It was a running joke at Stanford that nobody really understood what Palantir, Peter Thiel’s data analytics company, actually did. It’s easy to understand what consumer product companies like Apple do: Apple makes iPhones and MacBooks. Even enterprise companies are often pretty straightforward: Box develops file-sharing tools for businesses. But Palantir? They make Gotham, an “operating system for global decision making.” They also make Apollo, an “operating system for continuous delivery.” They now offer AIP, which is described as helping “get AI into operations.”
Why does Palantir do so many disparate things, and why are all of those things so hard to explain? Well, it has something to do with their customer base: the government.
Selling to the government is hard.
For many years, tech startups didn’t sell products to the government. Compared to everyday consumers and even business customers, the process for selling technology to the US federal government is really, really hard. Instead of just selling a product to a buyer who wants to pay for it, and negotiating a price based on how much they need the product, companies selling to the federal government have to navigate a complex system of open contracts and cost-plus pricing to make a single sale. And until recently, that contracting system struggled to even support software purchases; I’ve heard stories about companies selling the DoD laptops that have their software pre-installed, rather than just selling software licenses, because it made contracting easier.
Palantir was the first company to really crack the process of selling tech products to the government. And there wasn’t some magic trick to make it work; they just had great technology and a hard-working sales team that expertly navigated their way through the federal government technology acquisition process. And once they accomplished this, they started making money because of one simple reason. The old defense contractors were able to sell bad products because the complex contracting system stifled competition.
Old-school defense contractors got complacent.
For decades, the so-called “defense primes” -- companies like Lockheed-Martin, Boeing, and Raytheon -- were some of the only companies selling technology to the US federal government, and especially to the DoD. Startups who tried to compete often got bogged down in the difficult contracting process and just gave up.
At the same time, most government contracts were being paid out on a “cost-plus” basis: the government reimburses the contractor for all of the contractor’s costs developing the product, plus some additional fixed fee. That meant that there was very little incentive for large defense contractors to make cheaper, better products! There were no startups competing with the defense contractors, and the defense contractors were not rewarded for operational efficiency, so most technology developed for the federal government was pretty bad.
This meant that when Palantir managed to actually start selling to the federal government, their customers loved it. For the first time, modern tech startup products were being delivered to the DoD and the three-letter agencies. Anduril later followed suit, building hardware where Palantir was only delivering software. From the outside, Anduril is just as much of an enigma as Palantir: they sell drones, but also submarines, sensor arrays, surveillance towers, and sensor fusion software.
Both of these companies have such disparate product portfolios for a simple reason: even though they’ve broken through the federal contracting process, selling to the federal government is still really, really hard.
Bundled technologies, one sales process.
Building up a sales team that’s capable of selling to the US federal government is a unique challenge that both Palantir and Anduril solved. Once you’ve sunk the time and effort into building that sales team, you want to leverage it as much as possible, by selling as many things to the government as possible. That’s why these companies have such broad product portfolios: they can leverage one strong defense sales team to sell many different products.
Ultimately, this unique situation gives these companies a number of advantages. First, because so much of their unique advantage comes from their sales relationship with the US government, their technology doesn’t have to do all of the heavy lifting. Ultimately, Palantir sells various data analytics solutions. Anduril sells drones and similar products. They’re good products, but Palantir’s data analytics products aren’t necessarily super differentiated from similar offerings across the industry from a technical perspective. The differentiation comes from the sales relationship with the feds.
This is a double edged sword for venture investors in the defense tech space. On one hand, a defense tech startup can be fairly successful taking existing tech and combining it with a sales process tailored to the US federal government. On the other hand, that sales process is really difficult, and a lack of technical differentiation can make it hard to sell to conventional enterprise customers.
I will note that not all defense tech startups fall into this category. One notable exception is Epirus, who have a unique high-power microwave technology. They sell to the DoD because the DoD is the best buyer: Epirus’s technology is incredibly good at knocking drones out of the sky.
Ultimately, though, many defense tech startups take existing technologies and bundle them with sales arms capable of selling to the DoD. It’s why a lot of defense tech companies end up with complex, bundled services rather than a single clear offering. And that’s why nobody understands what Palantir does.
Superb post. Some questions:
1. With respect to:
"the government reimburses the contractor for all of the contractor’s costs developing the product, plus some additional fixed fee. That meant that there was very little incentive for large defense contractors to make cheaper, better products!"
Isn't the fixed fee a percentage and thus the actually incentive is to inflate the cost to make the product/service.
2. I'm note entirely convinced the bundling and strong sales relations is the key. Isn't a lot of the success of Palantir and Anduril a feature that the founders are billionaires. In other words, both could operate at a loss for a long time while developing expensive technology they believed the US defense and national security agencies needed. This means that when Pentagon officials saw what they were doing and wanted to buy/deploy, Palantir and Anduril had the upper hand and could sell at a beneficial fixed cost (i.e., if they can lower their costs then the margin is all profit).
Brilliant post. Digging deeper into Palantir's moat on sales relationships:
1. What exactly goes on in the sales process? From what I understand, government sends out open tender invitation for a certain product (hardware or software), bids are evaluated on technical parameters and cost, and the supplier is selected accordingly. At least this is how it happens in India, where I'm from. What might be Palantir's differentiation on the sales process?
2. Given that Palantir is about undifferentiated technology sold to DoD in an efficient sales mechanism, how do they hire and train their sales staff? Are there certain "feeder" industries or organisations from where they hire?
Thank you for these wonderful insights! :)
P.S. Peter Thiel did a Reddit AMA 10 years ago. Someone asked him "Is Palantir a front for the CIA?" His reply: "No, the CIA is a front for Palantir."
https://www.reddit.com/r/IAmA/comments/2g4g95/peter_thiel_technology_entrepreneur_and_investor/?sort=confidence