Software Engineer and Finance Blogger
"An investment in knowledge pays the best interest." - Benjamin Franklin

Palantir (PLTR): Must-Own Growth Stock?Palantir: Must-Own Growth Stock?

By FinanceTLDR | Dec 21, 2021

Chord progression is a defining feature of Western music on which melody and harmony are built. A basic chord progression starts with a tonic chord that acts as a pillar of stability, followed by a predominant chord that sets harmonic direction, preparing the listener for a dominant chord. The dominant chord creates tension and instability, and draws the listener towards a tonic chord for resolution.

The dominant chord is an appropriate but imperfect analogy for Palantir's current state. There's an air of positive instability about the company. It's at an inflection point, where it has recently gone public and has just found significant traction in both its commercial and government businesses. The company seems to be on the brink of transforming from a burgeoning data analysis SaaS (Software as a Service) company to an industry powerhouse.

Before we explore Palantir's potential, we first need to understand what it does. Its product offerings are complex and highly technical, and I've noticed that many analysis articles gloss over how they work. In doing so, they greatly miss Palantir's story and either exaggerate or underestimate its potential.

A 30,000 Foot View

In the Lord of the Rings, Palantírs are indestructible crystal balls that were used to see anywhere in the world, at any point in time. It's a fitting name for a big data and business intelligence SaaS company. This type of company ingests massive amounts of data produced by modern corporations to generate actionable insights for managers. The big data and business intelligence (BI) market is huge. In 2020, the market was estimated to be valued at $198 billion. This is projected to reach $684 billion by 2030, according to research firm Valuates (source).

The big data/BI market's lucrativeness naturally results in intense competition. Fortunately, Palantir has carved a niche for itself in two ways.

First, it has a government business with almost zero competition. Tech companies with the technical prowess to compete generally have an ideological aversion to serving military and intelligence agencies. Palantir's commercial business is also unique in that it aims to deeply integrate with a company's data to generate high quality insights, while competitors prefer to build general and open-ended tools.

The company is growing at a rapid pace. Its Q1-Q3 revenue this year is 44% higher than the same period last year, reaching over $1.1 billion, and commercial customer count grew by 135%.

The stock is reasonably priced with a price to sales (PS) ratio of about 25. Although this is a high PS ratio for growth companies, it's on the low-end among competing data analysis SaaS companies. For example, $SNOW has a PS ratio of about 95 and $DDOG is at around 60.

Commercial Business

As mentioned above, Palantir's main differentiator in the commercial big data/BI business is data analysis software that deeply integrates with a company's data. Deep integration means transforming said data into a format that Palantir understands, enabling complex analysis. This is in contrast to generalized software that accomodates disparate schemas but can only perform shallow analysis (e.g. Snowflake for scalable data storage/processing and Tableau for data visualization).

Deep integration generates higher quality insights at the cost of higher onboarding friction.

Given the size and rapid growth of the overall big data and BI market, Palantir's commercial business is expected to drive the majority of its growth. Even though its government business is lucrative and lacks competition, governments move slowly and contracts come by slower than Wall Street would like.

Palantir's flagship commercial product is called Foundry. It's a feature-rich end-to-end data analysis system with a vision of being "The Operating System for the Modern Enterprise".

Foundry transforms disparate and messy company data into data that conforms to a custom ontology (i.e. data format). Once transformed, Foundry deeply understands the data and is able to generate unparalleled insights via visualizations, data breakdowns, and even forward-looking simulations.

A screenshot of Foundry being used by the UK's NHS
A screenshot of Foundry being used by the UK's NHS

Data Transformation

Data transformation is the toughest part of Foundry's onboarding process. A company often has many disparate data sources with inconsistent schemas that require significant work to gather, sanitize, and transform to conform with Foundry's ontology.

For example, consider a battery manufacturing company that wants to use Foundry to optimize its manufacturing output. The company probably stores its data in several disconnected databases that use wildly different schemas with different database engines. The sensor data from the factory floors is stored in one database, the maintenance data is stored in another database, the operational costs data is stored in yet another database, and on and on.

In order to find opportunities to increase efficiency, Foundry must holistically understand the manufacturing process. To do so, it needs to gather and transform all this data to conform with its ontology, correcting any inconsistencies along the way. This is not an easy feat, especially when it needs to be done for every new customer. Each company is different and follows different data practices. This results in countless integration edge cases.

This is why Palantir has a special class of software engineers called Forward Deployed Engineers that work closely with prospective companies to integrate Foundry. In addition, Palantir has invested heavily in streamlining large parts of the integration process. A prominent example is this low-code data pipeline orchestration system seen below:

A screenshot of Foundry's low-code data pipelining system
A screenshot of Foundry's low-code data pipeline orchestration system

Data Management (Versioning, Governance, and Provenance)

Companies have stringent data management requirements that Palantir is well aware of. Foundry is filled to the brim with features to meet these requirements and the company isn't shy about advertising these capabilities in its demo days.

To start, Foundry has a powerful data versioning system that allows engineers to manage datasets like code. This enables efficient concurrent collaboration on data engineering workflows.

Not all data should be available to everyone in an organization. Companies want strict control over who can access what. Foundry's data governance tools are state-of-the-art and enable fine-grained permissioning that extends from the original dataset to any derivatives down the pipeline. It also offers regulatory compliance support for data protection laws such as GDPR.

Finally, companies want to know where a piece of derivative data comes from. This helps with double checking sources to make sure the original data isn't corrupted. The process of connecting derivative data with its original source is called "Data Provenance". Foundry has deep support for this, allowing engineers to track the upstream source of a piece of data, as well as how it affects downstream artifacts. Working in concert with the data provenance system is a powerful automated data health check system that continuously monitors pipelines for anomalies and inaccuracies.


