Palantir (PLTR) results show enterprises are spending to put AI into their operations, and the commercial business earns credit for that. The stock reaction matters because it shows how much investors already expect from the company and from AI overall. Slower growth in new contract value gave investors a reason to look harder at what comes next. The reaction reads as investors questioning whether exceptional growth can last, rather than rejecting AI itself.
Successful deployments carry real weight, but the harder test is sustained expansion across many customers. The enterprise test is whether customers can show better results after counting the full cost of deployment. That is what supports continued spending once the first excitement fades.
The lock-in question
Alex Karp said that once you move to Palantir, you don't go back. That claim holds up now, though it would have seemed doubtful before. Most companies building with AI want openness and the freedom to switch, so they don't feel tied to one vendor. Over the past year or two, that view has shifted as sovereignty, cost efficiency for AI, and security have become bigger concerns. Locking into a trusted partner, or a couple of partners, now looks worth it. Palantir's partnership with Nvidia (NVDA) is one example.
If an enterprise wants to build AI it can trust to write back to its system of record, it needs to invest heavily in a company that knows how to build the engineering layer underneath. That layer lets the company work with its data in a deterministic way instead of a purely probabilistic one.
Forward deployed engineers and the scaling problem
Palantir's offering mixes software with engineers, known in the industry as forward deployed engineers (FDEs), who work alongside customers to make the product useful in their businesses. That explains the traction. It also raises a future-growth question: how much of that work becomes repeatable if you need teams of engineers to keep it running? And how easy is it to switch providers when you depend on Palantir's engineering team?
Adoption is not just running software. It comes with a team of people needed to run and support it to get results.
Standards and ontology
The entire data industry has spent years fighting over standards, and this matters for where Palantir fits. To succeed with current technologies, companies often have to unlearn what they have known for the last four or five decades. Part of that is building a semantic layer that understands the meaning of data, how it relates to other data, and the rules used to govern what you do with it. All of that has historically been tied up in ontologies, a term tied to decades of standards work.
Many enterprises still struggle to put that into practice. This is why the FDEs matter: they come in, look at your business and your data, understand how to build the ontology, and then maintain it. Maintaining it is the trick. Palantir has the technology and knows how to do this correctly, but scaling it is a challenge, because the FDE model is not one-and-done. It is a long-term relationship.
