Equinor
AI in industrial operations
Equinor said AI produced USD 130 million in value and savings in 2025
Problem
Unplanned failures of rotating equipment such as pumps and compressors on oil and gas installations are costly, and interpreting seismic data for exploration takes geoscientists a long time.
Approach
Since 2020 the Norwegian energy company has monitored more than 700 rotating machines through about 24,000 sensors across its facilities, using models that flag deviations before breakdowns. It also uses AI to interpret seismic data, covering 2 million square kilometres in 2025, and used it to find a design option for phase 3 of the Johan Sverdrup field. The company says it has identified more than 100 further use cases.
Results
Takeaway
Predictive maintenance accounted for USD 120 million of more than USD 330 million in AI value reported since 2020. The 2025 figure combined value and savings for Equinor and its licence partners, with no calculation method disclosed.
Categories
- Size
- Large enterprise
- Industry
- Energy and utilities
- Business function
- Production and qualityIT and R&D
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
Equinor reports USD 130 million in value and savings from AI in 2025 for itself and its partners, and over USD 330 million since 2020. It credits predictive maintenance on more than 700 machines with 24,000 sensors with USD 120 million since 2020, and cites a tenfold gain in seismic interpretation capacity and a USD 12 million saving on Johan Sverdrup phase 3. It says more than 100 new use cases have been found.
The Society of Petroleum Engineers' magazine reports Equinor's figures. It quotes the executive vice president for technology saying the company mainly uses traditional machine learning on operational data, with copilots and chatbots used by staff. The 2025 amount includes both value creation and savings.
Full details are in the original sources.