[Source Photos: Palantir and Getty Images]

By popular demand, this week I’ll do a deep dive into Palantir. In an industry full of largely anonymous and bland companies, led by “professional managers”, Palantir stands out in many ways.

Largely hated by the media and professional investors, the company has embraced it’s role as the contrarian, going as far as even playing along with the retail investor community.

Today we will try to answer several questions:

What does Palantir actually sell?
How does their product tie into the future of Enterprise adoption of AI?
Is this a good opportunity for tech sales?
The key takeaway
For tech sales: Palantir has an exceptional product and deep domain expertise in driving LLM and Enterprise-grade ML adoption within customers with large scale projects. It’s also a company that deeply despises sales reps. Avoid at all costs until a structural change at leadership level.
For investors: I’ll not speculate on where the price of the stock will go because all bets are off when we are talking about a “darling” of the retail community. From a purely product and customer integration perspective, the company is one of the 3 strongest players in AI adoption today. The continued mismanagement of the GTM team raises significant questions around how sustainable growth is on a long enough time scale.

So is there an actual product here?

Some say there are three major unsolved mysteries in this world, namely: who really built the Great Pyramid of Giza, what is the meaning of life, and does Palantir even have a real product?

I highly recommend to go trough the insider view of Palantir by Nabeel S. Qureshi. This is how he experienced their approach to customers:

When I joined, Palantir was divided up into two types of engineers:
1. Engineers who work with customers, sometimes known as FDEs, forward deployed engineers.
2. Engineers who work on the core product team (product development - PD), and rarely go visit customers.
FDEs were typically expected to ‘go onsite’ to the customer’s offices and work from there 3-4 days per week, which meant a ton of travel. This is, and was, highly unusual for a Silicon Valley company.
There’s a lot to unpack about this model, but the key idea is that you gain intricate knowledge of business processes in difficult industries (manufacturing, healthcare, intel, aerospace, etc.) and then use that knowledge to design software that actually solves the problem. The PD engineers then ‘productize’ what the FDEs build, and – more generally – build software that provides leverage for the FDEs to do their work better and faster.
This is how much of the Foundry product took initial shape: FDEs went to customer sites, had to do a bunch of cruft work manually, and PD engineers built tools that automated the cruft work. Need to bring in data from SAP or AWS? Here’s Magritte (a data ingestion tool). Need to visualize data? Here’s Contour (a point and click visualization tool). Need to spin up a quick web app? Here’s Workshop (a Retool-like UI for making webapps). Eventually, you had a damn good set of tools clustered around the loose theme of ‘integrate data and make it useful somehow’.
At the time, it was seen as a radical step to give customers access to these tools — they weren’t in a state for that — but now this drives 50%+ of the company’s revenue, and it’s called Foundry. Viewed this way, Palantir pulled off a rare services company → product company pivot: in 2016, descriptions of it as a Silicon Valley services company were not totally off the mark, but in 2024 they are deeply off the mark, because the company successfully built an enterprise data platform using the lessons from those early years, and it shows in the gross margins - 80% gross margins in 2023. These are software margins. Compare to Accenture: 32%.
Tyler Cowen has a wonderful saying, ‘context is that which is scarce’, and you could say it’s the foundational insight of this model. Going onsite to your customers – the startup guru Steve Blank calls this “getting out of the building” – means you capture the tacit knowledge of how they work, not just the flattened ‘list of requirements’ model that enterprise software typically relies on. The company believed this to a hilarious degree: it was routine to get a call from someone and have to book a first-thing-next-morning flight to somewhere extremely random; “get on a plane first, ask questions later” was the cultural bias. This resulted in out of control travel spend for a long time — many of us ended up getting United 1K or similar — but it also meant an intense decade-long learning cycle which eventually paid off.

There are two main themes to unpack here, one is about strategy and the other one is about execution.

From a strategy perspective, the goal of the company was to gain access to customers with complex challenges trough it’s consultancy approach, then build custom applications that solved quantifiable problems. These became the foundation of their current platform.

