The way Palantir talks about the CIA really rubs me the wrong way. For years, they would leak to journalists that Palantir "found bin Laden" when, of course, it had nothing to do with finding him. Several CIA employees died trying to find Bin Laden, all for some schmucks in Silicon Valley to try and capitalize on their sacrifice.
If you want to give a Silicon Valley company kudos for Bin Laden, give it to Cisco, VMWare, and Equinix.
Palantir's whole "CIA" marketing schitck appeared to be a ploy to build a strong reputation to help hiring.
At the end of the day, they're just another Datalake company that makes money off professional services, except Databricks and Snowflake can actually execute.
What more do you expect from a project from Peter Thiel, which is named after the most evil guy's magic all seeing orb from LoTR, which is explicitly made for governments to target whatever they want to call "bad guys" by slurping up as much data as possible from people who shouldn't be collecting it in the first place?
Dude has a dictator complex. Of course he fully the embraces the "just fucking lie and make money" ethos
What's funny is that Thiel is a believer of nominative determinism: https://en.wikipedia.org/wiki/Nominative_determinism
Critical support to Palantir in their quest to steal CIA valor.
There are issues with innovation in the DoD and DHS, but a lot of this is offloaded to private sector vendors anyhow.
I notice how the article didn't mention any of the companies I personally know doing stuff in the space, nor actually sourced from members of the VC, Business, or Defense community.
The fact that the author took Palantir's marketing at face value is proof enough - the CIA let their contract with Palantir lapse a couple years ago (and I think they only even bought it because of their stake in In-Q-Tel), and they haven't had great success selling to the Fed.
I actually work in this space btw.
-----
The bigger stumbling block is procurement.
Software Procurement by Federal standards is relatively straightforward so a Series E+ startup can make it if they spend around $7-10M and 1-1.5 years on a dedicated roadmap for FedRamp and FIPS compliance.
Once you step out of software, procurement becomes paperwork hell. Throw in the paperwork hell from Grantmakers like the DoD and DoE, and you end up with a quasi-Soviet procurement system. Ironically, most of these compliance and regulatory checks were added for good intentions - primarily to minimize corruption and graft, yet it basically clogged up the entire system, and dissuades startups and innovators from working directly with the Defense community.
Some projects like DIUx and and In-Q-Tel are trying to change that, but it's too little too late, and our defense base is entirely dependent on firms like Microsoft, Cisco, Crowdstrike, Zscaler, etc acquiring promising startups to evangelize their innovations internally.
Fundamentally, this is why I dislike the New America/Khan/Chopra vision of anti-trust. It doesn't actually help innovation from a federal standpoint, as small companies and startups have no reason to work with the Fed given the amount of red tape that exists.
If the same effort was put to harmonizing and simplifying procurement across the Federal Government, you could directly make demands on competition.
This is what China does, and is a major reason their MIC was able to grow leaps and bounds in just 20 years.