Hyperscalers have committed to more than $600B in capex spending in 2026, with more than $1T expected in 2027. While much of this is going to hardware and chips, a large portion is going to the necessary real assets and infrastructure that are proving to be the true bottleneck of AI: Electrification infrastructure, energy, and raw materials. Rising electricity demand is not just owed to data centers, but the appetite for electricity from AI data centers will require a multi-decade investment in the generation, transmission, and resources. Most cap-weighted large-cap portfolios are inherently concentrated in the companies driving rising electricity demand while being woefully underexposed to the sectors and industries that stand to benefit from it.

The conflict in Iran is likely to have a lasting impact on global commodities markets, regardless of the timeline for a return to “normal” for the Strait of Hormuz. Countries around the world have realized the risk of being dependent on geopolitically unstable geographies, and countries beholden to the straight are likely to invest heavily in the infrastructure needed to circumvent the strait. This will put additional pressure on already fraught global commodities markets as stockpiling and global supply chain reorientation efforts are undertaken.

While the comparisons with the Dot Com bubble are often misguided, the truth is extended market valuations, high equity market concentration, and the expected Initial Public Offerings (IPOs) of private companies at trillion-dollar valuations does bear watching. These dynamics create a potential risk-reward dynamic that cap-weighted US large-cap portfolios are ill-equipped to remediate.