Besides powerful data management tools, Foundry is one of the most secure data analysis products on the market. This comes from Palantir's roots in developing software for the government. In fact, Palantir is one of only five SaaS companies authorized for Mission Critical National Security Systems (IL5) by the U.S. Department of Defense (DoD). IL5 is the DoD's second highest security classification level.

Deep Integration Can Scale (In the Long Run)

A top concern with Palantir's commercial business is that it doesn't scale since it's expensive to onboard new companies to Foundry. Because Foundry works best with deep integration, each new company requires the attention of several Forward Deployed Engineers to work through edge cases.

However, I posit that deep integration does scale; it just takes longer to reach escape velocity. The more companies Palantir onboards to Foundry, the more edge cases it experiences, and the more it can automate the handling of these edge cases. For example, Foundry has a feature called "Archetypes" that provides workflow templates "that can be deployed from a configuration wizard in minutes and drive real-world outcomes for these organizations" (Palantir Blog). There are templates for optimizing supply chains, manufacturing, and many other common business problems.

The scalability of deep integration isn't just a hypothesis, it's already happening.

An example was given in a demo day earlier this year about a supply chain company with a complex data system ("dozens of data sources including 27 separate ERPs") that wanted to leverage Foundry for optimization.

Palantir claims that in just a few hours, the company had an integrated view of its supply chain. In two days, Foundry was proactively alerting on potential bottlenecks. In two weeks, Foundry had helped the company increase working capital by $50 million (source).

In addition, scalability is not just seen in one-off examples, it's showing up in business metrics. In Q3 2021, Palantir grew its quarterly commercial revenue by 37% year-over-year, with US quarterly commercial revenue growing by a whopping 103% year-over-year. Its commercial customer count grew by 46% from Q2 2021. Along with this is a significant reduction in costs leading to a $605 million improvement in adjusted free cash flow from Q1-Q3 this year as compared to the same period last year.

Government Business

Palantir is one of the most competent high tech firms to provide data analysis and software deployment services to the military and intelligence agencies of Western governments. Fortunately for Palantir, there is a lack of competition in this area because US tech firms have an ideological aversion to serving such institutions (which is a nice moat).

For example, in 2018, 3,000 Google employees signed a petition protesting the company sharing its AI technology with the US military as part of a Pentagon defense program called Project Maven. Facing significant internal pressure, Google backed out (source). Interestingly enough, it was Palantir that happily took over the contract (source).

Gotham and Apollo

A screenshot of Gotham being used for intelligence operaetions
A screenshot of Gotham being used for intelligence operaetions

Palantir has two primary products tailored for government institutions called Gotham and Apollo.

Gotham is a data analysis platform designed for battlefield and intelligence operations (Foundry is Gotham but for commercial customers). It's capable of ingesting vast amounts of information from disparate military data sources such as Humvees on the battlefield to satellites in space. Leveraging AI, it provides real-time insights to military commanders to significantly improve situational awareness and decision-making. Gotham is also used by intelligence agencies for information sharing and analysis. For example, US intelligence agencies were already using Gotham in 2015 to connect databases across departments. Prior to Gotham, these databases were siloed (source). Fast forward to today, Gotham is no doubt significantly more integrated with the data workflows of intelligence agencies. Palantir's Wikipedia page is a good resource on the myriad of government agencies that have used or are using Gotham.

The other product offering tailored for governments, Apollo, is a highly secure and flexible software deployment system perfect for military and intelligence agencies. Apollo's website advertises that it "enables continuous deployment across all environments", "[f]rom the back of a humvee to the hull of a submarine". Apollo is even capable of deploying software to physically disconnected environments which is most often run by intelligence services that want maximum security (source).

Put together, it's easy to see Gotham and Apollo propelling Palantir to become the top software provider for Western governments. There are few other tech companies competent enough, and willing, to supply governments with the type of software that Palantir develops. Gotham is a single point of truth for all military and intelligence data, augmenting operator decision-making with powerful data analysis systems and AI. Apollo helps government agencies securely keep their software up-to-date in a wide variety of ordinary or specialized environments.

The Best Customer

There are many upsides to having governments as customers. For one, they are slow to change and once they're integrated with Palantir's software, it's incredibly hard to switch away. This is a powerful moat that creates perpetual and highly dependable customers. In addition, governments (especially the US government) can quite literally print money. Who wouldn't want a money printer as a customer?

At the end of the day, Palantir's government business serves as a dependable source of income with a relatively low growth rate (still impressive at 34% year-over-year growth in Q3 2021). Even though the reliance of Western governments on Palantir will only increase over time, governments move slowly and contracts take time to finalize. Stable government revenue will act as a backstop against the more volatile commercial revenue.

A Good Stock to Own?

Palantir is an anomalous company. There are few data analysis SaaS companies building deeply integrated data analysis software and fewer serving government institutions as faithfully as Palantir does. This "Think Different" mentality has resulted in a diverse product line with strong moats.

Nevertheless, Palantir's biggest challenge for growth is scaling its commercial business. It's costly to onboard companies to Foundry, which requires deep integration. However, if 2021 business metrics are any indication, Palantir seems to be on the verge of surmounting this challenge, if it hasn't already.

At a 25 PS ratio, Palantir is fairly priced or even cheap when compared to the PS ratios of similar companies (i.e. 95 for $SNOW and 60 for $DDOG). Its uniqueness and high growth potential makes it a very attractive data analysis SaaS stock to own, especially in a diversified growth portfolio.

This is not financial advise, but simply my own opinion on $PLTR based on the analysis I shared above.