[E7] Palantir: Peter Thiel's vision for software
Start with a big problem in a niche market and dominate it
Dear Reader,
This profile in Product | Strategy | Innovation provides insights into a co-founder and investor whose company is using artificial intelligence / machine learning (AI/MI) to protect governments from terrorist threats and identify deep insights for multi-national corporations. Unlike Google and Facebook who first pursued vertical integration for AI/ML within their own companies, Palantir hires top talent and provides AI/ML platforms to build software analytics and business intelligence into other company’s computer systems. Palantir’s first customer was the US Central Intelligence Agency (CIA) and their first external investor was the venture capital arm of the CIA know as In-Q-Tel.
We will explore the following in this profile:
Why would Palantir initially focus on government agencies and intelligence?
What is Palantir’s value proposition for corporate clients?
Can Palantir maintain a competitive advantage as evolving technology and talent allow companies to build out solutions on their own?
Background
Peter Thiel founded the startup Confinity that later merged with Elon Musk’s X.com to form PayPal in 1998. Thiel was the PayPal CEO when it was sold to EBay for $1.5 billion in 2002. And Thiel was also the 1st outside investor in Facebook and still serves on its Board of Directors.
After EBay acquired PayPal, Peter Thiel started Founders Fund with two other PayPal co-founders to raise and invest venture capital. Founders Fund investments included Spotify, AirBnB, Palantir, Stripe and Lyft. Through Thiel’s experience co-founding and leading PayPal and then investing in many leading startup ventures, he started teaching about best practice for early stage startups for a computer science class at Stanford University. The popularity of this class led to a best-selling book Zero to One by Thiel.
This background is important because Thiel co-founded Palantir with a Stanford Law School classmate, Alex Karp. Ex-PayPal engineer and Palantir co-founder Nathan Gettings built the initial prototype platform. Stanford students and Palantir co-founders Joe Lonsdale and Stephen Cohen eventually helped Gettings build out more of the platform. Thiel served as Palantir’s Board Chair until just recently with Alex Karp still serving as the CEO. Thiel’s principles likely played a significant role in the formative years at Palantir. What does that tell us?
Fraud detection and cyber-security were key requirements at PayPal when Thiel served as its CEO. The 9/11 terrorist attack also happened during this time in 2001. These events set the early product direction for Palantir. Startup principles advocated by Thiel include creating a monopoly effect with core technology that is at least 10x better than any competition and using niche opportunities to build out the key product and services to dominate an initial market. PayPal focused on payment between EBay Sellers and their customers. Facebook focused on Harvard and then included some other leading colleges to acquire users for its initial social medial platform.
Government-contracted software solutions provided these initial criteria for Thiel and Palantir where a beachhead market was created at the CIA to tackle intelligence and national security from 2005 through 2008. Success at the CIA then led to additional contracts at the FBI, NSA and US Army. Palantir provides these customers a unique opportunity to deploy software and services to address their biggest challenges with top talent only a leading tech startup can attract and retain. National security also provided this top talent meaningful problems to solve vs. optimizing ads for corporate clients at Google or Facebook.
Corporate clients were eventually recruited by Palantir from insurance, healthcare, energy and manufacturing industries to scale up key talent through Palantir versus on their own. This may be Palantir’s core strategy to lead computer science and engineering top talent arbitrage against Google, Facebook and other tech companies and then sell the software and services the hired top talent builds and operates for governments and dated industries who can’t recruit that talent on their own.
But this also presents a challenge any business-to-business software company faces. The total addressable market for enterprise software is much more limited than consumer markets. Leading enterprise software companies Oracle, SAP and Salesforce are bound to market caps around $200 billion whereas Microsoft with only 36% of revenue coming from business productivity has a current 10x market cap with just under $2 trillion with the addition of consumer-facing revenue.
However, if Palantir expands from its beachhead government customers into key corporate industries using business intelligence to position the company as a core strategic partner, the services provided could be priced according to the value created. Maybe there is an opportunity to redefine enterprise software with a state-of-the-art artificial intelligence / machine learning platform to take on the biggest opportunities and threats.
Palantir
Company: Palantir Technologies, Inc.
Founded: May 2003 initially by Peter Thiel
Founders: Peter Thiel, Nathan Gettings, Joe Lonsdale, Stephen Cohen, Alex Karp
CEO: Alex Karp
NYSE: PLTR
2020 Revenue: $1.1B (+47% Year-over-Year)
Estimated number of customers: 238 companies, mostly in the US with at least 1,000 employees and over $1 billion in revenue.
Palantir’s mission is “To help our users, the people doing the hard work on complex, real-world problems. We do this by writing software that enables effective analysis against complicated, data-driven problems.”
