Why AI could weigh down oil prices in the next decade, according to Goldman Sachs

Three oil and gas rigs owned by Transocean Ltd sit idle in the Grand Harbor in Valletta, Malta October 22, 2015.

The artificial intelligence boom could make oil prices fall even as the technology boosts power demand, Goldman Sachs analysts said.

The analysts point to AI’s improved logistics capabilities and boost to oil recovery, which could increase oil supply and drive down prices in the next decade or so.

That forecast stands in contrast to predictions that energy prices will rise as demand for AI requires more power for data centers.

“The debate on the impact of AI on energy and metals has mostly focused on the demand side given the expected boost to power demand. However, we highlight that AI may eventually weigh on oil prices over the next decade by boosting oil supply through two main channels,” the analysts said in a Tuesday research note.

The analysts point first to AI’s potential to reduce oil production costs by improving logistics and resource allocation.

They point to an example of oil rigs in North Dakota, where an oil company’s use of AI-driven predictive and automated drilling reduced drilling times by 30%.

Assuming a 25% productivity gain, other early AI adopters could save $5 per barrel in costs, the analysts estimate.

The analysts say AI could also boost recoverable resources, with a hypothetical 10%-20% jump in recovering US shale, which could increase oil reserves by 8%-20%, or up to 30 billion barrels, they estimate.

Those supply gains greatly outweigh any potential boost in demand, the analysts wrote. Even as AI requires more power, that demand likely won’t effect oil demand since oil is “scarcely used as a source of energy for power generation today.”

Instead, power generation demand will likely boost power and natural gas markets, they said.

Any AI-related increase in oil demand will likely come from higher incomes, which could raise prices by $2 per barrel, the analysts estimate.

But that impact is minimal compared to how much electric vehicle adoption and lower natural gas prices will weigh on oil demand, while oil supply increases via the use of AI.

“Taken together, we believe that AI would be a modest net negative to oil prices in the medium-to-long term as the negative impact from the cost curve (c.-$5/bbl) – oil’s long-term anchor – would outweigh the positive impacts from the demand side (c.+$1-2/bbl),” the analysts wrote.