Norges Bank Investment Management (NBIM), which manages Norway’s $2 trillion Government Pension Fund Global, officially integrated Anthropic’s Claude AI into its core investment and compliance operations.
While the fund had been testing AI throughout 2025, the February 2026 deployment specifically targets the detection of corruption, fraud, and ethical violations across its massive portfolio of over 9,000 companies.
1. The “Invisible Pattern” Detector
The fund uses Claude to perform what CEO Nicolai Tangen calls a “24/7 Forensic Audit” of its global holdings.
- Corruption Red Flags: The AI parses news reports, regulatory filings, and leaked documents in dozens of languages to identify subtle correlations that humans often miss—such as a sudden change in a company’s local subcontractors in high-risk regions.
- Fraud Detection: Claude is connected directly to the fund’s data warehouses. It monitors for market manipulation (like front-running) and identifies “behavioral biases” in the fund’s own internal traders to ensure total transparency.
- Deep Research: Portfolio managers now use a “Conversational Data” model. Instead of waiting 48 hours for an IT team to pull a report, they ask Claude: “Find all companies in our emerging market portfolio that have had a 20% turnover in senior compliance staff in the last six months.”
2. ESG & Ethical Screening
The move to AI is also a strategic defense against increasing geopolitical pressure.
- The “Rigor” Shield: In February 2026, the fund faced backlash from the Trump administration over divestment from certain U.S. and Israeli firms. By using Claude to screen for human rights violations and environmental crimes, the fund aims to provide a “systematic, evidence-backed” rationale for its decisions, deflecting accusations of political bias.
- Scale: Claude enables the fund to scrutinize all 9,000 companies simultaneously. Previously, the Council on Ethics could only perform “deep dives” on a few dozen high-risk firms per year.
3. Performance & Productivity
The results of this “Agentic AI” transformation have been record-breaking for a public institution:
| Metric | Result (as of Feb 2026) |
| Productivity Gain | 20% increase across the organization. |
| Hours Saved | 213,000+ annual hours unlocked for high-value research. |
| Trading Costs | AI-driven efficiency has reportedly saved $100 million annually in trading costs. |
| Accuracy | Claude was chosen over other models for its low hallucination rate and its ability to signal “uncertainty” in financial data. |
4. “The Tangen Mandate”
The rapid adoption is driven by CEO Nicolai Tangen’s strict policy: “AI is not an option, but a must.” * The Career Link: Tangen reportedly told staff that AI proficiency is now a mandatory requirement for promotion within the fund.
- Internal Tools: The fund has deployed “Claude Code” for its technology teams and an internal “Investment Simulator” to help portfolio managers “dare to be contrarian” by identifying herd behavior in real-time.
Why it Matters: This represents the largest-ever enterprise deployment of AI in institutional finance. If successful, it could set the global standard for how sovereign wealth funds manage ethical risks in an increasingly polarized world.
