Predictive maintenance and industrial IoT platforms

Listening to the machine, and catching a fault before it fails

A good mechanic can tell that a motorcycle engine is in trouble from its sound. Machines in a factory give the same kind of warning before they fail. A vibration changes. A bearing runs hot. A pump draws more power than usual. Software now listens for these signs, all day, on thousands of machines at once.

Why it matters. When a machine fails without warning, the whole line can stop. The factory loses output every hour, and an emergency repair costs more than a planned one. So factories want to know in advance which machine will fail, and when. This is called predictive maintenance.

How the listening works

Sensors on the machine measure vibration, temperature, pressure and electric current.

A gateway, a small computer near the machine, collects the readings and sends them on. Connecting machines to the internet this way is called the industrial internet of things, or industrial IoT.

Software studies the readings. Each reading comes with its time, so the data is called time-series data. The software learns what normal looks like for each machine. It raises an alert when the readings drift away from normal. Such a drift is called an anomaly.

The maintenance team gets the alert, with a view of what is likely to fail and when.

How the listening works: sensors measure the machine, a gateway sends the readings on, software learns what normal looks like and spots an anomaly when the readings drift away from it, and the maintenance team gets the alert.

The industry calls this kind of software asset-performance management. An asset is a machine or a plant.

Two ways to get the data. Some companies fit their own sensors on the machines. This works even on old machines that record nothing. Others use the data a plant already collects, in its control systems and historians. (More on those in Market Segment 37.1, Industrial automation and control software.)

Two ways to get the data: companies such as Augury and Shoreline AI fit their own sensors on the machines, while Falkonry and Cognite use the readings a plant already collects; either way, software such as Octave, UptimeAI Inc. and Avathon studies them to predict failures.

Who buys, and how. The buyers are factories, refineries, power plants and mines. The software is usually sold as a subscription. It is a field of many small companies, and India has produced a good number of them.

Every company named here has posted software engineering jobs in India. Famous companies that don't actively hire software engineers in India are left out.

Keep reading, free

One more section is on this page: Sensors, platforms and AI for maintenance. Sign in to read it here, in full.

Continue with Google

Sensors, platforms and AI for maintenance

Sensors, platforms and AI for maintenance

Asset-performance software

Octave

with engineers in Hyderabad, sells asset-performance and maintenance software. It is the business that Hexagon separated from itself in 2025. Octave has the most job postings of any company covered mainly in this Market Segment. (More on Hexagon in Market Segment 37.2, CAD and PLM software.)

Their own sensors on the machine

Augury

with engineers in Bengaluru, and Shoreline AI, with engineers in Pune, fit their own vibration and temperature sensors on machines. They predict failures from the readings.

The data the plant already has

Falkonry

with engineers in Mumbai, finds anomalies in the time-series data that plants already collect. (More on it in Market Segment 6.3, Observability and monitoring platforms, in Industry Vertical 6, Devtools companies.)

Four Indian companies that bring AI to the plant

UptimeAI Inc.

with engineers in Bengaluru, sells an AI expert that works out what is wrong with a piece of equipment.

Detect Technologies

with engineers in Chennai, uses cameras and AI models to watch plants for safety risks and corrosion, the slow eating away of metal.

Raven

with engineers in Bengaluru, makes interactive diagrams of plants, with an AI assistant. (More on it in Market Segment 37.2, CAD and PLM software.)

Avathon

formerly SparkCognition, with engineers in Bengaluru, sells industrial AI for maintenance and operations.

Platforms and devices

Datoms

with engineers in Bengaluru, and HBeonLabs Technologies, with engineers in Delhi, sell IoT platforms and gateways for monitoring machines.

ScratchNest

with engineers in Delhi, makes RFID readers and temperature loggers. An RFID reader reads small radio tags fixed to goods. A temperature logger records the temperature over time, for example inside a cold truck. (More on it in Market Segment 30.4, Shipment tracking and supply chain visibility platforms, in Industry Vertical 30, Supply chain and logistics.)

Accurate Technologies Inc.

with engineers in Bengaluru, makes tools for the engineers who develop engines. The tools record data from an engine and adjust its settings.

Two more companies work in this area: Bytebeam, in Bengaluru, and Bert Labs. (More on Bert Labs in Market Segment 40.7, Carbon accounting and sustainability software, in Industry Vertical 40, Energy and climate tech.)

Two companies covered mainly in other Market Segments also work here:

Cognite

with engineers in Bengaluru, sells an industrial data platform, which gathers a plant's data in one place. Cognite has more job postings than any company covered mainly in this Market Segment. (More on it in Market Segment 8.1, Data integration and governance platforms, in Industry Vertical 8, Analytics and data providers.)

Nextpower

formerly Nextracker Inc., with engineers in Hyderabad, writes the software that steers its solar trackers. (More on it in Market Segment 40.1, Renewable energy and battery storage, in Industry Vertical 40, Energy and climate tech.)

The maintenance engineer's assistant.

Predictive maintenance promised for years to foresee failures. Often it buried engineers in alerts instead. AI models now explain an alert, search the manuals and the plant's history, and suggest the fix. The value is moving from the sensor that detects a problem to the assistant that decides what to do. India's start-ups in this field are building such assistants.

That is the business of listening to machines: sensors that measure, software that spots an anomaly, and AI that suggests the fix. It is a field of many small companies, and a good number of them are Indian.

Who they hire

Who these companies hire, and for what, is on What manufacturing tech companies hire for.

← Previous in Manufacturing tech

Manufacturing execution systems (MES)

Next in Manufacturing tech →

Industrial machinery and robotics