Energy giant Schneider Electric is to acquire Cognite, an industrial data and AI software company, in an all-cash transaction valued at $3.1 billion. Schneider plans to integrate Cognite into AVEVA, its wholly-owned industrial software business, upon completion of the deal.

The goal is to build an AI industrial champion by combining the historical capability of AVEVA and the potential that Cognite brings, according to Schneider CEO Olivier Pascal Blum:

This new cycle has one simple goal, which is really to advance energy tech to the next level. It means really being this company in the market, which will be able to connect the physical and to the digital world by electrifying, automating and digitalizing every industry business and home with a goal to drive efficiency and sustainability everywhere.

We [are] entering in a new era, a new era where we believe intelligence is key, a new era where AI will require more compute, more compute will require more energy and energy and industrialization will sit at the center of everything we do at Schneider Electric, not only because it’s important for Schneider Electric, but it’s important really for everything we do, our customer.

So for us, it’s extremely important that we create a company that is able to connect this physical and digital world, a company where we can capture data, structure data, contextualize data and deliver those data across the life cycle. We believe that the company which will be able to structure those data on top of physical assets will be the company that will win in the market. 

So what does Cognite bring to the party here? The firm uses a unified data model and knowledge graph to bring order to operational data, while its Atlas AI offering layers on generative and agentic capabilities that can take autonomous action across industrial processes. As per Blum:

Their vision about the industrial world and how AI is going to impact the industrial world is excellent. They are benefiting to a very, very strong pool of talent in AI and they have been able really to do a remarkable performance, to deliver remarkable performance over the past years.

He breaks the Cognite core technology stack into three distinct layers:

The first layer is what we call Cognite Data Fusion, which somehow is a core foundational piece. It provides a cloud-native industrial data foundation. It ingests and contextualizes data across IT, OT (operational technology), engineering systems, creating a unified knowledge graph. So all together, this is what enables customers to unlock significantly greater value from their industrial data within a fully open ecosystem.

The second layer is really Atlas AI. That brings an industrial AI agent layer, a unique one in its industry. Through a low-code workbench, customers can deploy AI agents on top of this data foundation to automate workflow and accelerate decision-making, which drive very tangible business impact and generate efficiency for customers.

And if you go to the upper part of the stack, you have Cognite Flow, which acts as the execution layer. It allows customers, people on site to rapidly build and scale production ready, AI native workload, which empower the frontline team to translate expertise into enterprise-wide value.

So all together, those 3 layers create already stand-alone a powerful end-to-end industrial platform that turns data into intelligence and intelligence into action

Unified

It will provide a unified platform for the next phase of industrial AI, predicts AVEVA CEO Caspar Herzberg:

In industrial software, Cognite is the largest and probably quite unique company in that space, which is why we’ve been after it, frankly, for many, many years. It complements us so well. If you’re looking at the more analytical level, I would look at the companies like Databricks, Snowflake and some of the hyperscalers as comparative.

Basically, what enterprises do with industrial data in the cloud in an AI agentic way is a new area, he contends:

When you look at the design, build, operate, and optimize lifecycle of industry and you translate this life cycle into data, then you will see on the design side, models and assets. These assets are at times very, very large, millions of data points if they are big process plants. Then on the operational side, you have the time series data that comes from how machines fare when they produce something. This is vibration data, the experience of the machine, if you like. And then, of course, you have optimization data, which is what you do with all of this data, asset and operational data and other IT and enterprise data.

The challenge, of course, with these enormous amounts of data is how to scale this, while then dynamically maintaining the context of that data to each other, the context of plant-to-machine, of time series data like a certain temperature on a certain day ten years back. [It’s about] how to maintain that and how to dynamically model that in the cloud?

What Cognite does for its customers is take this fragmented complex industrial data and integrate it into a single unified data model and most critically, a knowledge graph, supported by the Agentic AI workbench where you can use algorithms to dynamically model this without a lot of coding.

As to what’s in it for Cognite, Herzberg says:

AVEVA significantly accelerates Cognite because what we bring them is immediate scale across all of the large industrial segments that AVEVA and Schneider are currently supporting. And of course, [there is] the global footprint on top of that that Schneider Electric has with its countries and its strong go-to-market engine. In addition, we have an installed base of more than 23,000 customers. So that gives, in summary, Cognite a reach, a depth and a scale to accelerate adoption and to bring the innovation they have to, very often our joint, customers and increasingly as they move towards Life Sciences, out of the pure process industries, to new customers.

My take

The combination of AVEVA and Cognite is pretty much peerless, I believe, at this point in time.

A big move from some big players as industrial AI becomes a reality.

Game-changer? Let’s check back in a few months.