While unusual for a Silicon Valley company, it’s also not a shocking idea in the context of the industries they targeted. I used to work for a niche tech company focused on planning and optimisation software that had a similar approach, except that it started with the “base” software and then developed custom implementations on top of this stack that became industry-specific products. The reality was that even when utilising those “products”, there was still a heavy element of writing new code for the individual implementation. Unlike Palantir, the company ended up being acquired because it struggled to scale above the $200M revenue and it’s heavy reliance on consultants was seen as a negative for growth.

The second theme is about execution. Unlike the company that I worked for, Palantir focused on hiring exceptional talent and then pushing them to the limits in order to solve big problems quickly at their customers.

My first real customer engagement was with Airbus, the airplane manufacturer based in France, and I moved out to Toulouse for a year and worked in the factory alongside the manufacturing people four days a week to help build the version of our software there.
My first month in Toulouse, I couldn’t fly out of the city because the air traffic controllers were on strike every weekend. Welcome to France. (I jest - France is great. Also, Airbus planes are magnificent. It’s a truly engineering-centric company. The CEO is always a trained aeronautical engineer, not some MBA. Unlike… anyway.)
The CEO told us his biggest problem was scaling up A350 manufacturing. So we ended up building software to directly tackle that problem. I sometimes describe it as “Asana, but for building planes”. You took disparate sources of data — work orders, missing parts, quality issues (“non-conformities”) — and put them in a nice interface, with the ability to check off work and see what other teams are doing, where the parts are, what the schedule is, and so on. Allow them the ability to search (including fuzzy/semantic search) previous quality issues and see how they were addressed. These are all sort of basic software things, but you’ve seen how crappy enterprise software can be - just deploying these ‘best practice’ UIs to the real world is insanely powerful. This ended up helping to drive the A350 manufacturing surge and successfully 4x’ing the pace of manufacturing while keeping Airbus’s high standards of quality.

Funnily enough, my company was probably involved around the edges of these projects at Airbus but I don’t recall any discussions around the integration with the Palantir team or their influence on the day-to-day operations. From a purely execution perspective, this is an example of how much more value the Palantir team was able to generate by embedding themselves as part of their manufacturing team, rather than pitch solutions from the outside and then try to customise them during the demos.

FDEs tend to write code that gets the job done fast, which usually means – politely – technical debt and hacky workarounds. PD engineers write software that scales cleanly, works for multiple use cases, and doesn’t break. One of the key ‘secrets’ of the company is that generating deep, sustaining enterprise value requires both. BD engineers tend to have high pain tolerance, the social and political skills needed to embed yourself deep in a foreign company and gain customer trust, and high velocity – you need to build something that delivers a kernel of value fast so that customers realize you’re the real deal. It helped that customers had hilariously low expectations of most software contractors, who were typically implementors of SAP or other software like that, and worked on years-long ‘waterfall’ style timescales. So when a ragtag team of 20-something kids showed up to the customer site and built real software that people could use within a week or two, people noticed.
This two-pronged model made for a powerful engine. Customer teams were often small (4-5 people) and operated fast and autonomously; there were many of them, all learning fast, and the core product team’s job was to take those learnings and build the main platform.
When we were allowed to work within an organization, this tended to work very well. The obstacles were mostly political. Every time you see the government give another $110 million contract to Deloitte for building a website that doesn’t work, or a healthcare.gov style debacle, or SFUSD spending $40 million to implement a payroll system that - again - doesn’t work, you are seeing politics beat substance.

This also highlights how Palantir created something completely new by essentially approaching their customers both as a consultancy company that actually got stuff done AND a software vendor that was not shy of creating bespoke implementations. Fundamentally this is a strong conviction bet that by delivering outstanding value to your customer, you are no longer just a vendor that can be replaced but a key part of their vision as a company.

Now if we pull forward to today, the company has retained it’s approach of deeply embedding engineers on the ground with the customer with a strong focus on solving their data problems, but also has a high-margin software suite that has been built over the last decade from the repeatable use cases they’ve seen in their core markets. For this overview, the two relevant products are Foundry and AIP.

