Goldman Sachs issued a research report this week stating that the current discussion around the impact of AI on the energy industry is mainly focused on the demand side. For example, the increase in AI computing power demand will lead to an increase in power demand, which may push up energy demand and prices. However, the impact on the supply side is more complex. AI can greatly improve the efficiency of oil exploration and production, reduce oil production costs, and increase oil supply. Although AI may promote oil consumption on the demand side, efficiency improvements on the supply side may restrain oil price increases in the long term.
First, AI has the potential to reduce costs by optimizing the entire supply chain, such as logistics and resource allocation. This is particularly important in shale oil development, for example, because wells are often located in remote areas and transporting supplies is costly. AI can optimize the transportation paths of materials and equipment through large-scale data analysis, reducing transportation costs and time. In addition, AI can analyze oil well production data in real time and rationally allocate mining equipment and human resources.
According to Goldman Sachs estimates, AI has the potential to reduce the construction cost of new shale oil wells by about 30%, resulting in a reduction in marginal incentive prices of about US$5/barrel, meaning that oil companies can maintain production at lower costs, thereby boosting global oil supply. increase in quantity.
Secondly, AI is expected to significantly increase the recovery rate of US shale oil and expand the ultimate recoverable oil reserves. Shale oil reservoirs are often located in tiny fractures or pores, making it difficult for traditional mining techniques to effectively extract oil from these fine structures. Therefore, despite the vast amounts of oil stored underground, actual shale oil well recovery rates are low.
Goldman Sachs estimates that if AI technology can increase U.S. shale oil recovery by 10-20%, oil reserves may increase by 8-20%, equivalent to an increase of 10 billion to 30 billion barrels.
The positive impact of AI on the economy is mainly reflected in the improvement of production efficiency and the increase in income driven by innovation. This income growth may increase consumption levels, and people may increase consumption of services that rely on petroleum products, such as transportation and tourism, thus driving demand for petroleum products.
Overall, Goldman Sachs predicts that while AI is likely to drive oil demand through revenue growth over the next decade, the amount will be relatively small at about 700,000 barrels per day, which could boost long-term oil prices by about $2 per barrel. .
However, the effect of AI in boosting oil demand is relatively limited. AI boosts the growth of electricity and natural gas demand more significantly. This is because AI technology relies on a large amount of computing power, which directly promotes the demand for electricity, especially the rapid popularity of electric vehicles. Will significantly reduce oil demand. At the same time, natural gas, as a relatively clean energy source, is often used for power generation. Therefore, the popularity of AI will also indirectly increase the demand for natural gas.
Goldman Sachs predicts that the positive effect of AI on oil demand is not enough to offset the negative impact of electric vehicles and natural gas substitution on oil demand. As the world gradually shifts to electric vehicles, oil demand is expected to decrease by about 8 million barrels per day in the next 10 years. Falling natural gas prices are expected to reduce oil demand by about 2 million barrels per day.
Taken together, Goldman Sachs believes that AI may have a moderate net negative impact on oil prices in the medium to long term. Since the downward pressure on prices brought by increased supply (a decrease of US$5/barrel) significantly exceeds the price increase caused by increased demand (an increase of US$2/barrel), the net impact of AI on oil prices tends to be negative. Therefore, with the widespread application of AI technology, the global oil market may enter a long-term price downward cycle.