这份《FT News Briefing》1月2日（周五）的访谈回顾2025并展望2026，聚焦「美股七巨头」（Alphabet、Amazon、Apple、Meta、Microsoft、Nvidia、Tesla）在指数中的极端集中度。Gillian Tett 一年前预测它们不会崩跌，但也难以再大幅齐涨；核心理由是历史上过高集中往往会回撤，而AI热潮对获利的落地速度仍充满不确定。

2025年的实际表现呈现明显分化：Alphabet 上涨超过60%，Nvidia 上涨超过30%，Tesla 约上涨20%；其余多数成员涨幅低于60%，Amazon 甚至只有个位数。Tett 将Alphabet的反弹与「约18个月前」市场认为其在AI竞赛落后、但后来研发实力回证相关，并指出投资人开始更挑剔地判断谁会是赢家，因为「不太可能全部都赢」。

对2026的风险与催化因素，她提出三条主线：第一，AI 资本支出愈来愈多由「债务」而非自由现金流支撑，若利率上升将冲击估值；第二，政治反弹可能升温（例如部分民粹对所谓“broligarchs”的不满）；第三，若更便宜、耗能更低的替代路线（含中国的开源取径与各种低成本方案）跃迁，可能颠覆七巨头的既有框架。她认为要再现2024式的同步大涨不易，除非非科技公司出现可量化的AI生产力/营收提升，且美国能在电网与输电（而不仅发电）上完成大规模升级。

In the Financial Times’ Jan 2 (Friday) episode, Sonja Hutson interviews Gillian Tett about the “Magnificent Seven” (Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, Tesla) and whether their dominance can persist into 2026. Tett’s prior view was that the group would not collapse, but would struggle to keep rising in lockstep, because market concentration in major indices had reached historically extreme levels and AI hype was running ahead of proven earnings impact.

The 2025 scorecard showed sharp dispersion rather than uniform gains: Alphabet rose more than 60%, Nvidia more than 30%, and Tesla around 20%, while several others lagged and Amazon delivered only single-digit gains. Tett links Alphabet’s rebound to perceptions roughly 18 months earlier that it was losing the AI race, followed by evidence of deep research capacity and a more cautious rollout. The broader trend, she argues, is a more selective investor mindset: history makes it unlikely that every member will be a lasting winner.

For 2026, Tett frames upside versus multiple risks. A key financial risk is higher interest rates because AI capex is increasingly funded by debt rather than free cash flow. A political risk is backlash against AI “elites,” even if top-level leadership is supportive. The largest competitive risk is platform leapfrogging—cheaper, less energy-hungry approaches (including open-source momentum in China) challenging proprietary bets, especially as the US faces electricity constraints for data centers. A synchronized surge like 2024 looks unlikely unless non-tech firms show meaningful, measurable revenue/productivity gains from AI and the US rapidly expands grid and transmission capacity.