Software, Gas, and the New Muscle of Industrial AI

Big software deals and new gas power projects show how refiners are gearing up for AI by rewiring control systems, data flows, and energy supply

8 Apr 2025

Three deals announced in early 2025 point to a clear shift in how refiners and heavy industry are preparing for the age of artificial intelligence. Together, they link digital control, cloud analytics, and raw power supply into a single, fast-moving strategy.

First came Emerson’s move to fully absorb Aspen Technology. In March, the automation giant paid about US $7.2 billion to buy the shares it did not already own. The goal is less about scale and more about control. By fusing AspenTech’s optimization software with Emerson’s sensors and valves, the company wants refineries to act and adjust in real time, not hours later. Closed-loop control, long promised, is now closer to routine practice.

A few weeks later, AVEVA struck a very different kind of partnership. Its deal with Databricks connects factory-floor data to cloud-based AI tools. Announced in San Francisco, the tie-up aims to break down the old wall between operational technology and IT. Early tests suggest that when process historians feed directly into Databricks, predictive maintenance models train far faster, cutting time by more than 50 percent. For operators, that means quicker insights and fewer blind spots.

The third move tackles a problem that sits outside the refinery fence but increasingly shapes it. AI needs power, lots of it. In January, Chevron joined Engine No. 1 and GE Vernova to plan up to 4 gigawatts of gas-fired generation built next to data centres. These so-called power foundries are designed to bypass grid bottlenecks and keep servers running around the clock, with first power expected in 2027.

Seen together, the message is blunt. Control systems are going vertical, data platforms are going open, and energy supply is being locked in. For downstream operators, the payoff starts with better analytics and higher uptime. Over time, this blend of on-site control, cloud AI, and dedicated power could reset what it means to run an efficient, AI-ready refinery.