wall Street are grappling with an unprecedented systemic risk that challenges traditional diversification. For instance, Monte Tarbox, chief investment officer at the New York City Retirement Systems (NYCRS), recently passed on AI-heavy funds, warning that the technology's pervasive reach across public equities, private equity, corporate debt, and infrastructure has turned conventional asset-class diversification into an illusion.

Data indicates that AI infrastructure firms and hyperscalers now make up roughly 40% of the S&P 500's market cap, while driving a major share of investment-grade debt issuance, venture capital funding, and controversial circular financing deals. Cross-asset models from JPMorgan and warnings from Morgan Stanley Wealth Management emphasize that institutional portfolios show strong positive correlations to AI risk factors, stoking fears that excessive infrastructure spending and inflated valuations could trigger severe market fallout once the cycle turns, as hinted by the July tech rout.

To better quantify undefined AI exposures, institutional allocators like CalPERS, LACERA, Australia's Aware Super, and Marsh are adopting the Total Portfolio Approach (TPA) and advanced analytical frameworks to assess cross-asset vulnerabilities and theme-level correlations. While portfolio managers acknowledge that prematurely capping AI holdings could drag down overall returns during an extended rally, protecting millions of pensioners requires tightening risk guardrails and continually recalibrating their exposure against an evolving technological wave.
2026-09-22 (Tuesday) · 5c9c82c44cab4ba478c0cddf31e75eb11d750447