Procore is the default. Palantir is the replacement. In Q4 2025 alone, Palantir signed $1.34 billion of new US commercial contracts, a 67% jump year on year. Its construction-specific offering went live earlier this year. Thomas Cavanagh Construction, a Canadian contractor with 350 employees and $120M in revenue, has quietly ripped out its ERP and rebuilt its entire operation on Palantir Foundry. “97% of our employees use Foundry every day,” their AI lead said on stage at Palantir’s Paragon 2025 conference. “Every other software must justify its existence. So far they haven’t been able to.” That is a direct shot at every ERP, project management platform, and point tool in the stack. And the top contractors in America are listening.
Table Of Content
- The Case Study That Should Worry Every ConTech CEO
- Why ConTech Incumbents Are Structurally Vulnerable
- The Pricing Objection Is Already Losing
- What Happens to Procore, Viewpoint, Autodesk, and the Rest
- The Boring Counterargument (And Why It Is Wrong)
- Palantir vs. the Construction Software Stack: Side by Side
- Frequently Asked Questions
If you run IT or operations at a tier-one contractor, you have probably had three meetings about AI agents this quarter. You have also probably had zero meetings about the company most likely to eat your back office over the next five years. That company is Palantir. And it is not pitching you a new Procore feature. It is pitching a full replacement for the fragmented stack of ERP, project controls, document management, and scheduling tools you have been duct-taping together for twenty years.
This is not a projection. Palantir for Construction launched as a formal offering in 2026, joining the company’s existing verticals across defence, manufacturing, healthcare, and energy. Its first published reference customer, Thomas Cavanagh Construction, is not a Fortune 500 builder. It is a Canadian civil contractor with 350 employees. If Palantir can displace the ERP at Cavanagh’s size, the playbook scales upward fast. And the economics support it: US commercial revenue at Palantir grew 137% year on year in 2025, with a single quarter of commercial bookings now larger than Procore’s entire annual revenue.
The construction industry has been having a quiet software crisis for a decade. Procore plateauing. Viewpoint and Sage still charging $100,000 to $400,000 in first-year costs for interfaces built in the 2000s. Trimble bolting products together through acquisition. None of this has been solved by another vertical SaaS point solution. Palantir’s bet is that it will be solved by a platform that treats the contractor’s entire business as a single, queryable data model. The early evidence suggests they are right.
The construction software incumbents have spent a decade on product roadmaps that amount to “our old thing, plus a chatbot.” Palantir is quietly skipping that category entirely and selling contractors a new operating system. The top builders are taking meetings, the first deployments are live, and the numbers say the rest of the market is next.
The Case Study That Should Worry Every ConTech CEO
Cavanagh rebuilt a contractor on Palantir and did not need Procore or Viewpoint to do it
Thomas Cavanagh Construction is the kind of contractor most software vendors chase. Mid-market civil infrastructure, Canadian operations, steady book of work, roughly $120 million in annual revenue. Exactly the profile that a Procore sales rep would plan a site visit for. In 2023, Cavanagh selected Palantir Foundry to replace its legacy analytics and business intelligence systems. By 2025, it had gone significantly further, publishing a co-authored case study with Palantir describing the full removal of its ERP from daily operations in favour of what the team called Total Operations Management, or TOM, built entirely on Foundry and Palantir AIP.
The public framing from Cavanagh’s AI lead Joe Patrois at Palantir’s Paragon 2025 conference was explicit: the company has gone “all in so much so that every other software must justify its existence. And so far they haven’t been able to. 97% of our employees use Foundry every day. Foundry is our operating system.” Read that one more time. A contractor with 350 people operates inside a single software environment that is not a construction-vertical product. It replaces the analytics layer, the ERP layer, and increasingly the operational workflow layer. Dispatch, trucking, site management, cost reconciliation. All on one platform.
The Cavanagh framework rewires the basic economics of running a contractor. Instead of manual cost coding and reconciliation after the fact, the TOM model captures activity-based costing as work happens. Instead of dashboards reviewed in weekly meetings, reasoning sits on top of live operational data, with large language models ready to take autonomous action when appropriate. The ontology layer, which Palantir has spent two decades building for government and Fortune 500 clients, makes contracts, resources, and costs queryable in real time. That is not a feature Procore can add to its roadmap. It is a fundamentally different category of product.