Palantir states that it is an engineering culture. Everyone regardless of background are focused on solving the hardest problems the company can find. Palantir provides three guiding ideas at the company. 1.) The best idea wins. 2.) Nothing is permanent. 3.) Keep focused on the mission.
Product
: Gotham: Integrate, manage, secure, and analyze all of your enterprise data. Gotham was the original unstructured data product with a government intelligence agency focus to provide advanced software tools to project leaders, analysts and operations. Foundry: Open source, interoperable platform for end-to-end data transformation mostly for multi-national corporations. Apollo: Continuous delivery software that powers SaaS platforms, Foundry and Gotham, in the public cloud and beyond.
Strategy:
Palantir sought out a niche problem space related to fraud detection and security threats based on co-founders’ experience building out fraud detection at PayPal. The objective was to build a beachhead market using this initial niche opportunity, dominate it and then expand. Intercepting and preventing terrorists threats was selected as the problem to solve. The CIA was recruited as the first customer. Palantir focused just on the CIA for 3 years before they expanded to other national security agencies including the FBI, NSA and then eventually the Armed Forces.
Terrorist threat detection was also an ideal use case for artificial intelligence and machine learning. This not only leveraged past experience at PayPal, but was a strategic area to recruit and hire top engineering talent. The CIA recruits top engineering talent directly, but cannot compete with the Google and Facebook opportunity and pay. But Palantir could as a Silicon Valley startup gunning to lead artificial intelligence and machine learning.
So a key part of Palantir’s core strategy is to lead top computer science and engineering top talent arbitrage against Google, Facebook and other top tech companies using the biggest problems they can find in government agencies and non-tech companies in dated industries to build out AI/ML software solutions. But it appears Palantir doesn’t just build, customize and sell software and services. Palantir is using the top talent and long-term contracts to solve big problems that require vertical integration between Palantir’s platform(s) and a customer’s own computer systems. So the result with 238 customers would be 238 unique integrated software systems that each leverage 30-60% Palantir AI/ML platform technology embedded with 70-40% custom software development, system integration and system operations with long term contracts.
Innovation:
The strategy to leverage a common AI/ML platform across many customers to build custom vertically integrated software systems offers a unique opportunity if Palantir is successful with its top talent arbitrage against Google and Facebook. Palantir can solve a broader set of problems that offer different challenges for AI/ML systems to solve. That means the core AI/ML platforms at Palantir can accelerate advancements vs. other tech companies that are more narrowly focused. Palantir can also leverage other customers big data to advance training the machine learning models faster and more reliably than it could do on its own or with a limited customer size.
But a more speculative innovation Palantir could leverage is a system to recruit and retain the very best engineering talent as illustrated in Fig E7-2. Recruiting new talent would benefit from open-source technology to evaluate candidates and hire based on prior experience with these open-source tools. But after a period of new hires building out solutions on open-sourced platforms like Foundry, top talent among this bottom tier could be recruited for training on more proprietary tools and systems that are unique to Palantir.
For top projects, proprietary tools that may outpace the performance of open-source technology could be used by higher-tiered Palantir employees who are on aggressive employment retention programs. And then the very top tiers could be recruited to an even more select group that builds out the Palantir proprietary tools. Even more aggressive employment retention programs are used to keep these apex employees. The objective is to prevent the top talent and proprietary tools leaking out to Google and Facebook. This could eventually disrupt these disruptive innovation leaders.
1. Why would Palantir initially focus on government agencies and intelligence?
Palantir’s first customer was the CIA. This is partly due to the problem focus on intercepting and preventing terrorist threats. The CIA would have the richest data set. They would also have the funding to support protecting the national security of the United States with a solution at least 10x better than any competing options. And lastly, the CIA recruits engineering talent, but not so much when Google and Facebook would be competing for a candidate. Google can offer a top engineer a 7-figure salary. The CIA cannot, but Palantir could with a long-term contract for the right candidate.
So to summarize, Palantir chose government agencies and intelligence in particular because:
Primary stakeholders to prevent a big problem for national security
Big & reliable datasets from many sources to train & challenge AI/ML algorithms
Well-funded customer for the right solution
Long-term contracts are common for stable revenue source
Subject matter experts on the problem are abundant
Big problem that would be of interest to top engineering talent
But government agencies would have difficulty hiring top talent directly
Big problem that is not going away to sustain product development
Competition is more consulting firms who routinely contract with he government vs. tech startups
Good prospect to create a beachhead market and a monopoly for the right solution to block others out of this beachhead.
Government agencies also have multiple branches that do similar work. So starting with the CIA allowed the AI/ML platform and talent to be built out over time. And success with the CIA enabled expansion to the FBI, NSA and Armed Forces and eventually other problems within these agencies. But the key is these government agencies would have difficulty recruiting and retaining top talent with the ability to build out the right solution for such a big problem within the required time frame. Alternatives would likely be Accenture, Cognizant, Deloitte, McKinsey and other professional service firms.
