KONE Commits to Connecting 1.8 Million More Lifts and Escalators

KONE and Amazon Web Services announced a seven-year extension of their strategic collaboration on 28 January 2026, aimed at AI-powered digital innovation in smart building solutions. The agreement covers accelerating cloud connectivity for 1.8 million additional devices. KONE states that more than 40% of its global device network — hundreds of thousands of elevators and escalators — is already connected through its digital platform.

That connectivity underpins the company's connected services offering, its real-time monitoring and predictive maintenance product, and KONE is migrating its product data management suite to the cloud. Stated priorities include embedding AI across the global service portfolio and deploying machine learning and analytics for operational efficiency. KONE reported €11.2 billion in 2025 annual sales with more than 60,000 employees across nearly 70 countries.

The number worth sitting with is 1.8 million. Vertical transport is one of the largest installed bases of safety-related mechanical equipment in the built environment, it is maintained under statutory inspection regimes, and it has historically been serviced on a calendar. Connecting it changes the service model from scheduled visits to condition-driven ones, and at this scale the economics of that shift are substantial — a technician not sent is a saved visit, and a fault found before a entrapment is a different kind of saving.

For integrators the strategic reading is about who owns the layer. This is an OEM building its own connected platform across its own installed base, which is the pattern across vertical transport, and it means third-party connectivity work in this sector is mostly either interoperability with an OEM stack or displacement of one — rarely greenfield. A building operator with lifts from three manufacturers will have three platforms, and the integration question is whether they can be surfaced together in the building's own systems.

The other observation is more general. A device population of this size, connected and monitored, generates exactly the kind of comparative dataset that makes condition monitoring work: identical equipment, many instances, long history.