One hundred percent of critical operational decisions now supported by specialist systems
In Itabira, the Brazilian city where Vale was founded more than eight decades ago, the mining giant has reopened a transformed processing plant — one where artificial intelligence now governs the rhythms of ore and output that once depended entirely on human hands. The Conceição2 facility, reborn as a Model Usina, marks a deliberate reckoning with the limits of industrial tradition: a wager that machines watching four hundred variables in real time can do what no single workforce could sustain alone. The move speaks to a broader tension of our age — the pursuit of efficiency, safety, and sustainability through the quiet displacement of human judgment by algorithmic oversight.
- A plant that once relied on manual oversight and human exposure to hazardous zones has been fundamentally reimagined, with over 7,300 automated instruments and 100+ cameras now running what workers once had to enter dangerous spaces to manage.
- The urgency was industrial and existential: unplanned shutdowns, inefficient mineral recovery, and worker risk were bleeding productivity and threatening Vale's competitive position in a decarbonizing global steel market.
- In under two years, the AI-driven system delivered a 25% productivity surge, a 40% jump in premium pellet output, and a 26% drop in iron content lost to waste — results that validated the eighteen-month, fifty-one-solution overhaul.
- Water recirculation now reaches 92%, and critical operational decisions are fully supported by specialist systems, signaling that the facility has crossed from pilot experiment to replicable model.
- Vale and partner ABB are now positioning this Itabira blueprint as a template to be scaled across the company's broader mining network, turning one plant's transformation into an industry-wide ambition.
Vale inaugurated its first high-tech processing facility in Itabira, Minas Gerais — the city where the mining company was founded 84 years ago — marking a deliberate turn toward operations governed by artificial intelligence rather than human judgment alone. The Conceição2 plant, now called the Model Usina, was rebuilt over eighteen months through fifty-one integrated solutions designed to eliminate bottlenecks and modernize every stage of ore processing.
The scale of the transformation is considerable. More than 100 surveillance cameras monitor the complex, nearly 7,300 instruments have been automated, and an AI layer now supervises over 400 production variables in real time — adjusting processes based on incoming mineral characteristics and anticipating equipment failures before they cause shutdowns.
The results arrived quickly. After less than two years in pilot operation, productivity rose 25% and annual output capacity climbed from nine million tons to 11.2 million. The share of premium direct reduction pellets — critical for greener steel production — jumped 40%. New real-time ore analysis technology reduced iron lost in waste material by 26%, and the plant now recirculates 92% of its water.
Perhaps the most human dimension of the change is what workers no longer have to do. Remote operation systems and mechanical arms now handle tasks that once required physical presence in hazardous zones, allowing operators to manage the facility from control rooms at a safe distance. When on-site access is necessary, real-time data helps workers plan with precision, minimizing time in risky conditions.
Vale built this transformation alongside Swiss automation specialist ABB, which integrated new technologies with existing systems rather than replacing them wholesale. The partnership proved efficient enough that Vale now plans to replicate the model across its other mining operations — carrying what it learned in Itabira into the broader future of the industry.
Vale opened the doors of its first high-tech processing facility on Wednesday in Itabira, a city in Minas Gerais where the mining company was founded 84 years ago. The Conceição2 plant, now reborn as what the company calls its Model Usina, represents a deliberate shift toward operations run by artificial intelligence and data systems rather than human judgment alone. The transformation took eighteen months and involved fifty-one separate solutions designed to eliminate bottlenecks and improve how the facility works.
The modernization is substantial in its scope. More than one hundred surveillance cameras now watch the complex. Nearly seventy-three hundred instruments have been automated, including new measurement devices and sensors throughout the plant. An artificial intelligence layer now supervises more than four hundred variables across every stage of ore processing, making real-time adjustments to the production process based on the characteristics of the incoming mineral and the desired output. The system can anticipate failures before they happen, preventing unplanned shutdowns that would otherwise halt operations.
