Key Takeaways
- • Crypto analyst Michaël van de Poppe has ceded active trading authority over his $80,000 altcoin portfolio to Claude AI, following a 50% loss from his original $160,000 investment.
- • Van de Poppe built a proprietary framework integrating Claude with technical indicators (sigma deviations, RSI triggers, Bitcoin correlation analysis) to generate automated buy and sell signals rather than discretionary trades.
- • The AI framework has already overruled van de Poppe's trading instincts, most notably directing him to hold a NEAR position representing 41% of his portfolio allocation until specific exit zones (€1.75–€1.85) are reached, despite his discomfort with the concentration.
- • Executed trades include a Renzo position sale at 7 cents, which van de Poppe characterized as favorable, demonstrating early proof-of-concept for the AI-driven approach.
- • This case illustrates institutional-grade adoption of large language models for quantitative portfolio management in alternative assets, moving beyond advisory commentary into autonomous execution frameworks.
Summary
Michaël van de Poppe, a recognized crypto analyst, has implemented an experimental framework delegating portfolio management authority to Claude AI following significant realized losses in his altcoin holdings. Van de Poppe's $80,000 current portfolio balance represents a 50% drawdown from his initial $160,000 capital deployment, prompting a fundamental shift in his investment methodology away from discretionary decision-making.
The framework integrates Claude AI with a custom technical analysis system built on sigma deviations, Relative Strength Index (RSI) triggers, and Bitcoin correlation coefficients to generate systematic buy and sell signals. Rather than using the AI model for general market commentary or advisory purposes, van de Poppe has architected a rules-based system where the AI actively generates and enforces specific position entry and exit parameters. This represents a move toward quantitative, algorithm-driven portfolio management in the altcoin space.
The most significant test of the system occurred when van de Poppe sought to liquidate a position, but Claude's framework analysis determined the asset had not reached its designated exit zone. Van de Poppe complied with the AI's directive to hold his NEAR position, which currently represents 41% of his total portfolio allocation. The framework's exit zone for NEAR is pegged between €1.75 and €1.85; the AI indicated that additional price appreciation levels remained before position trimming should commence. Van de Poppe acknowledged the psychological discomfort of maintaining such concentrated exposure to a single asset while deferring to algorithmic authority rather than his own market instincts.
The framework has executed at least one trade independently: a Renzo position sale at 7 cents, which van de Poppe assessed as a successful execution. This early result provides preliminary validation of the AI-driven approach, though the portfolio remains underwater from inception. The experiment reflects broader institutional interest in deploying large language models and AI systems for quantitative portfolio management and autonomous decision-making in alternative asset classes.
The framework integrates Claude AI with a custom technical analysis system built on sigma deviations, Relative Strength Index (RSI) triggers, and Bitcoin correlation coefficients to generate systematic buy and sell signals. Rather than using the AI model for general market commentary or advisory purposes, van de Poppe has architected a rules-based system where the AI actively generates and enforces specific position entry and exit parameters. This represents a move toward quantitative, algorithm-driven portfolio management in the altcoin space.
The most significant test of the system occurred when van de Poppe sought to liquidate a position, but Claude's framework analysis determined the asset had not reached its designated exit zone. Van de Poppe complied with the AI's directive to hold his NEAR position, which currently represents 41% of his total portfolio allocation. The framework's exit zone for NEAR is pegged between €1.75 and €1.85; the AI indicated that additional price appreciation levels remained before position trimming should commence. Van de Poppe acknowledged the psychological discomfort of maintaining such concentrated exposure to a single asset while deferring to algorithmic authority rather than his own market instincts.
The framework has executed at least one trade independently: a Renzo position sale at 7 cents, which van de Poppe assessed as a successful execution. This early result provides preliminary validation of the AI-driven approach, though the portfolio remains underwater from inception. The experiment reflects broader institutional interest in deploying large language models and AI systems for quantitative portfolio management and autonomous decision-making in alternative asset classes.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.