Open an AI chatbot and the interaction feels almost weightless. A prompt leaves your phone, a response returns, and somewhere between the two sits a word called “cloud.” The word hides an awkwardly physical reality: acres of servers, voracious cooling equipment, high-voltage connections and a power system that cannot blink. ABB does not make the chips or train the models. It makes much of the industrial layer that lets those things function.
That layer reaches well beyond computing. In a cement kiln, a cruise ship, a hospital, a paper mill or a municipal water plant, electricity must be distributed safely, motors must turn at the right speed, and processes must respond to thousands of signals without drama. ABB sells the breakers, switchgear, motors, drives, controllers, instruments and software that do this work. It also sells the engineering and long-term service that keep them useful after the launch event has been forgotten.
The Zurich company was created in 1988 by the merger of Sweden's ASEA and Switzerland's BBC Brown Boveri, but both family trees reach back to the 19th-century electrical age. This gives ABB an odd dual identity. It is young enough to be a product of modern corporate strategy and old enough to have helped build the infrastructure that modern strategy assumes will always be there.
The machinery behind the machinery
A useful way to understand ABB is to imagine a factory as a body. Electrification is the circulatory system, moving power from the incoming connection to the exact place it is needed. Motion is the muscle: motors, generators and drives turn electricity into movement. Automation is the nervous system, reading instruments, making decisions and adjusting the process. Software watches the whole organism for waste, wear and unusual behavior.
The invisible industrial stack
Consider the variable-speed drive, an uncelebrated box with an outsized job. Without one, a pump or fan may run at full speed while a valve or damper restricts the output - the industrial equivalent of driving with the accelerator down and controlling speed with the brake. A drive matches motor speed to actual demand. The result can be less electricity, less mechanical stress and finer control. Multiply that logic across factories, mines, buildings and data-center cooling plants, and energy efficiency stops being a slogan and becomes an operating decision repeated every second.
Who pays, and what they are buying
ABB's customers are organizations for which downtime is expensive, unsafe or publicly embarrassing. Utilities use its grid and distribution products. Machine builders put ABB controls and drives inside equipment sold under other names. Manufacturers automate lines and monitor assets. Mining and process companies rely on control rooms and instruments. Ship operators use propulsion, power and automation systems. Building owners install electrical distribution and smart controls. Data-center operators need power protection, cooling control and a path to add capacity without rebuilding every room.
The sale begins with hardware, but the business model stretches across an asset's life. ABB can help specify a system, commission it, connect it to monitoring software, supply spare parts, train operators, repair equipment, modernize controls and wrap the arrangement in a service agreement. An industrial motor or distributed control system may remain in service for decades. That longevity creates repeat work and makes trust more valuable than a clever quarterly promotion.
The scale is substantial. ABB reported $33.22 billion in 2025 revenue, more than 160 manufacturing sites and 111,900 employees. Its customers are spread across more than 100 countries. In the second quarter of 2026, orders reached a record $12.04 billion, up 30 percent from a year earlier, while quarterly revenue rose 14 percent to $9.48 billion. These are not startup numbers, but the demand drivers sound surprisingly current: electrification, grid modernization, industrial productivity and the power appetite of artificial intelligence.
A physical growth signal / USD billions
Breadth is useful only when it connects
ABB competes with different companies in each aisle of the industrial store. Schneider Electric, Siemens and Eaton meet it in electrical distribution. Rockwell Automation, Emerson, Honeywell and Yokogawa compete in control and automation. Motor and drive makers fight over efficiency, reliability and channel reach. Until the agreed sale of ABB Robotics closes, Fanuc, Yaskawa and KUKA remain robotics rivals. No comparison works cleanly across the entire portfolio.
Its distinction is the ability to connect multiple layers with application knowledge. A competitor may make an excellent component. ABB can often offer the component, the controller, the digital monitoring and an engineer who understands why the combination behaves differently on a mining conveyor than on a ship. That is especially useful when electrical and operational technology converge. It is also difficult to reproduce quickly because much of the knowledge lives in field experience, installed equipment and customer relationships rather than a patent database.
The moat is cumulative
A broad catalog matters, but the harder asset to copy is the loop between installed equipment, service experience, engineering know-how and the next product revision.
The company is trying to make old industrial systems more open without asking customers to gamble a running plant. Automation Extended, launched in 2026, separates mission-critical control from a more flexible digital environment. The idea is to add analytics, AI and new applications while preserving the deterministic core that keeps a process safe. For an office app, a restart is annoying. For a chemical process or power station, continuity is part of the product.
A portfolio in motion
ABB's current shape requires one important footnote. In October 2025, it agreed to sell its Robotics division to SoftBank at a $5.375 billion enterprise value. Robotics continues to operate within ABB as a standalone discontinued operation while the transaction moves toward an expected mid-to-late 2026 close, subject to approvals. ABB has not stopped making robots overnight, and the distinction matters: a signed transaction is not a completed one.
At the same time, ABB is buying deeper into its remaining thesis. The company completed acquisitions in energy management, power converters, Chinese wiring accessories and renewable power electronics. In July 2026 it agreed the terms of a recommended cash offer for Rotork, a British maker of electric actuators, flow-control equipment and instrumentation, at an enterprise value of about $5.5 billion. If completed, the deal would add more field devices to ABB's automation stack - the equipment that physically moves valves and measures what is happening in pipes and processes.
The swap is telling. Robotics is visually compelling and increasingly linked to physical AI. Flow control is easier to overlook. ABB's decision suggests it values fit and shared routes to market over fame. The intended company is organized around three connected areas - Electrification, Motion and Automation - with a decentralized operating system called the ABB Way. Divisions are expected to act close to customers and carry clear responsibility for performance.
Where ABB fits now
ABB sits between component specialists and giant diversified engineering groups. It is not a pure software company, though roughly half its R&D workforce now focuses on software, AI and digital work. It is not simply a machinery maker, because much of the value lies in power electronics, control logic, data and service. And it is not a utility-equipment company in the old sense, because its customers increasingly include AI campuses, smart buildings and flexible factories.
Its climate role is similarly practical. Renewable generation must be converted and connected. Grids need protection and automation. Existing motors can use less power. Old equipment can be modernized instead of scrapped. ABB reported that 98 percent of the electricity it purchased in 2025 came from renewable sources and that its own Scope 1 and 2 emissions were 79 percent below the 2019 baseline. The larger potential impact, however, sits with customer operations, where small efficiency gains repeat across enormous installed fleets.
There is no need to romanticize the company. Industrial cycles turn, projects slip, acquisitions carry integration risk, and the competitive field is crowded. A sprawling catalog can create complexity as easily as it creates cross-selling. The Rotork offer and SoftBank sale also make 2026 a year of unusual portfolio execution. What makes ABB interesting is more grounded: the fashionable parts of the economy keep colliding with constraints that old engineering companies understand.
Every larger AI model intensifies a chain of physical questions. Is there enough grid capacity? Can the facility distribute power safely? Can cooling respond efficiently to volatile loads? Can operators see trouble before equipment fails? ABB will not answer all of them, but it can answer enough to be invited into the room early. That is where the 19th-century electrical lineage meets the 21st-century data center - not as nostalgia, but as a purchase order.