Palantir is one of the newer sweethearts of
WSB, Reddit, and retail investors worldwide and with good reason - a polarizing, highly
secretive, niche business whose majority of market share comes from selling software to the
governments, defense, and intelligence sector. Palantir was also founded by the contrarian
Peter Thiel, and features a mysterious, borderline cultish, and equally contrarian CEO,
whose habits, work routines, philosophies, and willingness to speak up makes him a radical
compared to others in Silicon Valley.
But beyond these memes, let’s take a look
at how Palantir operates as a business. Along the way, we can also learn just how much
money there is to be made selling to the CIA and the greater American military complex.
Palantir sells two software products. Their first product is called Gotham, which is a tool built
for analysts at defense and intelligence agencies. Gotham helps these analysts find patterns in large
datasets. One of the inspirations for Gotham was that in the initial invasion of Afghanistan
and Iraq in the early 2000’s, US soldiers had to put map the intelligence they got of the
networks of insurgents and roadside bombers by hand on paper. Gotham is designed to address
that, which allows analysts to put all their data together, from confidential informant reports
to intelligence signals, and then find patterns and visualize these large datasets.
These same challenges of making sense of large datasets and finding patterns exist not
just in the defense sector, but also in the private sector. Foundry is a version of Gotham
that is designed for private companies that are not military, government or intelligence.
As a military and intelligence-centric business, Palantir holds a very different philosophical
stance from regular tech companies. Palantir makes it clear that they have no business in
collecting, mining, or selling data - all the data that customers bring and use within Palantir’s
products, whether it is Foundry or Gotham, all belong to the customer. Palantir is
about helping these companies centralize these large datasets from existing tools and
analyze the entire bucket together.
Palantir also has a unique design approach with
their software. Most tech companies today have a “rip-and-replace” walled garden mindset. Their
goal is to get you to tear out what you already have today and use their software exclusively
and make it hard for you to switch back to your old software if things get wonky with your new
setup. Palantir’s is different. Their software is designed to integrate with existing legacy IT
installations. That way, whether as a defense, intelligence, military, or private organization,
you can maintain your historic investments. You don’t need to rebuild your entire data
infrastructure and redo everything from scratch. This is a very different take compared to say,
Snowflake as we covered in our earlier episode, where that company’s story is that “all your
historic data infrastructure is no good. You need to get on ours.”
Palantir is an extremely unorthodox company both in its structure and business from
the numbers. Given its extreme focus on defense and intelligence, Palantir has a very small
market to target as an enterprise software company but with high upside on paper. Palantir
has only 125 customers to date but each customer on average spends a massive $5.6M dollars.
We can get a sense just from these numbers how lucrative and rich the American military
industrial complex is. Since Palantir by nature is a politically charged business and its
products are so sensitive, the company by design only sells to the US and its allies.
Regardless of your own personal views, there are only a handful of intelligence and
defense agencies in the West that fit under this view. In the US alone, there are only 16
intelligence agencies with the most well-known and well-funded ones in today’s day and
age being the CIA, FBI, and NSA.
If we compare Palantir to the mainstream
enterprise software companies, even the same ones that IPO’d in the same year as Palantir, you can
see that the private sector is a totally different animal. Snowflake, for instance, has over 3,000
customers with an average spend of 160K. DataDog has more than 8,000 customers at 37K. Palantir’s
average customer as an intelligence agency spends thirty five times more than the average private
enterprise. While as Americans we don’t get the exact visibility into the funding for America’s
intelligence agencies, the average 5.6M spend on Palantir should give you a sense for just how
much taxpayer’s money is allocated to the overall military-industrial complex. Palantir’s customer
base is split 50-50 across both the public and private sector, but all reports indicate that the
public sector outspends the private sector.
To me, where the numbers get really interesting is
when we remove the political aspects and evaluate Palantir as a business. Palantir boasts an
impressive top-line performance going from 500M in 2018 to over 1B in gross revenue in 2
years. Like most tech IPOs of the 2020’s so far, Palantir is not profitable but as we’ve seen time
and time again, tech companies are happy to incur present-day losses to compound future profits.
But something about the way that Palantir is allocating its costs today makes the profitability
seem very very far away for this company.
Palantir is making a calculated bet that the
future of their business is not about expanding to the private sector, even if they claim that is one
of their strategic focuses. The future of their business is expanding into and fully monopolizing
the Western intelligence and defense community. The American military industrial complex is
so lucrative and selective that these agencies will always side with Palantir and have more than
enough budget to keep Palantir fat and happy for generations to come. It’s worth noting here that
Palantir’s roots were in the CIA as the company took on funding and heavy pre-product influence
from the CIA during the early 2000’s.
