At CES 2026, Caterpillar put AI at the center of the physical foundation the digital world depends on: the Cat AI Assistant, Helios, autonomy, and the NVIDIA edge. Every one of those programs is software that has to be built, maintained, and shipped faster. Here is a short overview of where Cursor fits.
Cursor was named a Leader in the 2026 Gartner® Magic Quadrant™ for Enterprise AI Coding Agents, with the highest completeness of vision among all vendors evaluated.
Cursor is a frontier AI development platform. An agent-native environment that reasons across an entire codebase, plus Cloud Agents that run work asynchronously, Bugbot for automated code review, and access to every leading model. It is built for large, complex engineering organizations, and it stays model-neutral so the choice of tool never locks your AI roadmap to one vendor.
On the CES 2026 keynote stage, Joe Creed framed Caterpillar as the invisible layer of the tech stack: the physical foundation every device, data center, and AI model ultimately runs on. That layer is now gaining intelligence with sensors, software, connectivity, and AI. Sitting behind all of it, alongside NVIDIA, is a growing body of software.
A multimodal group of AI agents on the Helios digital platform, going live this quarter to help operators, technicians, and fleet managers across every job site.
More agents, more surfaces, more code. Cursor accelerates the platform and feature work behind it, and Bugbot holds quality as agent-written code volume climbs.
The assistant runs at the edge on NVIDIA Thor so it works in remote, off-grid, and harsh environments, with Isaac and Jetson powering on-machine perception and control.
Edge software, control logic, and model integration all have to be written, tested, and maintained across a fast-moving stack. That is exactly where Cursor works.
Caterpillar previewed the next era of construction autonomy: five new autonomous machine types built on a digital nervous system of perception, edge computing, and computer vision.
That is a huge, safety-critical software surface. Cursor helps engineers navigate and safely change large, interconnected codebases as autonomy expands.
Mining is a top Caterpillar priority and Cat leads in autonomous haulage with MineStar Command for hauling. Moving faster toward fully autonomous trucks means shipping more on a big, safety-critical codebase, and Cursor helps engineers change it fast and safely.
Proven: NVIDIA committed 3x more code across 30,000 developers after standardizing on Cursor, with flat bug rates.
Helios is the future single platform for dealer and fleet data, replacing several fragmented systems. Shipping it faster pulls forward prioritized service events, and the service revenue tied to them that each business unit carries. Cursor speeds the build and makes moving off the old systems safe.
Proven: Wayfair used Cursor to cut its ML model costs by 90%, twice.
VisionLink is Caterpillar's remote monitoring and diagnostics layer, fed by dealer data through Helios. Faster dealer onboarding is what turns that data into predictive maintenance and service-event leads. Cursor speeds the apps and integrations behind it.
Proven: Coinbase reduced time from idea to production by 90% with Cursor.
AI and data center demand is driving large new power deployments, like the roughly 600 gas-engine build in Wyoming with a dedicated site team standing up. Each rollout needs strong data connectivity, predictive analytics, and parts readiness, and downtime is costly under availability commitments. Cursor speeds the software behind deployment, monitoring, and uptime.
Proven: Amplitude ships 3x more production code with Cursor across its engineering team.
Caterpillar carries years of software debt and a backlog that hiring alone will not fix. Cursor refactors legacy systems, builds context across the codebase, and automates routine developer work, including alongside your existing SAP tools with no migration.
Proven: Dropbox indexed 550k files in one monorepo to build an AI-native SDLC with Cursor.
Caterpillar is targeting $30B in services revenue by 2030, up from $24B.
That gap is software-dependent: fleet management, predictive maintenance, the Cat AI Assistant, autonomous equipment, and digital twins.
Every product requires more code, shipped faster.
CEO Joe Creed put AI at the center of Caterpillar's strategy at CES 2026 with NVIDIA, framing the invisible layer gaining intelligence across the Cat AI Assistant, Helios, and autonomy.
16,500 engineers are building against a record $63B backlog, and Caterpillar cannot hire its way out of it.
The tooling should match the ambition.
Caterpillar's complex stack spans Java, Python, TypeScript, C/C++, and SAP. Cursor uniquely understands your entire codebase, including the legacy applications and languages, and modernizes it agentically.
The leading AI model changed five times in 2025. Cursor is model-neutral. No lock-in on your AI roadmap.
Presented autonomy at CES 2026 and owns the technology behind the Cat AI Assistant and the NVIDIA edge work. Cursor keeps the software under all of it moving as fast as his teams can build, maintain, and ship it, from in-cab intelligence to construction autonomy.
Introduced the Cat AI Assistant and Helios on stage at CES 2026. Every Cat Digital surface, from enterprise data and AI to connectivity, equipment management, and eCommerce, is software that ships faster when engineers can understand and safely change any codebase.
Owns the data, AI, and GenAI platforms powering customer and dealer-facing apps, including the agents behind the Cat AI Assistant. Cursor accelerates that platform and feature work, and Bugbot holds quality as agent-written code volume climbs.
Recently stepped in with a mandate to scale AI and analytics globally, after consolidating tools and running cloud migration for a 600-person org at Toyota. Cursor is the build layer for that kind of modernization, and stays model-neutral so the AI roadmap never locks to one vendor.
An illustrative, attribution-based estimate. Caterpillar runs roughly 16,500 engineers, so the model is capped there. Adjust the inputs to your rollout. All figures are model estimates, not a commitment.
A focused, time-boxed proof of concept on one Cat Digital program, run with the teams that own the work, with success measured against your own baseline.
Anchor it on a crown-jewel program: autonomous mining, or a Helios milestone tied to prioritized service events.
A two to four week hands-on pilot with a small team on a real Caterpillar codebase, with success criteria agreed upfront.
Measure against a number Caterpillar already tracks, like autonomy code velocity or a Helios delivery milestone, against your own baseline.
Review the results together, then scale what works to the teams building the invisible layer.