Palantir Foundry, introduced in 2016 as the commercial evolution of the company's government-focused Gotham platform, is an enterprise data operations system that uses an ontology-based architecture to create structured representations of organizational data. The platform builds upon Palantir's experience in government intelligence work, applying similar data integration principles to commercial applications while adding features like version control, low-code development tools, and integrations with common business intelligence software like Tableau. Since its launch, Foundry has become Palantir's primary commercial product, used by organizations ranging from manufacturing companies to financial institutions for managing complex data operations and analytics workflows.

Palantir's Artificial Intelligence Platform (AIP) is an enterprise software system that integrates large language models with organizational data systems through the company's ontology-based architecture, which has been a cornerstone of their data integration approach since their founding in 2003. The platform builds upon Palantir's experience with government and commercial clients, offering containerized deployment of AI models within private networks while connecting to existing data warehouses like Snowflake and Google BigQuery. While relatively new compared to Palantir's established Foundry and Gotham platforms, AIP has gained traction in 2023-2024, with the company offering a free tier through build.palantir.com that allows small teams to test AI workflows with limited ontology objects.

Palantir and the right time, right place

There aren’t many companies that were better prepared for the introduction of LLMs to the technology stack. Leveraging their deep expertise of data analytics and ML, the pivot towards integrating LLMs to their custom implementations is, well, a perfect fit.

Source: Palantir Q3’2024 business update

Ryan Taylor: In this winner take all AI economy, the divide is widening between those who are leveraging AIP and those who are not. At a leading global insurance organization, AIP has helped automate key underwriting workflows, reducing the typical underwriting response time from over two weeks to 3 hours. We implemented over 10 business use cases in just nine months at Associated Materials, increasing its on time in full delivery rates from 40% to 90%. AtTrinity Rail, it took just three months to get to a functional workflow with a $30 million impact to its bottom line.

Intensity, intent and quantifiable results.

Ryan Taylor: Last quarter, we closed 104 deals, over $1 million. The evolving deal cycle as we take customers from prototype to production is having a particularly phenomenal effect on the growth of our U.S. Commercial business, which continues to see AIP driven momentum both in expansions and new customer acquisitions. In U.S. Commercial, we closed nearly $300 million of TCV and customer count grew 77% year-over-year compared to 37% year-over-year in Q3 2023. To highlight a few notable deal cycles, a large American equipment rental company expanded its work with us less than eight months after converting to an enterprise agreement, increasing the account ARR 12 fold.

It’s difficult to find out the precise sales headcount at Palantir because the company is extremely resistant towards hiring them. It’s quite possible that last quarter 30% of the reps closed at least one $1 million deal.

Ryan Taylor: A bottled water manufacturer, a specialty pharmaceutical company and an agricultural software provider all signed seven figure ACV deals less than two months after their initial boot camps. In our U.S. Government business, we are outfitting our warfighters with advantages over our adversaries. Last quarter marked our U.S. government businesses continued strength through the end of the U.S. government fiscal year. It was our strongest sequential growth in 15 quarters, driven largely by our DoD businesses 21% quarter-over-quarter growth.

Palantir's boot camps, launched in 2023 as part of their expanded commercial sales strategy, are intensive one-to-five day training sessions where potential customers work directly with Palantir engineers to explore practical applications of the company's software platforms.

This hands-on sales approach, which combines technical workshops with customer success stories and real-world data examples, represents a significant shift from Palantir's traditionally selective engagement model - the company conducted over 500 boot camps in 2023, averaging about 5 events per day globally.

The program's success in driving commercial adoption led to a reported 70% increase in US commercial revenue in Q4 2023, with demand necessitating the deployment of European engineers to support US events.