Why ConTech Incumbents Are Structurally Vulnerable
The stack that won 2015 is the stack that loses 2026
Procore closed out 2025 with $1.323 billion in revenue, up 14% year on year. That is a respectable number, but it is a long way from the 30%-plus growth the company posted in its IPO years. The stock has drifted sideways since mid-2021. Tooey Courtemanche stepped aside as CEO in November 2025. His replacement, Ajei Gopal, inherited a company that is profitable but no longer defining the narrative. As Bricks & Bytes covered at the time in our Procore Slows, Trimble Soars piece, the slowdown is real and the bathtub of AI hype is not fixing it.
The incumbent ERPs are in worse shape. Viewpoint, now inside the Trimble Construction One suite, still reportedly costs mid-sized contractors between $100,000 and $400,000 in year-one implementation and licensing, according to procurement data from Vendr. Sage 300 CRE carries a three-year total cost of ownership between $47,000 and $132,000 for mid-market users, with a reputation for a dated interface and a steep learning curve. User reviews on Software Connect describe documentation as “completely out of date” and warn of a $500,000 floor for a serious setup. This is what contractors are replacing when they look at Palantir. Not modern cloud tools. Twenty-year-old systems bolted onto incremental acquisitions.
The broader construction software market is equally exposed. CMiC, Acumatica Construction Edition, Jonas, Foundation. All credible products, all competing with each other on feature parity while Palantir pitches a completely different product category. A decade of vertical SaaS consolidation created an industry of mid-sized software winners, none of whom are structurally built to compete with a horizontal ontology platform that gets smarter every quarter. When the same conversation was happening in defence, Palantir won. When it happened in pharma, Palantir won. Construction is not magically different.
We’ve gone all in so much so that every other software must justify its existence. And so far they haven’t been able to. 97% of our employees use Foundry every day. Foundry is our operating system.Joe Patrois, AI Lead at Thomas Cavanagh Construction, speaking at Palantir Paragon 2025
The Pricing Objection Is Already Losing
“It’s too expensive for construction” stops working when owners are writing the cheque
The standard objection is price. Palantir’s average commercial contract is measured in millions, with many enterprise deals in the $5 to $10 million range. That is a tough sell for a regional contractor with 2 to 3% net margins. But the objection misses two structural points. First, the customer base that matters most for Palantir in construction is not regional contractors. It is owners. Hyperscalers building data centres, sovereign wealth funds running megaproject portfolios, energy majors coordinating capital programmes across continents. For these customers, a $10 million software contract is a rounding error inside a $10 billion capex programme. And they are already paying Palantir for other workflows.
Second, the cost of the legacy stack compounds. A tier-one contractor running Viewpoint, Procore, Primavera P6, a separate BI tool, an estimating platform, and two middleware connectors is easily spending $1 to $3 million per year in licences and implementation services. Layer in the hidden cost of the eight separate systems that do not talk to each other, and the number gets worse. Replacing that with a single Palantir deployment does not reduce spend to zero. It changes which vendor captures the margin, and it consolidates workflow in a way that point tools structurally cannot.
The other pattern worth watching is land-and-expand. Palantir does not typically start at $10 million. It starts at a specific operational problem, proves out the ontology against that workflow, and expands into adjacent processes. Cavanagh began with analytics and BI. Two years later they are replacing the ERP. The commercial momentum is built into the product, not the sales pitch. Every contractor that lets Palantir in the door for a narrow use case ends up looking at a TOM-style deployment three years later.
What Happens to Procore, Viewpoint, Autodesk, and the Rest
Three scenarios, one clear trajectory
We are recording an upcoming episode on this exact shift
The Palantir-into-construction deep dive drops on Bricks & Bytes soon. Subscribe to get the episode, the guest list, and the breakdown the moment it is live.
Join 2500+ ReadersThe Boring Counterargument (And Why It Is Wrong)
Yes, implementation is hard. No, that does not protect the incumbents.
Every ConTech analyst who dismisses this thesis says the same three things. Palantir contracts are too expensive. Implementation takes too long. Construction is too conservative. All three are true. And all three missed what happened in every other vertical Palantir has won in.
Defence contractors said Palantir was too expensive. Now the US Navy just awarded Palantir up to $448 million to modernise shipbuilding supply chains. Pharma said implementation would take years. Now Palantir Foundry is production infrastructure at multiple top-10 pharmaceutical companies. Banking said regulators would never approve it. Now Morgan Stanley, BP, Airbus, and dozens of others run major operational workflows on Foundry. The pattern is identical across every sector where the company has landed: the first deployment is painful, the second is faster, and by the fifth the incumbent software vendors are no longer invited to the renewal meeting.
