Cleveland-Cliffs launches three-year AI rollout with Palantir

Cleveland-Cliffs has signed a three-year agreement with Palantir to deploy artificial intelligence across its entire operations chain. For the American steelmaker, this is not a one-off pilot but an attempt to make plant modernization systematic.

Why the contract matters

The deal runs for three years, and the term alone shows the scale of the task. This is not a one-time test at a single site but a gradual rollout of AI tools across production, planning and day-to-day operations. For Cleveland-Cliffs, it is a way to speed up the renewal of its industrial base without building everything from scratch: companies typically get part of the benefit from better use of data, optimized equipment loading and faster decision-making.

For heavy industry, this matters most when margins depend not only on metal prices but on how precisely plants keep to schedule, how quickly disruptions are fixed, and how much raw material, energy and time goes into producing each ton. In this context, AI is not a showcase for investors but a tool that has to deliver an economic result in real operations.

Where the impact is expected

The companies have not disclosed which use cases will be part of the first phase. But in comparable industrial projects, AI is usually introduced first where there is plenty of historical data and a clear KPI that can be improved within the first months. This lowers project risk: teams first pick areas where the effect is easier to measure in money, time and output stability. Such cases are also easier to scale across several sites later.

• Scheduling of equipment and shift loading to reduce downtime and bottlenecks
• Equipment monitoring to detect failures earlier and prevent breakdowns
• Quality control and detection of deviations in process parameters
• Optimization of logistics for raw materials, semi-finished and finished products
• Analysis of production losses, energy consumption and other costly resources

If Cleveland-Cliffs and Palantir follow this route, the main task will be not simply to connect models but to build them into the decisions of foremen, dispatchers and site managers. In industry, AI delivers results only when recommendations reach the shop floor and change what people do on shift, rather than remaining polished analytics on a separate dashboard. Otherwise the system stays an adviser with no noticeable influence on output.

Why the bet now

The partnership matters not only as a fact but for its timing. American industrial companies are increasingly looking for ways to modernize existing capacity faster and more cheaply than long capital projects allow. Against this backdrop, AI becomes a way to squeeze more out of assets already running: improve predictability, cut losses and shorten the decision cycle. After the wave of AI enthusiasm, businesses increasingly demand not presentations but concrete returns on live production lines.

Such agreements are also telling for Palantir. The market is less interested in abstract promises around generative AI and more in projects that tie the technology to operational efficiency, output and financial impact. Deals with industrial groups therefore work today as a maturity test for AI platforms: if the result can be measured at the plant, it is much easier to defend before a board of directors and replicate at other sites. This is where a platform's viability is tested.

The companies have not yet disclosed the specific toolset, the geography of the rollout or the target metrics that investors and management will track. But it is already clear that Cleveland-Cliffs sees AI as part of a broader production modernization, not as a separate experiment inside the IT function. That raises the stakes: the project will be judged on whether it improves output, reliability and the economics of the business. These metrics will show whether AI has made the move from presentations to production practice.

What it means

For the market, this is one more signal that AI is leaving its status as an office assistant for good and taking hold in heavy industry. If Cleveland-Cliffs can show a measurable effect over the three-year cycle, similar programs at other manufacturers will accelerate, because the question is no longer AI fashion but plant competitiveness. For technology vendors, it means growing demand for solutions that can work with shop-floor data, not just office tasks.