Caterpillar’s AI Vision Takes Shape, but Data, Cloud Focus Was the Precursor

Caterpillar has plans for artificial intelligence (AI)–enabled features in its construction equipment; autonomy at scale and machine learning with computer vision, sensors, and edge computing; and Cat AI Assistant, which enables customers to engage with the company’s machines. The bigger story, however, may be the years of transformation, data preparation, and technology infrastructure leading up to Caterpillar’s big CES 2026 splash. 

The company has built its stack around three key technology partnerships: NVIDIA is key to Caterpillar’s AI plans, Caterpillar’s cloud and software stack runs on Microsoft, and Snowflake is the data platform that has enabled the company’s AI efforts. 

Caterpillar’s transformation highlights how there are no overnight successes in enterprise AI. Without a focus on data and cloud transformation, Caterpillar’s AI splash at CES 2026 wouldn’t have happened. 

Get this customer story as a PDF | More research

“The digital world depends on a physical layer most people never think about. Every device in this room that you have depends on minerals that had to be pulled from the ground. Every data center behind the AI you’re going to see this week or you’re going to hear about this week was constructed from the ground up, and it stays online with power systems that provide reliable electricity,” said Caterpillar CEO Joe Creed, speaking at CES 2026. “Every road, every port, every power line connecting our economy had to be built. That’s the invisible layer of the tech stack, the physical foundation for modern technology.” 

Caterpillar CEO Joe Creed speaking at CES 2026
 

Creed wants people to see Caterpillar differently. He noted that Caterpillar is a technology company and added that the company has a connected fleet of more than 1.5 million assets in the field. Caterpillar is a physical AI play, and its assets in the field “generate data so that we get better with every hour of operation,” he said. 

Here’s what Caterpillar announced to highlight how it plans to layer AI throughout its products, which are used for real-world construction. 

Caterpillar is partnering with NVIDIA to roll out AI-driven features in its products; manufacturing systems; and network of customers, dealers, and employees. Caterpillar highlighted upgrades including:

• In-cab AI features with an intelligent operator assistant that’ll provide insights, coaching, and safety alerts for operators 
• Autonomy embedded across its construction and mining machines 
• Machine intelligence via sensors in the field across Caterpillar’s fleet and job sites
• Excavators, loaders, haul trucks, dozers, and compactors that will have AI-enabled features, and Cat VisionLink and Cat MineStar systems that will connect fleets and coordinate them 
• The Cat AI Assistant (built on NVIDIA’s Riva open speech models), which will be embedded into Cat digital and onboard products 

Caterpillar is using its NVIDIA AI Factory to transform its manufacturing and supply chain operations with automation. Caterpillar also said the NVIDIA AI Factory will improve manufacturing processes, forecasting, and scheduling.

The company is also building digital twins of its factories, using NVIDIA Omniverse libraries and OpenUSD.

Caterpillar’s Platform

Ogi Redzic, Caterpillar chief digital officer and senior VP, outlined Helios, an event-driven  cloud platform that brings together Caterpillar’s global fleet and 1.5 million connected assets. Helios “sends and receives thousands of messages every second and triggers millions of data pipelines a day to ingest, organize, and deliver this data where it needs to go. It now houses more than 16 petabytes of reusable high-quality data,” he said. 

Helios was the linchpin of launching Cat AI Assistant. “We’re building this with Helios on top of the latest AI tech. We’re building a fleet of AI agents that can see the state of our customers’ fleets, understand what’s happening, and take actions through our applications or APIs [application programming interfaces]. What matters most, however, is what this all feels like in real life. For a customer, Cat AI Assistant is like a proactive partner. It flags machines that need attention, provides custom insights, and makes actionable recommendations,” said Redzic. 

Caterpillar is leveraging NVIDIA’s Thor AI robotics platform to run models on equipment even when a cloud connection isn’t available.

Jaime Mineart, senior VP and CTO at Caterpillar, said the company is pushing toward autonomous equipment across its network. Caterpillar has invested in technologies and delivered Level 4 autonomy across its fleet for more than a decade. 

“Caterpillar’s autonomous mining fleet is one of the most proven and largest in the world, moving over 11 billion tons of material, traveling more than 385 million kilometers autonomously. That’s over twice the autonomous mileage of the automotive industry, without a single reported injury,” said Mineart. “We did it because the industry was changing fast and with growing demand, labor challenges, and a constant need for greater productivity. Autonomy helps our customers scale up, control their costs, and protect their people.”

