Salt Is the New Oil: Critical Minerals, the 2026 Investment Playbook

Morgan Stanley says salt is the new oil. GE Vernova builds gas turbines for AI data centres. Copper, lithium, and sodium are the new strategic reserves. Inside the commodity race defining the next decade.

When Morgan Stanley’s commodities research team declared in June 2026 that salt is “the new oil,” it was not making a literal claim about sodium chloride replacing petroleum. It was making a structural argument about how the economics of energy transition — combined with the extraordinary power demands of AI infrastructure — are elevating previously overlooked materials into strategic commodities with genuine supply constraint characteristics.

The argument has three layers. Sodium-ion batteries, which use sodium rather than lithium as their charge carrier, are emerging as the cost-competitive alternative for grid-scale energy storage applications that do not require the energy density of premium lithium chemistries. Sodium is abundant — effectively unlimited in supply relative to lithium — but the manufacturing infrastructure to process it into battery-grade inputs remains concentrated and immature. As AI data centres create an insatiable demand for power storage to manage grid intermittency from renewable sources, the industrialisation of sodium-ion chemistry is becoming an investment thesis with multi-decade horizon.

GE Vernova and the AI Energy Infrastructure Buildout

The more immediate energy story is less speculative. GE Vernova is building massive gas turbines specifically to power the AI data centre boom — a business that positions the industrial conglomerate at the intersection of AI infrastructure and energy security in a way that few traditional manufacturers had anticipated when the generative AI wave began in 2022.

The numbers behind data centre energy demand are staggering. A single large-scale AI training cluster requires as much electricity as a small city. The International Energy Agency projected that data centres would double their electricity consumption between 2022 and 2026, with AI workloads driving the acceleration. Utilities, grid operators, and industrial equipment manufacturers have found themselves with unexpectedly urgent, unusually large orders — and constrained capacity to fulfil them.

GE Vernova’s gas turbine order book reflects this structural demand. Grid-scale gas generation is being positioned not as a fossil fuel legacy but as the dispatchable backup to renewable intermittency that allows AI infrastructure to operate at the reliability levels hyperscale data centres require. The climate policy tension is real but commercially secondary to the infrastructure imperative.

Memory Shortages: The Apple-Microsoft Crisis and the Smaller Players at Existential Risk

The AI infrastructure boom has created a secondary supply crisis that has reached into consumer electronics: a global shortage of high-bandwidth memory chips that is described as an “existential crisis” for smaller players while causing significant disruption for Apple and Microsoft. The shortage stems from the same AI training cluster buildout: Nvidia’s H100 and H200 GPUs require HBM3 chips at volumes that have outpaced Samsung, SK Hynix, and Micron’s ability to expand production capacity.

South Korean chipmaker SK Hynix announced plans to raise $29 billion via a Nasdaq listing as soon as July 2026, in part to fund capacity expansion. The capital raise would represent a substantial bet on sustained AI chip demand — and a significant addition to a global technology equity market that has already absorbed SpaceX’s $86 billion fundraise.

The Buy Now, Pay Later Credit Cycle: Blue Owl, KKR, and the Untested Model

One of the more pointed warnings from Bloomberg’s coverage in June 2026: billions are flowing from firms like Blue Owl and KKR into Buy Now, Pay Later companies — an untested credit model in a higher-rate environment where consumer balance sheet stress is rising. Sceptics worry about what happens in a downturn to BNPL portfolios funded by private credit. The companies themselves argue that their underwriting technology is superior to legacy consumer credit models. Both arguments rest on assumptions that have not yet been stress-tested by a genuine recessionary cycle.

The convergence of alternative credit providers, private capital, and consumer fintech is one of 2026’s most significant financial structural experiments. If the Iran war recession scenario materialises, it will provide the empirical data that models have been unable to generate.