Vol spike hits quant funds
A sudden volatility regime shift has whipsawed systematic shops such as Citadel and Point72, exposing ML models trained on unusually tranquil data and prompting renewed focus on robustness, stress testing, and out-of-sample validation. Firms are revisiting model governance and hiring profiles to bridge research and production resilience. (alpha-maven.com)
Millennium Management reported roughly $1.5 billion in losses for the week through March 6, Citadel’s Wellington fund fell about 2% that same week, and Point72’s platform declined roughly 1.1% in the March 1–6 selloff tied to strikes on Iran. (bloomberg.com)) The Cboe VIX surged into the mid-30s before retreating, closing around 25.50 on March 10 and averaging about 25.09 for March 2026 as measured by monthly indices. (financialcontent.com)) Coverage framing the episode as a regime shift singled out systematic and multi‑strategy shops—calling attention to concentrated losses and the market dislocation that produced them. (alpha-maven.com)) Operational responses documented this month include renewed emphasis on formal AI/model governance and human-in-the-loop checkpoints as part of model rollout and monitoring programs for trading models. (arcesium.com)) Quant teams are reported to be adding technical layers such as regime‑switching early‑warning models and unsupervised clustering of volatility/correlation states to trigger out‑of‑sample checks and adversarial stress scenarios. (sciencedirect.com)) Hiring and org shifts reflect the pivot: Point72’s Cubist unit is actively recruiting quantitative researchers and machine‑learning specialists via its 2026 Quant Academy and job listings, while industry pieces note multi‑strategy firms are prioritizing cross‑discipline hires from academia and HFT to strengthen research-to-production pipelines. (job-boards.greenhouse.io))