The boot camp strategy follows a structured approach:

Identify impactful use cases
Deploy MVP with customer data
Demonstrate ROI through workflow development
Scale deployment across organization
Build AI-powered foundation for ongoing customer success

Shyam Sankar: The divide between AI haves and have nots is rapidly accelerating in this winner take all AI economy. What will differentiate the AI haves from the have nots is the ability to maximally leverage these models in production by capitalizing upon the rich context within the enterprise. That's why our focus on delivering proof, not proof of concepts continues to pay-off. Years of foundational investments in our infrastructure and in on ontology have positioned us uniquely to harness and deliver on AI demand.
This is Palantir's focus. The market has been focused on AI supply the models.
We see this clearly in the progress, but also in the capital sunk into these models. Indeed, the models continue to improve, but more importantly, the models across both open and closed source are becoming more similar. They are converging, all while pricing for inference is dropping like a rock. This only strengthens our conviction that the value is in the application and workflow layer, which is where we excel.

Last week I talked about AWS’s challenge of trying to move from a sales cycle focused around “providing building blocks” and technical insights to the customer, towards a more “point of view” approach where they try to identify and tie this to the customer use cases.

Shyam Sankar: Tapping into this rapidly expanding pool of leverage from AI labor means more than just saving money. It means a massive acceleration of results for our customers. As Ryan mentioned, we have automated the insurance underwriting process for one of America's largest and most well-known insurers with 78 AI agents, taking a process that took two weeks to 3 hours. More than the labor savings, this presents the customer with an asymmetrical advantage in the marketplace to bind contracts before the competition has even gotten through 15% of their process.

Currently the best performers in Enterprise adoption of AI in 2024 are Azure, Databricks and Palantir. All 3 organisations put a significant focus of bringing on their own expertise and interoperation of how to leverage LLMs and Enterprise-grade Machine Learning to drive significant transformation of existing applications or solve new problems.

Shyam Sankar: In U.S. government, we automated the foreign disclosure process for sharing critical and timely intelligent with allies from three days to 3 hours. The center for Security and Emerging Technologies at Georgetown published a study on Maven that showed how the entire targeting and fires process can be done in Maven with 20 people it used to take 2,000. There is a huge opportunity for our customers to automate the tail and liberate capital to reinvest in the tooth across government and commercial, we see enterprise autonomy as a key theme in our proof. Our deep investments in CJADC2 Combined Joint All Domain Command & Control continue to meet their moments.

What’s more interesting when it comes to Palantir is not only the technology and execution advantage, but also the emotional attachment the company demonstrates towards it’s install base. This is very unusual for how a modern tech or consultancy company operates, but is essential to the success of Palantir.

Shyam Sankar: We are investing aggressively to expand the perimeter to give our warfighters the unfair advantage they deserve, advanced multi-INT sensor fusion, integrated logistics into fires and large scale command and control of swarms of autonomous systems. We announced warp speed last quarter, our modern American manufacturing operating system. We as a nation must reindustrialize to prevent escalating conflict and regain deterrence. Before the fall of the Berlin Wall, only 6% of major weapons systems spend went to defense specialists, the so called primes. 94% went to dual purpose companies who were invested in both freedom andprosperity.
Chrysler built cars and missiles, Ford built satellites until 1990 and General Mills, the serial company made weapons. Today, that 6% has become 86% when including firms whose only commercial exposure is in aerospace. We won World War II and the Cold War with an American industrial base, not a defense industrial base and we need to bring that back at warp speed. And in addition to working with new champions like Anduril and Shield AI, we're also working with L3Harris and two other of the big primes to help them bend atoms better with bits.

Earlier in this article I talked about the divisive reputation of Palantir. The company is one of the rare examples where they have earned the right of being a contrarian - every single thing that they have achieved today is due to making high conviction bets and sticking to them regardless of the relentless negative feedback.

David Glazer: Our U.S. government business grew 40% year-over-year and 15% sequentially, a seven fold increase compared to the prior year period growth rate and the strongest growth we've seen in 15 quarters. On the back of this strength, we are increasing our full year revenue guidance midpoint to $2.807 billion representing a 26% year-over-year growth rate.
We delivered these outstanding top line results, while expanding adjusted operating margin to 38%, highlighting the strong unit economics of our business. Our revenue and profitability drove a 4 point sequential increase to our Rule of 40 score from 64 in the second quarter to 68 in the third quarter.

As any other public tech company, there is a certain degree of asymmetry between the work of the technical and GTM teams and the “outcomes” on the financial markets.

