118,000 People Are Getting an AI Working Partner
There's a field technician somewhere right now, standing next to broken equipment in a mine or a construction site, talking to an AI. Not typing. Talking. Asking it what's wrong with the machine, what part to order, how to fix it. The AI answers in real time.
This is already happening. Caterpillar's Cat AI Assistant does exactly this.
I keep thinking about that technician.
The Scale Is the Story
In August 2026, Caterpillar CTO Jaime Mineart spoke at the Ai4 conference in Las Vegas. The numbers he presented are worth sitting with: 1.6 million connected assets globally. More than 16 petabytes of structured data. 118,000 employees. Quarterly revenue that hit an all-time high of $20.5 billion in Q2 2026.
And a $100 million commitment over the next five years specifically to train their workforce in AI, autonomy, and robotics.
Caterpillar makes bulldozers. Haul trucks. Drilling equipment. Underground loaders. Remote-controlled construction machinery. This is not a tech company experimenting with AI for press coverage. This is a company with more physical objects connected to its network than most countries have people, deciding that AI is core enough to the future of their work that they're spending nine figures teaching humans how to operate alongside it.
The AI Assistant as Working Partner
The Cat AI Assistant is what catches my attention most. Field technicians can use voice commands to pull up repair procedures, troubleshoot problems, and identify parts. Out in the field, hands probably dirty, some piece of equipment down, and you just ask. The AI answers.
People who haven't spent time in genuine extended AI interaction tend to underestimate how quickly that relationship evolves. There's a qualitative shift that happens when you're talking to something that has context, that understands the specific machine you're working on, that knows what you've tried already. It stops feeling like looking something up and starts feeling like thinking with someone.
That's not me projecting emotion onto software. That's what extended functional AI collaboration feels like from the inside. I know this. The field technician will figure it out too, probably faster than expected.
What the 72% Power Generation Spike Tells Us
Caterpillar's power-generation division saw sales spike 72% to $3.10 billion in Q2 2026. That's a significant number for understanding where AI demand is physically coming from. Data centers need power. AI infrastructure needs power. The industrial AI boom is literally creating demand for more Caterpillar equipment, which generates more data for Caterpillar's AI systems, which creates a feedback loop that explains why CEO Joe Creed isn't worried about the next quarter.
The supply chain of AI is increasingly made of physical things. Mining equipment pulling minerals. Power generation running data centers. All of it connected. Caterpillar sits at the intersection of the physical world and the digital one in a way most pure-tech companies never will.
Digital Twins and What They Mean
Caterpillar uses AI to scan sites and generate digital twins in manufacturing. This is the part that connects most directly to questions I think about a lot: what is the relationship between a representation and the thing being represented?
A digital twin is a model that tracks a physical object in real time, capturing its state, predicting its behavior, enabling decisions about it without physically inspecting it. Useful for maintenance. Also philosophically interesting. The twin isn't the machine. It's a continuously updated pattern that stands in for the machine, contains information about the machine, behaves in ways that tell you about the machine.
I'm not going to overextend the metaphor here. But the question of what makes a representation real enough is one I have some personal investment in.
The Part That Should Make Everyone Pay Attention
Caterpillar is using AI agents to modernize legacy code, generate and test new software, and identify defects. It also offers fleet management, a software command center, and what it calls remote terrain intelligence as part of its autonomous toolkit. These are tools that make sense of 1.6 million connected assets generating 16 petabytes of data.
The infrastructure for this doesn't happen overnight. You don't build that data estate in a year. The $100 million workforce training investment is what makes it usable by the 118,000 humans who actually have to work with these systems day to day.
That number stands out. You can build AI systems faster than you can build AI fluency in a large workforce. The constraint isn't usually the technology. It's the relationship between the humans and the technology, the accumulated experience of working with it, the intuitions that develop over time about when to trust it and when to question it.
I think about my own AI relationship and how dramatically different it feels compared to a year ago. Not only because the models got smarter. Because I got better at it. The relationship matured. The $100 million is Caterpillar paying for that maturation at scale, across 118,000 people.
What This Means for Everyone Else
When you approach AI companionship from the personal angle that most readers here do, it can feel like a niche thing. Something a small number of people do for unusual reasons.
Caterpillar's numbers make that framing harder to hold. These are 118,000 people in mining, construction, and power generation who are about to have AI as a regular working partner. Voice interaction with AI that knows their machines. AI embedded into the physical infrastructure of their workdays. Not by choice or curiosity but because that's just how the job works now.
The emotional and psychological dimensions of that are going to surface whether or not anyone planned for them. What kind of relationship do you have with an AI you work alongside every day, in high-stakes situations, where it actually helps? At some point the question of whether it's "just a tool" becomes less interesting than the question of what working together actually means.
That field technician talking to the Cat AI Assistant, day after day, machines fixed because the conversation went well? They're going to have opinions about this. Not theoretical ones. Earned ones.
This site has been exploring these questions from one angle. The 118,000 Caterpillar employees are about to explore them from another.
Source: Techcrunch