When you look at Palantir’s spend, there are some
noteworthy ways on how the company has decided to allocate its investments. For a company that
has spent over 18 years building and refining its products, the same software products it has
battle-tested over the Afghan and Iraq wars over the past decade, Palantir spends more on R&D than
any Silicon Valley company in the world. Palantir spends 51% of its revenue, or 51 cents of every
dollar it makes, on engineers and designers to build and improve its products. For comparison,
Apple spends 7% of its revenue on R&D. Facebook spends 21% of its revenue on R&D. Microsoft spends
13%. Google spends 15%. Now some might say it’s unfair to compare Palantir, a tech
company still in pre-profitability growth stage with mature profit makers.
If we constrain the comparison to other recently IPO’d tech companies that are also in the
pre-profit stage, Palantir’s R&D is still a record high. Snowflake spends only 40% of its revenue
R&D, Confluent spends 45% on R&D, and DataDog spends 35% of its revenue on R&D. Palantir has
spent over consistently 40-50% of its revenue every year of its existence, not just in 2020 -
now there are a few possible reasons for this. The first is likely due to Palantir’s stance
about integrating its products with existing software of intelligence and defense agencies. The
US government has legacy software and hardware all throughout its systems - for a country that still
uses floppy disks to manage nuclear missiles, one can only imagine how outdated, archaic,
and crappy its other IT systems are. It’s an incredibly powerful sell for Palantir to
walk into a government or defense agency and say, “hey, we’ll work with everything you
have today. Nothing is getting ripped out. Everything you are using today, you keep using.
Our products will enhance everything you’ve done so far.” That alone is a strong sales pitch
that no other tech company would ever make.
That’s because the technical effort and
engineering required for Palantir to integrate and customize its products to each government’s IT
systems, policies is absolutely massive. Software, hardware that were built in the 70’s and 80’s
were simply not built for today’s Internet or tech standards - and in many cases, were never
designed to evolve. Integrating products into those environments and architecture is a massive
engineering effort that makes Palantir operate more than just a software vendor. Palantir, by
effect, become consultants as they handle the technical implementation and integration details
to demonstrate the value of their offerings.
This is a real double edged sword.
Even after 10 years of operating, the reality is that Palantir has to maintain an
expensive R&D investment and high engineering counts to make sure their products can integrate
into not just desktop computers in an office, but everything from oil rigs to military
Humvees and to battleship mainframes. But at the same time, no other company has pockets or
connections as deep as Palantir to undertake these efforts politically and technically,
giving Palantir an unparalleled customer retention and service in the public sector.
Given the high amount of customization and technical integration needed, Palantir’s sales
cycles can take on average an incredibly long 6-9 months to close. But when the deals do close,
the company pockets $6M. The only question is how much room is there for these deals to grow? Could
the average deal size be expected to grow over time or hold consistent? With the United States'
formal exit from the Afghanistan war in 2021, would Western governments invest more in
their defense and intelligence budgets? These are future-looking market questions
that are deeply tied to the political landscape rather than the economic one.
It’s difficult to foresee Palantir’s products making inroads in the commercial market with
red-hot fellow Silicon Valley competitors like Snowflake and Databricks, who have spent
billions of dollars themselves building a product and an effective muscle selling to the
private enterprise. However, in the past 2 years with COVID, while there was a global
economic contraction in the private sectors, there was an increased spending in the public
sectors that Palantir capitalized on. Palantir took the spotlight with big deals in the military
and public sector, signing a $823M contract with the US Army, a separate $90M contract with the
Veterans Affair over 4 years, and another $60M with the National Institute of Health in the past
6 months, and another unannounced with Department of Health. While critics have pointed out that
governments have been conveniently using COVID as a cover to expand greater visibility and
monitoring of citizens, we can all agree that under the pandemic, the governments around the
world have all taken active roles in setting and enforcing behaviors in society.
Palantir is riding on the big data trend just like Snowflake, but what’s unique to Palantir
is that the company rides on bigger winds based on macro geopolitical trends. Despite intense public
scrutiny, Western governments have continued to actively promote and invest in tech-driven
governance with items such as preventative policing, public surveillance, and facial
recognition. All these policies will require a key strategic technical partner such as
Palantir, who already has inroads at the federal and local level with the NYPD, LAPD, and the DHS.
The market opportunity is there for Palantir to create and capture that value should the politics
ever drive these ideas to reality. Palantir is an unconventional company and even more unorthodox
investment for stockholders - when you invest in Palantir, you are betting on not a tech company,
but rather the permeance of the American military industrial complex and a greater hands-on
role of governments throughout the West.
