Palantir

Palantir shares are up more than 10% after second-quarter results delivered record growth rates in US commercial revenue (150%) and US government revenue (90%), along with all-time highs in operating margins (62%), net retention (157%), and free cash flow margin (63%).

Why it matters: Palantir has the best growth and margin profile in our software coverage, but its valuation is demanding. Ostentatious growth is needed to placate investors, and we got that this quarter. We think the current share price makes sense, but there’s little margin for error. For an investor to achieve a 10% return, Palantir must benefit from orderly multiple compressions over the next five years while growing at roughly a 45% average annual rate. This growth rate would place it in the top quintile of previously realized growth rates for our software coverage since 2010. Difficult, but not impossible. While most companies would salivate at adding 35% more US commercial customers year over year, we are more impressed by net retention, US commercial revenue, and expanding margins. This combination suggests the platform is seeing deeper deployments and scaling incredibly well.

The bottom line: We maintain our $153 fair value estimate for narrow-moat Palantir. We see the business thesis strengthening as customers demand model portability, data sovereignty, and governed actions in artificial intelligence deployments, but we believe the share price is adequately capturing Palantir’s impressive trajectory. We believe Palantir’s ontology—the topological map that automates advanced decisioning by connecting customers’ data silos—is differentiated and hard to reproduce with frontier model solutions or cloud platforms today, but given the rapid pace of model advancements, we assume growth normalizes beyond the next 10 years.

Palantir announced intriguing partnership expansions in July. A “Sovereign AI” partnership with Nvidia aims to give customers full control over AI deployments, rather than handing data and control to a frontier lab.

These partnerships support Palantir’s total addressable market and solidify its position as the leading provider of secure, on-premises artificial intelligence stacks—an end market where we expect strong growth. In the Nvidia partnership, Palantir seamlessly integrates its governance and decisioning operating system (the software layer) with Nvidia’s chips and open-weight models (Nemotron), so security-conscious customers never need to send data externally to the public cloud or risk their data being used against them by a frontier lab that moves downstream. The suspected Iranian cyberattacks on critical water infrastructure this month further underscore the importance of resilient government infrastructure, where Palantir and Nvidia are likely to play leading roles.

We also like the expansion of the relationship with the US Army, where Palantir was named the primary cloud data platform for the Army’s Next-Generation Command and Control program, an initiative aimed at replacing disconnected military technology systems with a unified digital environment. While Palantir has long been the preeminent operating system for the US military, we see this as evidence of immense switching costs associated with mission-critical operations and a continued vote of confidence in Palantir over newer AI-enabled systems from frontier model providers.

Trading at an enterprise value of 60 times sales, the growth rates likely needed to deliver a 10% return require performance above the interquartile range of some of the most influential technology companies, but we think this is achievable given Palantir’s current growth trajectory, margin expansion, and cash generation. The debate is not whether Palantir is a high-quality company. It’s about whether the durability of growth is strong enough to withstand multiple compressions. We think Palantir can deliver.

We have also been monitoring the industrial and capital-spending cycle, particularly core capital-goods orders, which signal business investment in equipment. Core capital-goods orders have risen to roughly two standard deviations above their 10-year trend, indicating that US companies are building more domestic production capacity. As factories and supply chains grow and become more complex, demand for efficiency-driving software like Palantir should increase. Overall, we see reindustrialization impulses like this as a potential upside lever to our valuation.

Our fair value estimate of $153 implies a 2026 enterprise value/sales multiple of 48 times.

In our opinion, the primary driver of the stock’s value is the total addressable market Palantir’s software can ultimately serve. TAM size is truly a trillion-dollar question that is unfortunately laden with assumptions. Our base case has Palantir’s TAM growing to $1.4 trillion by 2033. We forecast 52% average annual growth between 2026 and 2028. Our analysis concludes that we are in the early innings of an AI revolution. In our base case, we expect Palantir to have a growth profile similar to that of innovative software companies like Salesforce in the late 2010s. Salesforce drove efficiency by creating a standardized workflow and logging process for enterprises accustomed to bloated sales teams compiling data in disparate locations. We expect Palantir to similarly drive efficiency among enterprises that now rely on large information technology teams to interpret and present data to support decision-making.

If Palantir executes well, it could become essential for businesses seeking to aggregate and analyze data, much as Salesforce became essential for sales teams. As such, we believe most of Palantir’s growth will come from the US commercial business, and we expect net revenue retention rates from existing clients to remain strong. Our bull case scenario calls for a TAM larger than $1.6 trillion and penetration of nearly 3%, resulting in a valuation of approximately $300 per share.

Since going public, Palantir’s gross margins have risen from an initial low of 68% in 2020 to 84.5% in late 2025. The 2020 low gross margin stemmed from the accumulation of restricted stock units that vested immediately upon the company’s public offering. Since then, the company has steadily improved its gross margin by scaling revenue faster than growth in salaries, hardware costs, stock-based compensation, and payments to cloud hosting services. We project gross margins to remain in the 83-85% range over the next 10 years as we balance the onboarding of higher-margin enterprise customers with the potential for cloud costs to rise amid ever-increasing demand for computing resources.

Palantir has spent heavily on research and sales in the past. In 2023, the firm’s research and sales spending combined totaled approximately 52% of revenue. That said, Palantir’s top-line revenue has been growing faster than research and sales costs, ultimately benefiting the bottom line. We like Palantir’s rollout of a ‘bootcamp’ based sales approach, which encourages rapid turnover of sales touchpoints at a lower customer acquisition cost, all while still providing a customized solution that organizations expect from Palantir’s ontology framework. Palantir’s capability to grow the top line while keeping major costs in check has been a positive trend that should continue.

