From Concentration to Selectivity: Why We Reduced Legacy AI Exposure and Prepared a Broader July Portfolio

On Friday, 17 July 2026, Final Resurrection Ltd. implemented a targeted portfolio restructuring designed to reduce exposure to positions showing weaker relative strength while preserving the strongest long-term holdings in artificial intelligence infrastructure. The purpose was not to abandon the AI investment theme, but to improve capital efficiency, reduce reliance on a small number of technology stocks and prepare the portfolio for a more balanced final phase of July.

The technology sector has delivered exceptional long-term opportunities, but periods of rapid expansion frequently produce significant differences between individual companies. Some businesses continue to demonstrate strong earnings growth and market leadership, while others experience slower momentum, changing expectations or prolonged recovery periods.

Against this background, the portfolio review conducted on 17 July focused on three questions: which positions still offered the strongest short-term opportunity, which holdings were absorbing capital without producing sufficient momentum, and where newly available capital could potentially achieve a better balance between growth and risk.

The first decision was to reduce exposure to Oracle and Marvell Technology. Both companies remain strategically relevant to the broader digital infrastructure market, but their positions had developed into substantial portfolio laggards. Partial sales were therefore executed during market strength rather than through indiscriminate selling into weakness.

Oracle shares were sold in two tranches as the stock recovered, while additional shares remained subject to a higher staged selling level. Marvell was also reduced in stages, with the first sale completed and further orders positioned above the market. This approach was designed to release capital while retaining some participation should the recovery continue.

The portfolio also fully exited a leveraged Micron-related call warrant. Leveraged products can amplify positive movements, but they also introduce time decay, volatility risk and a greater dependence on short-term market timing. Closing the position removed a source of accelerated risk and converted the remaining value into deployable capital.

Importantly, the restructuring did not involve increasing the already substantial holdings in Arista Networks or NVIDIA. Both remain central long-term positions, but adding further capital would have increased concentration in the same AI infrastructure theme. Instead, the next allocation phase was directed toward companies representing different sources of earnings growth.

Three candidates were selected for staged purchase orders:

  • Palo Alto Networks, representing cybersecurity and AI-related digital protection;
  • Danaher, representing life sciences, diagnostics and high-quality industrial healthcare;
  • Dollar Tree, representing a consumer-sector recovery and defensive retail exposure.

The orders were deliberately placed in several price stages below prevailing market prices. This avoids paying any price merely to become invested and allows the portfolio to benefit from ordinary market pullbacks.

The broader lesson from the 17 July restructuring is that successful portfolio management is not simply a matter of identifying attractive companies. It also requires controlling concentration, reassessing opportunity cost and directing capital toward the strongest combination of momentum, quality and diversification.

The strategy for the remainder of July is therefore based on selective participation rather than maximum exposure. Strong AI holdings remain in place, weaker positions are being reduced in an orderly manner, leveraged risk has been lowered and new capital is being prepared for deployment across cybersecurity, healthcare and consumer retail.

Leave a Comment

Your email address will not be published. Required fields are marked *