Edge AI for real-time decisions

Industrial organizations make thousands of operational decisions every day. The speed and quality of those decisions determine everything from productivity and energy efficiency to equipment reliability, safety and operational resilience. Edge AI enables real-time decision making by performing AI inference close to industrial assets, reducing latency and allowing faster operational responses.

Web Story

6min

2026-07-20

01

Edge AI enables a 27 percent improvement in manufacturing efficiency,  eliminating cloud latency and enabling faster operational responses. [1]

02

Real-time data processing is the top driver for edge AI adoption, cited by 72 percent of enterprises, as industrial IoT sensors generate massive data volumes that cannot be efficiently transmitted to the cloud. [2]

03

ABB delivers real-time decision-making capabilities through its portfolio of AI-enabled edge solutions, which help industrial customers reduce downtime, achieve up to 30 percent energy savings, and maintain data sovereignty.

WHY EDGE AI, AND WHY NOW?

Every second, industrial assets around the world generate millions of data points. But waiting for data to travel to the cloud and back can introduce latency, increase operational risk and reduce efficiency. Edge AI addresses this challenge by running AI models close to the physical asset, enabling real-time decision making where it has the greatest impact. By running machine-learning models directly on, or next to, the asset - in the substation, the cooling hall, the transformer yard - industrial operators can act on intelligence in real time, even when the cloud is unreachable.³

The rapid growth of electrification, automation and industrial digitalization is generating unprecedented volumes of operational data. While cloud computing remains essential for enterprise analytics and model training, many industrial decisions cannot wait for cloud round trips.⁴

Edge AI keeps intelligence close to the asset.⁵ Data is filtered, compressed and acted upon locally; only relevant insights travel to the cloud or enterprise system. This results in decision-making speed increases of 37 percent in time-critical scenarios, operational efficiency improvements averaging 21.3 percent, and bandwidth reductions of 35–40 percent in industrial IoT (Internet of Things) deployments.⁶ 57 percent of respondents to an ABB study indicated the Industrial IoT has had a “significant positive effect” on operational decision-making.⁷

Megatrends driving edge AI in industry

Electrification and grid complexity

As renewable energy penetration increases, grid operators face a huge increase in the number of distributed assets to monitor. Edge-deployed protection is essential to maintain stability, isolate faults and restore supply at speed.

Industrial digitalization

Industrial IoT sensors on a single manufacturing line can generate gigabytes of data per hour. Shipping all of it to the cloud is neither practical nor cost-effective.

Data sovereignty and cybersecurity

Regulated industries and critical infrastructure operators are increasingly wary of sending operational telemetry to third-party cloud servers. Privacy-first, on-premise AI architectures, where no raw data leaves the site, are becoming a commercial requirement, not just a preference.

HOW ABB DELIVERS REAL-TIME DECISION MAKING WITH EDGE AI

These are some of ABB’s AI-enabled solutions for real-time decision making for industrial customers.

PRODUCTION AND INDUSTRIAL OPERATIONS

Production and manufacturing environments generate a constant stream of operational data, but data alone does not improve performance. Value comes from turning that data into timely decisions. By analyzing information close to the process, ABB`s AI enabled solutions help operators to identify anomalies, optimize performance and respond to changing conditions before they result in downtime, waste or quality issues.

ABB Genix™ Edge AI

An industrial edge computing platform this solution enables real-time data processing and analytics directly within industrial networks. Running machine learning algorithms at the edge reduces latency to sub 10ms, minimizes cloud dependency, and enhances operational efficiency by processing data closer to its source within secure industrial networks, making it a key enabler of smart, data-driven industrial operations.⁸

GRID INFRASTRUCTURE AND ELECTRICAL ASSETS

Power and distribution transformers are among the most critical, and valuable, assets in industrial facilities and electrical networks. By monitoring transformer condition in real-time, operators can identify emerging issues early, plan maintenance based on actual asset performance and extend equipment life while improving network reliability.