Palantir is the premier player in AI decision-making software, and we are bullish on its growth and profitability going forward, but we are mindful that AI enthusiasm can decouple market pricing from our intrinsic valuation.

We believe that Palantir warrants a narrow Morningstar Economic Moat Rating, based on switching costs and intangible assets.

Companies like AWS, Snowflake, and ServiceNow have developed data analytics tools, but Palantir differentiates itself as the only AI company with a framework that organizes disparate datasets and facilitates optimized decision-making. This machine-learning framework, which identifies opaque yet significant relationships in data and creates solutions for the end user, is referred to as the "ontology framework." Palantir engineers a read-write feedback loop that enables connectivity throughout a business, creating an accessible analytical framework to drive nuanced decision-making that improves over time. Palantir often competes against internal information technology departments because it also analyzes data and creates information dashboards for interpretation. This traditional in-house IT and data aggregation framework often results in patchwork solutions that are cumbersome, difficult to improve, and costly to scale.

To differentiate, Palantir creates a comprehensive, closed-loop system in which data flows from individual sources and data users. Put simply, Palantir gathers customer data of all forms, from structured data (stored in rows and columns) to unstructured (lacking a defined schema like emails, reports, logs from IT systems, or video/image files) to intelligence data (satellite data, troop movements, criminal records, threat assessments) to operational information (supply chain, demand, production or output), to analog sources (physical-world data like temperature, pressure, sound, or vibration in factories or field environments). The software cleans this data, identifies relationships, and deploys machine-learning algorithms that derive operational insights on the ground, ready for interaction and implementation.

Palantir’s data-driven insights create actionable productivity enhancers throughout all roles and operations in an organization. As of 2023, Palantir provides an upsell artificial intelligence platform, or AIP, that acts as a large language model orchestration tool, lowering the barriers to adoption for all nontechnical users. The core ontology function and value proposition is that Palantir not only organizes and displays data, but it also creates prioritized, ranked data that can be quickly understood and interacted with, ultimately automating real-world efficiency gains. Whether it is aggregating legacy data lakes, understanding patient data to increase bed turnover, or digitizing analog data from around the world, the company is driving efficiency gains from data. The real-world efficiency gains accumulate and translate into switching costs for customers. Effectively, to stop using Palantir's ontology framework means massive inconvenience and a monetary cost through lost productivity. Simply buying the software and not renewing is not an option if customers want to maintain their newfound edge. Palantir provides a dashboard that continually updates and learns as time progresses. Without the software, ex-customers would need to rely on former systems that lack the complex ontology framework, and they would lose the simplified dashboard that brings the productivity gains together.

To quantify switching costs as a moat source, we start with the most widely accepted indicator: net revenue retention. Palantir has the highest NRR (approximately 120%) across other data-forward software companies like Snowflake, DataDog, and Splunk. Palantir's NRR has also been increasing rapidly quarter over quarter since the release of its AIP. NRR is effectively a measure of a company's ability to capitalize on touchpoints and expand interaction with the upselling of additional capabilities, and AIP is a great example. Palantir’s top-of-class and rapidly improving NRRs are an exhibition of not only customer stickiness but also expanding adoption. Customers see value in Palantir's framework, and they continue to invest in it. We expect this trend to continue.

The average Palantir customer spends more than $1 million a quarter on the software. Customers have sunk significant monetary, intellectual, and data capital into acquiring and learning the software and are reaping associated rewards. Switching often doesn't make sense.

To anecdotally quantify the high-stakes problems that Palantir is solving that contribute additional switching costs, one must look no further than mission-critical healthcare and military operations. For example, Tampa General Hospital partnered with Palantir and reduced patient length of stay by 30%, helping save lives. On the military front, Palantir's software was critical to accumulating intelligence data and providing the framework for capturing Osama Bin Laden. When complex issues involve human lives, institutions must rely upon the most robust frameworks and software solutions, which Palantir provides. Switching away from this novel platform could open the customer to both monetary and immeasurable human costs.

Customer growth has been impressive. Since bottoming in mid-2023, we have seen rapid year-over-year growth in the US commercial segment. Facilitating this growth is the recently introduced boot camp strategy for its sales and marketing efforts, which allows for a high number of touchpoints with prospective buyers. Once landing a contract, there is still a specialized onboarding period when Palantir engineers learn the ins and outs of the customer's data to tailor the optimal framework for complex processes. We think this unique customer acquisition and retention process deepens Palantir's relationship with the end user and boosts growth. We also appreciate that the company has been diversifying its revenue base by adding commercial customers, which often come with more lucrative contract terms compared with those in the government segment.

Overall, Palantir is an innovative AI company that offers nuanced and optimized solutions derived from the world's most complex datasets. The company is deeply ingrained within diverse end markets and mission-critical customer infrastructure and has top-of-class net revenue retention metrics that justify a narrow moat rating.

Palantir has developed the premier AI software, well-positioned to capitalize on trends toward digitization, automation, and reindustrialization. We believe Panatir's software maintains a strategic position in the AI value chain as a model orchestrator.

Palantir's ontological framework and AI orchestration allow for the democratization of machine learning. Its software is useful to employees at all levels of a business to drive efficiency enhancements.

Palantir stands to disproportionately benefit from a Golden Dome-led fiscal spending boom and lacks a clear competitor.

Palantir’s end markets are confined to entities that coalesce with Western ethos. This caps the total addressable market.

The declining cost of AI inference and improvements in agentic LLMs will lower barriers to entry in the AI decision-making software industry, which Palantir currently dominates. Anthropic and OpenAI could become viable competitors.

Palantir’s high valuation multiple leaves no margin for error in terms of execution. Any fears on the maintainability of growth will be met with sharp selloffs.