The world's physical resources are coming under pressure from three forces: conflicts in the Middle East and Ukraine are draining energy supplies, artificial intelligence is driving unprecedented demand for electricity and metals, and disruption to fertiliser and other inputs is threatening the next agricultural cycle.
Since the Middle Eastern conflict began, global oil inventories have been depleted by approximately one billion barrels. Proposed US data centres would require 200GW of power1, and farmers face renewed fertiliser supply pressure ahead of 2027. The strain is spreading across energy, industrial metals and food production, and years of constrained investment and increasingly fragmented supply chains could make it difficult to reverse.
Today's market is not a repeat of the China-led commodity supercycle of the 2000s. That period was defined by extraordinary demand growth. This one is being driven less by demand and more by constrained supply, geopolitical fragmentation and the growing importance of regional resource security.
Call it the revenge of the old economy. For years, investors have been captivated by the digital economy, yet AI, data centres and electrification all depend on the physical world. A data centre cannot be built without copper and aluminium, and it cannot be run without power.
Demand is solid, but the real story is supply. Producers have spent the past decade prioritising cash flow over growth, and that discipline is now colliding with geopolitical fragmentation and new sources of demand. We believe this cycle is still in its early stages.
The global oil market is facing three simultaneous crises: upstream production, refining capacity and tanker availability. An estimated 7 to 7.5 million barrels a day of crude and refined product supply is missing from the Middle East and Russia combined2, while some 8 million barrels a day of refining capacity is offline globally3.
The costs are showing up along the chain. The US Gulf Coast three-to-one refining crack has risen to approximately US$45 a barrel, against a more typical level of US$20. Tanker costs between the Middle East and Asia have climbed from between US$2 and US$3 to around US$30 a barrel, a more than tenfold increase4. Global oil inventories have fallen by around one billion barrels since the start of the Middle Eastern conflict, leaving observed stocks at decade lows5.
The three crises are compounding one another, so Brent at US$110 does not tell the whole story. The pressure is also visible in refining margins, freight costs and the price consumers are paying for diesel. Even if the conflict ended tomorrow, rebuilding a billion barrels of inventories would create additional demand for years. We think the market is underestimating how long a return to normal could take, and supply has not normalised yet.
After two decades of largely stagnant US electricity consumption, artificial intelligence and manufacturing reshoring are driving a new wave of power demand. Proposed US data-centre projects represent approximately 200GW of electricity demand, equivalent to roughly 200 times the power consumption of San Francisco.
The AI revolution is running up against a very physical constraint. Hyperscalers are planning data centres on an unprecedented scale, but grid capacity is constrained and cannot keep pace. These facilities need enormous amounts of reliable power, and they need it quickly, so hyperscalers are increasingly exploring direct power generation. That creates new opportunities for natural gas. The race to build AI infrastructure is also becoming a race to secure energy, because without the power to run them, even the most advanced data centres are just expensive buildings.
The same race is redrawing metals markets, although record prices are masking increasingly divergent conditions. Copper inventories have risen sharply as material is redirected into the US amid tariff uncertainty, while Chilean production has fallen to its lowest level in six years6. Meanwhile, approximately four million tonnes of new Indonesian aluminium capacity is expected over the coming year7, even as European inventories remain historically low.
Copper is sending particularly mixed signals. Some investors see a market heading for a significant deficit, while others see oversupply, and both can point to evidence supporting their view. The long-term fundamentals remain compelling, driven by electrification and growing demand for infrastructure. But a strong long-term investment case should not be confused with a straightforward market today. In the short term, at these prices, copper looks increasingly risky.
Central banks purchased a net 289 tonnes of gold in the second quarter of 2026, 62 per cent more than a year earlier, according to the World Gold Council8. Sustained official-sector demand reflects a broader shift towards reserve diversification.
Central banks are continuing to buy even at historically elevated prices, which suggests a structural shift rather than a short-term reaction to geopolitical uncertainty. Meanwhile, gold miners are generating extraordinary cash flows, but equity valuations have yet to fully reflect the improvement in profitability.
Disruption to Middle Eastern fertiliser supplies, higher energy costs and tightening agricultural inventories are raising concerns about global food production heading into 2027.
The Middle East is a major supplier of nitrogen fertiliser, while global potash production is concentrated in a handful of countries. In 2025, Canada produced around 15 million tonnes of potash, compared with 10 million tonnes from Russia and 6 million tonnes from Belarus, on a potassium oxide equivalent basis. Transport bottlenecks continue to complicate exports from Russia and Belarus.
Northern Hemisphere farmers largely secured fertiliser supplies for the 2026 growing season, but renewed purchasing ahead of 2027 plantings could place significant pressure on availability and prices. For nitrogen, the full impact of the supply shock may not emerge until farmers are forced to return to the market. Potash faces a different challenge, as concentrated production and logistical bottlenecks make it increasingly difficult to get material to customers.
Global grain inventories also appear to have peaked9, while a prolonged El Niño could disrupt harvests in 2027. A potential increase in US ethanol blending to 15 per cent could create a significant new source of demand for corn10. After years of plentiful grain supplies, the agricultural cycle may be beginning to turn.
The risk is that these pressures arrive together. Higher diesel and fertiliser costs are already squeezing farmers, and a prolonged El Niño could put further pressure on yields. We believe the conditions are emerging for a potential multi-year bull market, particularly across nitrogen fertiliser, biofuels and seeds.
Across energy, metals and agriculture, the same underlying pressures are emerging: demand for physical resources is growing, supply chains are becoming more fragmented and producers remain disciplined about bringing new capacity online. For investors, the opportunity extends beyond higher commodity prices to the natural-resource companies positioned to benefit from tighter markets and stronger cash generation.
The world is entering a period in which access to physical resources could matter as much as access to capital. Conflict is exposing the fragility of global supply chains, AI is accelerating demand for power and metals, and pressure on agriculture is raising the stakes for food security. Years of underinvestment mean supply cannot simply respond overnight, and the ability to pay for resources will become increasingly important.
This is about more than rising commodity prices. It is about a fundamental shift in the value of the companies that produce the energy, metals and agricultural resources the global economy depends on. The next investment cycle may be defined as much by what the world struggles to produce as by what it wants to consume.
1 World Bank, UN, BHP analysis, Ninety One, September 2025 and NextEra Energy, Ninety One. ISO/RTO forecasts, NERC ES&D, utility IRPs, ICF. Historical demand represents data from NERC ES&D from 2000 to 2023; 2024 represents forecast from NERC ES&D; Q1 2025 represents ICF's demand projections from 2025 to 2050. August 2026.
2 Morgan Stanley Research, EIA, BNP. September 2026.
3 Bloomberg, Ninety One. September 2026.
4 Bloomberg, Poten, Ninety One. September 2026.
5 JP Morgan, Bloomberg, Ninety One. September 2026.
6 Bloomberg, Ninety One. September 2026.
7 Bloomberg, Ninety One. September 2026.
8 World Gold Council, Ninety One. 30 June 2026.
9 Ninety One, Bloomberg. September 2026.
10 EIA, Ninety One, oxygenate plant production of fuel ethanol (annual sums of monthly data). 2025 supply/disposition: gasoline supplied 3,261mb; ethanol blended 326.8mb (10.0 per cent blend); exports 52.7mb. Scenario: production = gasoline pool × blend rate + exports (held at 2025 level) + 2025 residual. Gasoline pool grows 3 per cent a year from 2025; blend rate ramps linearly from 10 per cent to 15 per cent by 2031. 2026 shown as scenario year.
