Dangote Refinery Is First to Deploy Honeywell Industrial AI

Dangote Petroleum Refinery in Lekki will be the first site to deploy Honeywell Technologies’ Forge Plant Cognition, an industrial AI system built with Nvidia and Dell Technologies. The project puts artificial intelligence inside the operational decisions of one of the world’s largest single-train refineries, rather than treating AI as a customer-facing add-on.

Honeywell announced the launch on 26 October and estimated that the technology could generate more than $60 million in annual gross operational benefits at full capacity. That figure is a projection, not a saving Dangote has already recorded. Honeywell says it excludes implementation costs and depends on operating conditions, crude mix, utilisation, margins and the quality of the refinery’s data.

Forge Plant Cognition is designed to combine live plant data with engineering knowledge, automation and physics-based models. Honeywell says it can identify bottlenecks that limit throughput or product yield, flag emerging equipment problems and help operators decide how to respond. The aim is not simply to produce a chatbot’s answer. It is to make complex process information useful to engineers while the plant is running.

The system uses Dell AI Factory infrastructure with Nvidia technology. Honeywell says Nvidia Nemotron models help agents interpret operating data, while Omniverse libraries support a digital twin for simulation and planning. Vision AI tools are intended to detect anomalies and issue alerts. None of that establishes how much value the refinery will actually gain; the test will be how reliably the tools perform under real operating conditions.

Honeywell put its preliminary expected benefit at more than 29 cents per barrel, which it translated into the greater-than-$60 million annual figure at full-capacity throughput. Readers should not treat the estimate as net profit. Costs of deployment, integration, maintenance and training will matter, as will whether improved decisions translate into measurable increases in output or fewer interruptions.

Dangote says the tools should also make specialist process knowledge more accessible to its local engineering and operations teams. That may prove as consequential as any immediate efficiency gain. A refinery can buy advanced software, but dependable results require people who understand the plant, challenge recommendations and take responsibility for decisions. Honeywell says the project supports Dangote’s goal of building a workforce that is 70% Nigerian.

The deployment comes alongside a broader expansion plan. Honeywell and Dangote have said they are working to raise the refinery’s capacity from 650,000 to 1.4 million barrels a day over three years. That is a separate target and should not be attributed to the AI platform alone. TechBooky has previously explained how Nigerians can access the refinery’s share offer, another sign of the scale of public interest in the project.

The useful measure now is evidence from the site after deployment. Honeywell and Dangote can report whether downtime falls, yield improves and local teams gain practical capability, while separating gross operational gains from the cost of achieving them. Until then, the launch is an ambitious industrial AI trial with a compelling forecast, not a proven $60 million windfall.