AI becomes a boost for global oil demand, not a bridge to net zero

Artificial intelligence could help the energy industry tap an additional one trillion barrels of oil by making more of these barrels economical to extract.

Wood Mackenzie analysts made the prediction recently, adding fuel to an already heated debate: is AI going to accelerate the transition or, as the Wood Mac prediction suggests, hinder it by extending the world’s reliance on oil and gas?

For the energy industry, the question is rather irrelevant. From the perspective of companies in the business of extracting energy commodities, AI is a big positive that will help them continue extracting energy commodities for a longer period, for less money than would have been needed otherwise.

This is an important consideration in the context of energy security, and it became even more important in the past few months after forecasters began to change their predictions in the face of physical facts. Even the International Energy Agency has dropped its peak oil prediction and now expects oil to remain a dominant element in the global energy mix until 2050.

For climate activists, however, the role that AI could play in bringing hundreds of millions of new barrels to the oil market is that of a villain. While it is often touted as an important tool in the pursuit of a net-zero energy system, artificial intelligence now emerges as a tool for stalling the transition by opening up new oil deposits to producers.

Axios, in a report on the Wood Mackenzie prediction, cited a climate activist as lamenting the findings in the consultancy’s forecast and suggesting that the transition could have had a better chance if oil and gas had become harder to extract and more expensive, but alas, AI came and put an end to that.

“[F]ossil fuels were becoming too difficult and expensive to produce — until AI came along to make them profitable again,” the report cited activist Holly Alpine as saying. It is, in fact, true that in many places the production costs of oil extraction are going up, driven by factors such as general inflation and natural reservoir depletion. However, with oil, it has always been about demand driving supply, so as demand remains resilient, so will supply—with AI potentially helping producers optimize extraction.

Interestingly, the findings of the Wood Mackenzie report are based on artificial intelligence as well. The company developed its own AI-powered tool called Analogues and then used it to identify the reservoirs where the biggest gains could be made in already producing areas. The tool revealed that using already existing technologies in production, the energy industry could tap an additional 1 trillion barrels of oil from existing fields alone.

This should be good news as demand for oil is set to continue growing for longer than previously expected, again per Wood Mackenzie. Global demand for crude oil is going to continue on an upward trajectory until at least 2032, the consultancy said in a report last month, adding the world was way off track in meeting its Paris Agreement goals. In an ironic twist, AI may well contribute to a further veering off the Paris Agreement course because of its massive need for energy.

Energy consumption by data centers has become notorious in a matter of months. The facilities hosting AI infrastructure require so much electricity to run reliably that traders are rushing to bet on a natural gas rally as utilities plan new power plants to ensure there is enough supply. Forecasts point towards a sharp rise in electricity demand everywhere there are data centers, and in some cases even suggest a shortage of electricity may emerge, for instance, in Norway.

The energy consumption aspect of artificial intelligence has prompted serious concerns from climate change advocacy circles, as it has spurred a race to secure baseload power generation capacity, meaning gas, coal, and nuclear. Some from the industry, however, believe that AI itself is instrumental for the energy transition by helping transform the grid and improve power generation and distribution, while keeping energy affordable. The technology has yet to live up to this particular promise, but it is already helping oil and gas producers optimize their production processes.

Meanwhile, demand for oil continues to climb higher, despite all the efforts to replace it with demand for electricity generated by low-carbon technologies such as wind and solar. Two months ago, demand for crude oil booked an especially sharp increase, both on a monthly and an annual basis, suggesting those new forecasts for “stronger for longer” may well be right. Lower prices will help that, too, even though forecasters argue that their price predictions are bearish because of weak demand. The thing with oil is, when the price is low enough, demand springs higher—and what the Wood Mackenzie modeling suggests is that AI can put a ceiling on prices by optimizing drilling, ensuring supply remains affordable for longer.