Karthikeyan Natarajan, CEO of Infinite Uptime, thinks that cement plants could operate autonomously by 2030.
Global Cement (GC): Please can you introduce Infinite Uptime to our readers?
Karthikeyan Natarajan (KN): Infinite Uptime was founded in 2018 to provide digital solutions that help users to increase uptime in heavy industries. It began with non-AI systems, but the technology has since advanced greatly into AI-based solutions.
GC: How key is the cement sector for the business?
KN: Infinite Uptime is involved in steel, cement, paper and pulp and heavy processing engineering industries, including chemical, energy, oil and gas and many other sectors. Cement is in the top three, with our systems active in 165 cement plants operated by around 45 cement producers in India, the Middle East, Southeast Asia, the US, Western Europe and Africa.
GC: What areas of the plant do your systems cover?
KN: The dominant areas are kilns, mills, roller presses, preheater cyclones and separators, the pieces of equipment that cause the most havoc when they fail. Data accuracy is the foundation of any AI-based control system. As cement processing equipment can be quite slow-moving, you need high-fidelity sensors that can take data accurately.
This is why Infinite Uptime developed its own line of piezoelectric sensors to work alongside conventional ones. These sensors work by analysing the very small electrical charges generated by small crystals when they are compressed. They are very useful for slowly-rotating machinery like rotary kilns.
GC: How does Infinite Uptime use this data?
KN: The value is not in the hardware, but how we make sense of the data coming from the hardware. Infinite Uptime’s answer to this challenge is PlantOS, a prescriptive maintenance system that operates across five layers. The first, ‘Sense & Ingest,’ is the foundation. It collects the raw condition data of the rotating equipment in the plant from all sensors, regardless of the manufacturer or data protocols used. The second layer is ‘Edge & Connectivity,’ which ‘translates’ this data into a single format that can be processed.
The third level is ‘Platform & Execution,’ the key operational layer, which structures the raw data into digital insights that can be acted upon. The fourth level is ‘Analytics & Prescription,’ the AI model that converts the insights into evidence-backed prescription. The fifth layer is ‘Decision & Planning.’ It prioritises the prescriptions into decisions at the plant level.
GC: What are the benefits to cement producers?
KN: The aim of PlantOS is to achieve operational stasis and detect any faults at the earliest possible stage. As of 2026, PlantOS has prevented more than 160,000hr of downtime at more than 1000 industrial facilities in 29 countries. This is because our systems can typically detect future failures in equipment 4 - 48hr before conventional approaches are able to. In some cases, it can alert users up to 60 days earlier. This is powerful information that can be used to plan appropriate maintenance activities in good time and to avoid costly downtime. We have seen cases where plants have moved from 80% availability to 97 - 98% availability over two - three years of using Infinite Uptime.
GC: What about other targets?
KN: There are other reasons to use PlantOS. In Europe, sustainability is an important target. Elsewhere, the target is to increase uptime to reduce cost per tonne of cement. We have also seen considerable success on the energy efficiency side of our business. Star Cement, which operates in India and the UAE, was able to reduce specific power consumption by 7% by deploying PlantOS. Southern Province Cement in Saudi Arabia was able to increase production by 28,000t/yr with no increase in energy use.
These benefits are a result of PlantOS simply allowing the cement plant to run smoothly. This allows plants to operate at a high very level and tells users the moment something is out of place. They can prevent deterioration and maintain high standards, increasing efficiency.
We place a lot of importance on Trust Loops. This is where PlantOS makes a suggestion, the customer acts to implement the solution and then - crucially - informs the system that it has been implemented. Closing the loop adds to the training of the AI, reinforcing to it that the action that it suggested was appropriate and timely. Ideally, in the longer term, this will be a semi-autonomous or even autonomous loop, without manual intervention.
GC: Cement sector AI is a crowded market now. What is Infinite Uptime’s unique selling point?
KN: Many AI systems have false positive rates of 30 - 40%, which means that around one in every two or three recommendations will be incorrect, and does not actually need to be acted upon. Infinite Uptime has reduced PlantOS’s false positives to just 5%. Furthermore, its false negatives, meaning impending failures that are not detected by the AI, are at just 0.03%. This means that the system will alert users to 9997 out of 10,000 critical failures, enabling unparalleled reliability for equipment.
GC: What are the limits to AI-based plant control?
KN: Following discussions with clients in the sector, Infinite Uptime is of the opinion that, within the next few years – and certainly by 2030 - we will see this technology develop to the extent that cement plants will run entirely without human oversight as so-called ‘dark factories.’ Many other plants, perhaps as many as 50% globally, will be semi-autonomous, and the days of the cement plant operator as we traditionally know it will be numbered. In a business environment where the number of qualified operators is decreasing, this could be timely for cement producers, as it is increasingly difficult to find and retain high-quality staff.
Initially, human maintenance teams will carry out the recommendations of the AI systems. This is until humanoid robots become reliable and dexterous enough to carry out the work. Some may say that I am getting carried away here, but humanoid robots will arrive only slightly after full AI control of the plant becomes possible. Their arrival will mark the arrival of truly ‘dark’ cement plants.
GC: What’s next for Infinite Uptime?
KN: We are working on a battery-free measurement probe that attaches to less critical equipment. It is intended to help maintenance teams keep an eye on pumps, fans, motors, conveyors and hoppers, which will open further avenues for optimisation, efficiency and the avoidance of downtime. It will substantially reduce the need for routine inspections. The probes will be continuously powered by the equipment itself via piezoelectric crystals, as long as it is rotating faster than 100rpm.
We are also very excited about our AI Shields, which offer enhanced asset protection for the most critical - and expensive - parts of the cement, for example the kilns, mills and crushers. We are combining historical equipment reliability, operators’ notes, text, voicemails - everything relating to the equipment - and combining it for an even more holistic approach.
GC: What are Infinite Uptime’s longer-term aims?
KN: We are very ambitious. We aim to cover 10,000 plants across all industries by 2028. That’s 10 times more than the present volume. From there, we want to scale up to 50,000 plants by 2030. In the next 10 years, we anticipate that our solutions can be rolled out in more than a million factories. This how convinced we are of the benefits to be gained.
GC: That’s certainly ambitious. I wish you all the best with the expansion.
KN: Thank you for the opportunity.
About Karthikeyan Natarajan
Karthikeyan Natarajan is CEO of Infinite Uptime, a role he has held since June 2025. He has more than 32 years of experience in digital transformation of industrial enterprises, manufacturing, technology and telecoms. His previous roles include CEO and Executive Director of the Indian engineering and technology services company Cyient and at Tech Mahindra, a global technology consulting service and systems integrator. He is a graduate of the University of Madras and has a diploma in Business Management from the Symbiosis Institute of Business Management in Pune.