Government agencies have also led to vertical integration of a complex solution that spans cloud and edge computing with many nodes deployed for intelligence and military operations. This led to the development of the first product called Palantir Gotham. The product development and solutions built with it started in 2005 and continue today. Gotham is the flagship product.
2. What is Palantir’s value proposition for corporate clients?
Palantir appears to target dated industries for similar reasons it targeted government agencies initially. Corporate enterprises in dated industries are more likely to buy back stock and pay dividends rather than build out their own complex AI/ML computing systems even for a strategic problem at the company. The key opportunity is these companies are not hiring top AI/ML talent directly. So if key strategic problems in these companies are well suited for AI/ML, Palantir has opportunities to leverage its platforms and talent to take on building out solutions for these companies in dated industries.
Palantir can access top talent to scale teams to take on new work. But the target prospect companies would also have big datasets to inform building out the computer systems. And Palantir can leverage its own software platforms and experience from prior customers to build out a team to vertically integrate the end-user solution with the customer’s own systems generating key input data.
Corporate enterprises also offer the business model opportunity for value-based pricing where saving money or improving operations can provide upside revenue opportunities for Palantir. Five customers generating $200 million of recurring revenue annually each build out $1 billion in combined recurring revenue. With a 10x multiple on revenue that is a $10 billion boost to the market cap.
But the real upside to expanding into corporate enterprises is the diversity of data and problems that would be made available to Palantir’s AI/ML platforms and talent. This diversity will not be found with the typical vertical integration within a single company. But Palantir can pull all the insights with use of the platform(s) across many customer sites to continue to iterate development of the core platform(s) beyond the end-user software at each location.
3. Can Palantir maintain a competitive advantage as evolving technology and talent allow companies to build out solutions on their own?
Open-source software and growing interest in AI/ML will build out more computer science and engineering talent to build end-user computer systems. The open-source software tools and training datasets will continue to improve, too. The combination of these 2 factors will enable more companies to take on building out AI/ML platforms and systems. That provides growing competition from the build it yourself model vs. buying from someone else.
But that is more true for corporate enterprises vs. government agencies. Palantir can likely create very sustainable businesses from government agencies. But companies beyond Google and Facebook will become more active in building out AL/ML systems. Palantir will need to maintain a 10x performance advantage over competing systems and what companies can build themselves. This is a challenge, but doable with aggressive roadmaps and an ongoing desire to hire, train and retain the very best talent to win.
Conclusions
Palantir is a unique software company that initially focused on 1 customer for over 3 years to build out the needed staff and tools to dominate a niche opportunity based on its co-founder Peter Thiel’s startup principles. The focus was intelligence and national security to intercept and prevent terrorist attacks. The first customer was the CIA. That list then expanded with the FBI, NSA and Armed Forces. Gotham was the original platform Palantir built out to serve the needs of the CIA. Gotham is still used today, but mostly for government contracts.
Foundry enables Palantir to use more open-source tools to build out end-user solutions. Advanced simulation engines allow customers to use Foundry to identify opportunities to improve business operations. Open-source software tools will continue to evolve. That provides both an opportunity to recruit talent that is already proficient with these tools to enter Palantir and contribute almost immediately to active projects. But companies can also recruit talent and use evolving open-source tools to build their own AI/ML computer systems.
Palantir will need to continue recruiting top talent and build out the top AI/ML tools to stay at least 1Ox ahead on performance to any competition. But the ability to leverage top talent and tools to build out complex computer systems with multiple customers provides an opportunity to build systems beyond the capabilities of more narrowly focused vertically integrated solutions within a single corporation.
Google and Facebook also have significant opportunities to hire the best talent and advance their AI/ML capabilities. These companies will likely also create the next big tech companies that leverage AI/ML to do incredible things. That will provide additional competition for Palantir, but it will have a twenty or more year head start in a few years to any emerging new venture.
But Palantir’s biggest challenge compared to Google, Facebook and other tech companies is the limiting factor of only providing business-to-business services. Palantir will need to transform business-to-business services into more upside using AI/ML to drive deeper insights and more of an impact on customer operations to share in the upside of those value-based outcomes. If Palantir can play a key role in the $3.5 trillion healthcare market or other large total addressable markets, those can generate more revenue than the services alone. And recurring software revenue should bring a nice multiple for the valuation and market cap.
For full-disclosure, I follow Palantir with a primary interest in their products, strategies and innovation, but as a retail investor, I’m also long PLTR and GOOG mentioned in this profile.
Best,
Stephen
If you are not a subscriber, please consider doing so to get updates on this topic and more.
If you liked this post, please share with your networks so others can benefit, too.
Nothing in this post is intended to serve as financial advice. Do your own research.