As the stock attracted significant retail attention (379% return since Jan’23), there has been an intense interest in seeing the company “be exposed”. The leadership team has responded to these dynamics in a similar manner to how Musk handled “the shorters” back around the time of the Model 3 launch (taunts, jabs and aggressive counter-moves).

Many see the behaviour of their board and leadership team as unacceptable, particularly also due to the insider selling that gets reported every quarter.

Ana Soro: Thanks, Alex. We'll now turn to a few questions from our shareholders before opening up the call. We received a few questions on AI. How will Palantir differentiate its AI offerings from others, including the model creators? And how is AIP different? And how will Palantir maintain its competitive edge?
Shyam Sankar: Well, Alex talked about how the models yellow arms are commoditized, but if you look at the models, you see that they're getting better, which is awesome. But they're also getting more similar across both closed and open source models, while they're improving, they're converging upon each other, all while the price of inference is dropping precipitously.
And that's, so if you even look at these model companies, they have to build applications around these models to extract value. That's where we have a decade long head start.

At the end of the day, for anybody who is actually involved in Enterprise selling of AI, it should be clear that Palantir’s point of view reflects deep practical understanding of the challenges and opportunities:

Alex Karp: And in a weird way, even though the models are improving, they're meeting up against greater skepticism among clients because clients have tried them and it's just a high school experiment. And then if you get to -- so it's like there's the market and analyst seem to have put a lot of credibility into the models and we do too. We think they're very valuable when managed correctly, when used in a way that an enterprise can understand. And one of the problems that people have is if you're not involved in enterprise software, it's very hard to understand how an enterprise actually works.
You cannot take a large language model that gives you an ELO score of 1200 and use it on targeting on the battlefield. There's a security model. There's a way in which the data is understood. There's certain things you can't share. There's places you would use certain models, but not others. How do you bring that back to your corpus of truth to understand, in a lethal context who dies and who doesn't? And you have very similar use cases in underwriting and in healthcare.

This is not empty fluff, regardless of whether you “like” how it’s delivered. It reflects a deep understanding based on first principles.

Alex Karp: We believe that by investing and we know at this point, instead of trying to have 10,000 clients, all of whom hate you, that's kind of what people want, 10,000 clients that hate you, but they can't give you a product. We want a smaller number of the world's best partners that, quite frankly are dominating with our product. And the way you do that is by having by not blowing up your margin and getting 10,000 sales people, it's actually by going deeper on the product. And in fact, what we see is the deeper and the better the product, the more we drive sales, the more we have our cultural singular advantage as Palantir, not as a commodity product.
It's like we are not a commodity. We do not want our customers to be commodities, we want them to be individual Titans that are dominating their industry or the battlefield. And we reflect that in how we do things. We are not trying to be your average Harvard Business School preferred company with like that, that crush that reduces the margins, has a thin product and then has a lower rule of 40 and presumably higher growth.

“We don’t want 10,00 clients that hate you” is probably one of the harshest reflections of where most large tech companies find themselves in. How you respond to this statement is a great litmus test about whether you see yourself as a mercenary or as a player that can make a difference.

This sounds like a great tech sales opportunity, right?

Oh. It’s that part of the article.

Remember that “Rule of 40” chart? Well the way they get there is by underpaying, hiring predominantly young talent and then running them into the ground.

I would rate the odds of “got fired the day before their wedding” to be approximately 100% accurate.

At the end of the day, you can have a similar “mission focused” experience but actually get paid at Azure or Databricks. Let’s take a look at how the Chief Revenue Officer will tackle this difficult moment and re-shape the sales culture:

I would like to see a CRO with actual track record in establishing a high-performance sales culture, before I can recommend Palantir as a viable opportunity for tech sales.

The Tech Sales Newsletter #67: On Palantir and high conviction bets

The key takeaway

So is there an actual product here?

Palantir and the right time, right place

This sounds like a great tech sales opportunity, right?

The Tech Sales Newsletter #68: The state of AI Agents in late 2024

The Tech Sales Newsletter #66: Q3’24 in cloud infrastructure software