ABB Ability™ Uptime360 for Transformers (powered by OKTO GRID)

This is a real-time electrification asset management platform for power and distribution transformers of any make or age. By continuously monitoring sound, vibration, temperature and magnetic field directly at the asset, it provides early warning of developing faults and supports condition-based maintenance without costly shutdowns.

DATA CENTER COOLING AND ENERGY OPTIMIZATION

As computing demand grows, so does the energy required to keep data centers cool, which can account for a significant proportion of a facility's electricity consumption.. By continuously analyzing operating conditions and adjusting cooling strategies, edge AI helps operators reduce energy use, lower operating costs and support more sustainable, resilient data center operations.

OctaiPipe’s AI for Cooling Efficiency (ACE), offered by ABB

This on-premise AI solution continuously monitors cooling systems and recommends precise real-time control adjustments: no new hardware required, no operational data leaving the facility. Working with existing assets, AI for Cooling Efficiency helps data centers reduce cooling energy consumption by up to 30 percent, with impact visible within the first 90 days.

KEY CASE STUDIES AROUND THE WORLD

Italy - Lodi

Italy - Lodi
Italy - Milan
Finland

Real-time asset monitoring improves operational decisions at Italian power plant Sorgenia upgraded the monitoring of transformers and switchgear at its combined-cycle gas turbine plant in Bertonico-Turano Lodigiano using ABB Ability™ Uptime360 for Transformers and SWICOM diagnostics. Continuous, real-time visibility into equipment condition enables maintenance teams to identify developing issues earlier and make more informed maintenance decisions before faults impact operations. The shift from schedule-based to condition-based maintenance has the potential to reduce maintenance costs for key plant assets by up to 30 percent, while supporting reliable 24/7 operation of critical grid-balancing generation.

Real-time AI optimizes data center cooling and reduces energy costs An Italian co-location data center improved cooling efficiency using OctaiPipe’s AI for Cooling Efficiency (ACE), offered by ABB. The on-premise solution continuously analyses cooling performance and provides real-time recommendations that enable operators to optimize cooling while keeping operational data securely on site. The deployment reduced cooling energy consumption by 25 percent during the three-month pilot, improved Power Usage Effectiveness (PUE) from 1.52 to 1.49 within the first month, and is projected to save approximately €320,000 annually at a 2 MW facility.

Real-time transformer monitoring safeguards critical hospital power supplies The world’s 10th largest hospital, Helsinki University Hospital (HUS), strengthened the resilience of its electrical infrastructure with ABB Ability™ Uptime360 for Transformers, developed in partnership with OKTO GRID. Continuous monitoring of sound, vibration, temperature and magnetic field provides real-time visibility into transformer condition, enabling facilities teams to identify potential issues early without interrupting hospital operations. The solution supports 24/7 monitoring and early fault detection, replacing periodic manual inspections with continuous operational insights that help safeguard uninterrupted power for over half a million patients each year.

GETTING REAL WITH AI

Industrial AI only creates value when it enables better decisions at the moment those decisions matter.
Edge AI is not a future technology. It is already helping operators make better, real-time decisions, at power plants, hospitals, manufacturing facilities, and data centers. From protecting workers from arc flash in milliseconds to extending transformer life by decades and reducing cooling energy-consumption by up to 30 percent, these are practical applications delivering measurable results today.

ABB's portfolio of AI-enabled edge-solutions brings real-time capability to industries around the world, helping them reduce unplanned downtime, wasted energy and safety incidents. Because being engineered to outrun starts with making the right decision at the right time: real-time.

References:

learn more

The role of AI in energy optimization

Can artificial intelligence (AI) optimize energy usage to maximize benefits for people as well as the environment? ABB’s industrial AI solutions, including analytical AI and generative AI, (GenAI) are powerful tools to deliver AI energy optimization across industry, buildings, grids, and digital infrastructure.

The role of AI in predictive maintenance

Can artificial intelligence (AI) predictive maintenance transform equipment monitoring, anomaly detection and condition monitoring? How to employ AI to reduce MTTR and unplanned downtime.

Always On: Scaling the Future of Data Centers

ABB is helping global data center service operators outperform with a portfolio of innovative and digital technologies