In other words, Caterpillar’s heavy equipment is part of the data and AI platform now. 

Moving forward, Mineart said construction is a big opportunity for autonomous operations. Autonomy can improve safety as well as productivity. “Thanks to advances in computing and AI, many of them powered by partners like NVIDIA, machines can make split-second decisions based on billions of data points. This accelerates human validation and informed oversight, ensuring speed without sacrificing control even in the middle of chaos,” said Mineart. “This is the breakthrough that will transform construction forever, making sites safer, faster, smarter, and more efficient than ever before. And that transformation starts now.”

Caterpillar highlighted its autonomous equipment that will be rolling out in the weeks ahead. 

“Tomorrow’s operators won’t just run machines; they’ll orchestrate entire job sites, intervening when human judgment matters most. We’re creating new roles on the job site. Tech talent isn’t just for Silicon Valley. It’s in mines in Western Australia. It’s in quarries in Virginia and soon to be on construction sites in South Korea and road crews right here in Las Vegas,” said Mineart. 

Caterpillar’s use of data and AI, which revolves around productivity, the value chain, and autonomy 

How Caterpillar Got Here and Where It’s Going

Caterpillar’s enterprise and physical AI coming-out party was just a few weeks after executives held an analyst and investor day that outlined how the company was layering data and AI throughout its business. Caterpillar’s investor update in November cited gains since it laid out its strategy in 2017.

Creed detailed how Caterpillar had doubled its services base from 2016 to 2024, developed an ecommerce platform that now conducts $18 million in transactions each day, and built up its connected assets. Creed said Caterpillar’s competitive advantage is its 1.5 million connected assets that churn out data that creates new products and services. 

• Industrial AI investments start to pick up
• IFS teams up with Anthropic, Siemens, Boston Dynamics, 1X, showcases industrial AI use cases
• Anthropic: Claude eyes real world impact, industrial use cases

Bob De Lange, group president of services, distribution, and digital at Caterpillar, said technology will be critical to the company’s success and that the company is stepping up its investments. De Lange added that the company is partnering to keep up with the pace of AI. 

“As the technology environment is changing very rapidly—in some cases, even by the month—we’re also partnering with some of the leading companies in the world, such as NVIDIA, to bring AI to manufacturing and onto our machines. Microsoft has been a long-standing partner for everything related to cloud and productivity. And for organizing Helios, everything related to organizing our data, we are working with Snowflake,” said De Lange.

Caterpillar’s focus on long-term technology partnerships 

De Lange said the next stop is to get to two million connected and reporting assets by 2030 and grow ecommerce revenue by 50%. “We want to achieve at least 500,000 tech-enabled machines. And with that, I mean machines that have at least one or more of our advanced technologies on board. And then we want to triple the number of autonomous trucks,” he said.

Via customer value agreements, the data flywheel with customers aims to address multiple customer pain points. Here are few examples:

• Data from equipment has enabled better service. Caterpillar said sensors on equipment have streamlined prioritized service events (PSEs). Sensors can find looming service issues and offer predictive maintenance services. In 2021 Caterpillar didn’t have PSEs. In 2024, by contrast, Caterpillar had $1.1 billion in sales from PSEs. 
• Ecommerce sales have surged, because Caterpillar has integrated with the enterprise resource planning (ERP) systems owned by its largest accounts. The systems offer parts based on a customer’s specific serial number. 
• Caterpillar has 690 autonomous mining trucks in operation today. 
• The company has 16 petabytes of data in Helios, 30 data sources, and more than 50 billion records ingested monthly. Those data points are fed into Caterpillar’s AI engine to generate qualified sales leads sent to dealer customer relationship management (CRM) systems. 
• One-third of Caterpillar’s total connected fleet has a digital twin. 

A look at Caterpillar’s data flow for delivering value

According to Caterpillar, urbanization and its need for infrastructure, digitization with the need for data centers, and the need for power will all drive growth in the future. Caterpillar has three units for construction, mining, and power/distribution. Creed said customers in all three categories struggle with safety risks, labor shortages, sustainability, and costs. 

“We need to be the clear leader in advanced technology, both on and off machine, and we’re going to transform how we work internally to get the job done,” said Creed. “We’re going to get closer than ever to our customers. If we’re going to solve their toughest problems, we have to understand their operations and their job sites, and we can’t do that unless we get closer to them. We are going to lean into technology in a heavy way moving forward. And I believe by integrating our off-machine digital solutions with our on-machine technology, we’ll unlock tremendous value for our customers.”