The results have been striking. In less than two years of pilot operation, the facility increased its productivity by twenty-five percent and reached its designed capacity. In 2024, the plant produced nine million tons of ore. This year, it is already capable of producing 11.2 million tons annually. The composition of what the plant produces has also shifted. The share of pellet feed for direct reduction—a premium product essential for decarbonizing steel production—jumped by forty percent. The plant has also optimized its mineral recovery, squeezing more value from each ton of material processed.
Quality improvements came through new technology that analyzes ore content in real time as it moves through the beneficiation process. When the system detects variations, it immediately adjusts the treatment route, ensuring the iron in the material is captured as efficiently as possible and waste is minimized. By this year, the average iron content in the plant's reject material dropped by twenty-six percent. The facility now recirculates ninety-two percent of the water it uses, filtering and returning it to the operation after the final processing stages.
Perhaps the most consequential change is invisible to production numbers. Workers no longer need to enter many of the hazardous zones within the industrial plant. Remote operation systems, including mechanical arms, now handle tasks that once required human presence in dangerous environments. Electrical and mechanical equipment—motors, valves, and other components—have been automated, allowing operators to run the facility from control rooms at a distance. When workers must access the installations, real-time images and data help them plan their activities with precision, reducing time spent in risky conditions.
Carlos Medeiros, Vale's vice-president of Operations, described the modernized plant as enabling "a more integrated, highly monitored, and safer operation." Rafael Bittar, the company's technical vice-president, noted that one hundred percent of critical operational decisions are now supported by specialist systems. He framed the facility as a step toward "a safer, more sustainable, and more competitive mining operation prepared for the challenges of the coming decades."
Vale did not build this transformation alone. The Swiss multinational ABB, which specializes in automation and electrification, served as the strategic partner. Rather than ripping out existing systems and starting fresh, ABB worked to integrate technologies Vale already had in place, optimizing existing investments and accelerating results. The company acts as a technology integrator, ensuring different systems and suppliers can work together seamlessly. Vale plans to replicate this model across its other mining operations, using the same partnership approach to scale what it learned in Itabira.
Citações Notáveis
The modernization allows a more integrated, highly monitored, and safer operation, with artificial intelligence supervising hundreds of variables and making real-time adjustments to production based on ore characteristics.— Carlos Medeiros, Vale vice-president of Operations
The facility has reached a new level of digital maturity, with all critical operational decisions supported by specialist systems, combining innovation, technology, and technical knowledge to increase efficiency and reduce impacts.— Rafael Bittar, Vale technical vice-president
A Conversa do Hearth Outra perspectiva sobre a história
Why Itabira specifically? Was it just symbolic, or does the location matter operationally?
It's both. Vale was founded there eighty-four years ago, so there's a historical weight to it. But practically, Itabira is where they could test this model at scale without disrupting their entire operation. It's a proving ground.
The 25% productivity gain in less than two years—is that unusual for a mining operation, or is that the kind of improvement AI typically delivers?
It's significant. You don't usually see that kind of jump that fast. But this wasn't just AI bolted onto an old system. They rebuilt the entire operation around data—new cameras, new sensors, new instruments. The AI had clean information to work with from day one.
What about the workers? You mention remote operation and fewer people in dangerous zones. Does that mean fewer jobs?
The source doesn't say jobs were cut. What it emphasizes is that workers are no longer doing the most hazardous tasks. Whether that translates to layoffs or redeployment—that's not addressed. But the safety improvement is real and measurable.
The water recirculation at 92%—how does that compare to what other mines do?
The source doesn't provide a comparison. But ninety-two percent is high enough that Vale is highlighting it as a sustainability achievement. It suggests most of what comes through the system gets used again rather than being discharged.
Why is ABB the partner here, and not a software company?
ABB brings automation and electrification expertise, which is what a mining operation actually needs. They're not writing code in isolation—they're integrating systems that control physical equipment. That's different from a pure software play.
What happens next? Is this model ready to roll out everywhere?
Vale says it's a reference point for expansion to other operations. But they're being careful—they built an ecosystem of partners to handle integration, which suggests they know scaling this won't be simple. Itabira is the proof. The real test is what happens when they try to replicate it elsewhere.