This is a story from 2017. The NYPD had been with Palantir for 5 years, and wanted out. It had developed its own system (Cobalt) to work with an IBM product. The NYPD complained that Palantir was not cooperating with the extraction of analysis and other data that the NYPD had generated while using Palantir.

Palantir had used the NYPD contract as a marketing tool, and NYPD was deep into the supposedly profitable Scale portion of the Acquire/Expand/Scale model.

Big takeaways

1. 1. $3.5 million annually for a mature Scaled client of significant size seems far too low. This appears how PLTR lost $4B - all those government contracts were low dollar and inadequate to offset the heavy lifting costs of Palantir’s Implementation engineers on long-term deployments.

2. 2. This dispute, the part that saw the light of day, suggests the “moat” might just be Palantir’s obstruction of data transfer at relationship’s end. Remember Acqure and Expand are not where the moat is, even in bull cases. The Scaled client should be so enmeshed with Palantir that the switching costs are very high. But maybe not, if an internally developed interface - like the NYPD’s Cobalt- is ready and a new partner like IBM is standing by to help with the data transfer.

3. 3. The NYPD developed its own replacement for what Palantir does. No aspersions to the men in blue, but if the NYPD can do it on a government budget….

I highly recommend reading the story below, and consider joining the CU Palantir chat discussion.