Palantir spent two decades building software for the hardest operational environments on earth — war rooms, intelligence agencies, counterterrorism units. Its core argument was always the same: that the most complex real-world operations cannot be managed with generic software, and that the only way to build something that actually works is to model the real world with enough fidelity that the software can reason about it, not just store data about it.
Now Palantir is making that same argument about construction. And it has embedded its engineers inside McCarthy Building Companies to prove it.
The two companies announced a multi-year, multi-million-dollar strategic partnership on June 4, 2026. The product they are building together is called Pulse — described as McCarthy’s AI-native operations system, built on Palantir’s AIP (Artificial Intelligence Platform) and Palantir’s Ontology. The specific dollar amount of the deal has not been disclosed.
What McCarthy Is Building
McCarthy is one of the largest general contractors in the United States, headquartered in St. Louis. It has its own internal applications team — unusual in an industry where the default is to buy from Procore or Autodesk and configure around the edges. That choice matters here, because what McCarthy is describing with Pulse is not a new tool bolted onto an existing stack. It is a platform designed to run everything.
The scope is deliberately broad: field execution, estimating, contracts, bidding, QA/QC, logistics, and equipment planning. Pulse is meant to give field teams — superintendents and project managers — real-time insight, scenario planning, risk analysis, and decision orchestration across those workflows simultaneously. Dave Evans, a senior superintendent at McCarthy, puts it plainly: by cutting through the noise and identifying what is most critical, he can devote more attention to delivering exceptional results.
Tristan Gruska, Palantir’s Head of Energy and Infrastructure, has described what McCarthy is building as “an operating system for the way construction actually works, from design through site execution.” That phrase is doing significant work. An operating system is not a feature. It is the thing other things run on top of. Calling Pulse an operating system is either exactly right about the ambition — or exactly the kind of language that sounds precise but means less than it implies.
Why Palantir
The obvious question is why a general contractor partners with a defense AI company rather than an established construction tech vendor. The honest answer is that Palantir offers something the construction software industry does not: a platform already built to handle the kind of operational complexity where decisions have real-time consequences and the cost of being wrong is measured in bodies and geopolitical incidents.
That is not a perfect analogy for a job site. But it is closer than it sounds. Construction projects are, structurally, large-scale operational problems — dozens of interdependent workflows, compressed timelines, cascading risks, and field decisions made under time pressure with incomplete information. The scheduling, logistics, and sequencing problems on a major construction project are not simple. What the industry has mostly built to address them are databases dressed up with dashboards.
Palantir’s structural advantage is its Ontology — a data model that represents the real world of an operation rather than just its records. It allows the software to reason about relationships between entities: a delay in equipment delivery does not just log an event, it triggers a recalculation of schedule dependencies across the project. That is a different class of software than most of what construction currently runs on, and it is why Justin McFarland, McCarthy’s Chief Digital Officer, noted that “Palantir brings exceptional engineering talent and strategic thought leadership that has accelerated our ability to transform complex operational concepts into scalable solutions faster than we thought possible.”
The embedded-engineer model is worth noting. Palantir does not send a sales team and a configuration consultant. It sends its engineers to sit inside the customer’s technical organization. That is how it has always worked with government clients. The approach is expensive for Palantir and demanding for the customer — and it tends to produce software that is genuinely integrated rather than loosely connected.
The Ontology Is the Point
The construction industry’s data problem is not a shortage of data. Most large GCs have more data than they know what to do with — from project management systems, from field tools, from ERP, from design. The problem is that the data lives in disconnected silos with no shared model of what any of it means or how it relates to anything else.
Palantir’s Ontology is, in principle, a solution to that specific problem. It creates a unified semantic layer — a representation of people, assets, workflows, and their relationships — that allows AI and analytics to operate across data sources rather than within them. If McCarthy can actually build that model for construction operations, and if Pulse can maintain it across field execution, estimating, and logistics in real time, the result would be materially different from anything the industry currently has.
That is a large conditional. Building a coherent ontology for construction — with its project-by-project variation, its subcontractor complexity, its messy field conditions — is not a solved problem. Palantir has done it in environments where the operational entities are relatively stable: a military unit, a hospital network, a government agency. A construction project is rebuilt from scratch each time. Whether the same approach transfers is the central technical question the partnership has to answer.
What to Watch
The partnership raises a strategic question for the construction tech industry that goes beyond McCarthy. If a large general contractor can build its own AI-native operations platform — with Palantir’s engineers helping, on Palantir’s infrastructure — what does that mean for the vendors currently selling to GCs?
The traditional construction software market works because GCs are buyers, not builders. They buy Procore for project management, Oracle or Sage for ERP, Autodesk for design. They configure and integrate. They do not typically co-develop enterprise software with a publicly traded AI company. McCarthy is doing exactly that, and it has its own internal applications team capable of sustaining what gets built.
The specific figure to watch is not the investment amount — that has not been disclosed — but the scope of what Pulse actually connects over the next 12 to 24 months. If it remains a field-team tool, it is a sophisticated point solution with good marketing language around it. If it genuinely runs estimating, bidding, and logistics through the same ontology that governs field execution, it starts to look like something that changes how a GC operates at a structural level.
Palantir’s (NYSE: PLTR) whole history is a bet that the most complex operational environments eventually need software that matches their complexity. It built that case for twenty years in rooms most people never get access to. Whether a construction site is that kind of environment — or whether a simpler, cheaper stack would do the same job — is the question the industry will be watching McCarthy answer.
