In brief: Millennium and Anthropic are developing a Claude-based digital risk analyst that supports risk managers in assessing positions, while keeping human sign-off and auditability as fixed control mechanisms.
Hedge fund Millennium Management is building an AI-powered risk analyst together with Anthropic that is meant to support risk managers in assessing positions across various asset classes. The solution complements the already widespread use of Claude and Claude Code within the company.
Anthropic is developing a “digital risk analyst” together with Millennium, one of the world’s largest alternative asset managers. This is an AI system that operates under the supervision of the company’s risk managers, delivers new risk assessments and forms assessments of risk exposure across asset classes. Claude and Claude Code are already widely used at Millennium – from trading desks to engineering to central business functions. The digital risk analyst specifically extends this use to the analysis of risk positions.
The system stores and retrieves information over time and uses reasoning capabilities to explain daily changes in the risk profile. These results are then validated and supplemented by Millennium’s human risk managers. According to Anthropic, the system logs its reasoning steps, tests actions in sandbox environments, and requires evaluation and sign-off by human subject-matter experts – an approach intended to ensure traceability and auditability. The risk analyst is being developed by Millennium’s own risk team together with Anthropic’s research and applied AI teams in the company’s internal AI lab.
For CTOs in the financial sector, the case reveals a pattern that extends beyond this single instance: companies with already established Claude Code usage – at Millennium spread across more than 340 investment teams – use internal AI labs to specifically test models against demanding specialist applications before productive workflows emerge. The combination of proprietary company data and Claude’s model capabilities forms the basis for domain-specific applications, while human decision-makers, according to the companies involved, remain central to the decision chain.
Vlad Torgovnik, Chief Information Officer at Millennium, emphasizes that human judgment remains at the center of decision-making. Belinda Neal, Managing Director Financial Services at Anthropic, points to the need for trust in AI systems in the complex, regulated environments of the financial sector. Peter Nolan, Head of Asset and Wealth Management at Anthropic, highlights that Claude can think through risk positions, explain daily changes, and carry learned insights forward into subsequent queries, with the aim of saving time on automated recommendations.
Source: claude.com · Published August 5, 2026
Lumi AI News — AI-assisted curation pursuant to Art. 50 EU AI Act. Paraphrasing and classification by Lumi News Pipeline v1.8.3.