Equinor’s recent billion-kroner gains from artificial intelligence do not come from advanced language models. Instead the company relies on classic machine learning applied to its own data, according to a report from Digi.no.
The Norwegian energy giant has not disclosed exact figures for the AI-driven profits. But its annual reports show income from digital solutions rose sharply in the past two years. Executives say the gains come from optimizing production, reducing downtime and cutting maintenance costs across oil and gas facilities.
Equinor’s chief digital officer told Digi.no the company treats AI as a tool, not a standalone business. We build models on our own data, he said. The approach avoids the need for full data release and keeps competitive advantages inside the company.
Industry watchers note that Equinor has spent years collecting sensor data from platforms and pipelines. Those datasets feed the machine-learning systems that predict equipment failure and schedule repairs before breakdowns occur.
The strategy contrasts with public language-model projects that often share code or datasets. Equinor keeps its models and data proprietary, focusing on internal efficiency rather than open-source contributions.
Source: digi.no