Construction has one unique advantage the incumbents keep citing: project fragmentation. Every job is a one-off, every contract is bespoke, every site operates differently. That is exactly the problem Palantir’s ontology was built for. It is not selling a template solution. It is selling a framework for modelling complex operational reality, which is literally what governments and defence agencies use Foundry for. If you can model a carrier strike group, you can model a general contractor. And unlike the defence contracts, this one comes with a commercial tailwind in the form of data centres and hyperscaler megaprojects where the owners are already Palantir customers.
Palantir vs. the Construction Software Stack: Side by Side
What each layer looks like before and after
| Layer | Incumbent today | Palantir equivalent | Strategic risk to incumbent |
|---|---|---|---|
| ERP / financials | Viewpoint, Sage 300 CRE, CMiC | Foundry ontology + AIP workflows | High – directly displaced at Cavanagh |
| Project management | Procore, Autodesk Construction Cloud | Foundry operational apps on top of the ontology | High – becomes a data source, not a system of record |
| Scheduling | Primavera P6, Planera, ALICE | LLM-based scenario analysis inside Foundry | Medium-high – best-of-breed specialists survive, generalists die |
| BI / analytics | Power BI, Tableau, PowerPoint | Foundry Contour, Object Explorer | Already replaced at Cavanagh in 2023 |
| Document management | Egnyte, SharePoint, Bluebeam | Foundry document ingestion + AIP retrieval | Medium – retrieval becomes commodity, storage survives |
| Field capture | OpenSpace, Buildots, Track3D | Data feeds into the ontology, ideally via partnership | Low-medium – specialists remain valuable if they integrate |
Frequently Asked Questions
For operators, founders, and investors trying to read this correctly
Thomas Cavanagh Construction, the first published construction reference customer, has 350 employees and roughly $120 million in annual revenue. That is mid-market by any reasonable definition. The economics of a deployment at that scale will still lean on consolidating multiple existing software spends, not adding a new budget line. (Source)
Public disclosure is limited, but Palantir closed 204 deals of at least $1 million in Q3 2025, 91 deals over $5 million, and 53 deals over $10 million. The commercial sweet spot is low single-digit millions in first-year ACV with expansion. For owners running capex programmes, those are rounding errors. For smaller contractors, Palantir has historically started with narrower engagements before expanding. (Source)
The most likely outcome is that Procore becomes durable but not strategic. Its installed base is real and its $1.3 billion in ARR is not going away overnight. But if Palantir becomes the system of record at tier-one contractors and owners, Procore ends up as a best-of-breed project management product that feeds data into someone else’s ontology. (Source)
That was a fair critique five years ago. With the launch of AIP and the expansion of Foundry into operational apps and agentic workflows, the platform has moved well beyond data integration. The Cavanagh case study explicitly describes LLM-backed autonomous action on operational data as the next phase. That is not middleware. That is a category of product construction has never had. (Source)
Pick a side. Either build a deeply specific best-of-breed product that has clear value regardless of which platform ends up on top, or explicitly build to integrate with Palantir and position yourself as a layer on the ontology. The founders who will struggle are the ones trying to be a platform in the middle, because that middle is shrinking fast.
Two things. First, do not sign a seven-year Viewpoint or Sage ERP deal. Even if you end up needing one, the terms in 2026 should be shorter and more modular than anything vendors were offering in 2020. Second, if your project mix includes data centres, energy, or hyperscaler work, start the conversation with Palantir directly. The owners on those projects are often already customers, and there are co-deployment economics available to contractors in that orbit. (Source)
Bigger. The frame is useful as a headline, but the real story is that the construction software stack built between 2005 and 2020 is being replaced by a horizontal enterprise operating system architecture over the next five years. Procore is the most visible incumbent, which is why the comparison works. The actual market shift is broader than any single vendor.
Palantir Blog: Cavanagh and Palantir Are Building Construction’s OS for the 21st Century
Palantir for Construction offering page
Palantir Q4 2025 investor presentation and business update
Procore Q4 and full-year 2025 financial results
Thomas Cavanagh Construction at Palantir Paragon 2025
Viewpoint implementation cost benchmarks, 2026
Related Articles
Procore Slows. Trimble Soars. Plus the $1.5M OpenAI Employee Bonus
Procore’s AI Bombshell & the Death of Startups?
ICYMI: Procore’s Acquisition Playbook, Goldbeck’s 570 Software Tools, and Oracle’s Construction Bet
