Palantir Technologies: anatomy of a data empire

By Kike Gandia · Research · 58 min read

Palantir Technologies, Inc. was founded in 2003 (incorporated in May of that year) by a group of former PayPal employees and Stanford students led by Peter Thiel, together with Alex Karp, Joe Lonsdale, Stephen Cohen and Nathan Gettings. The name Palantir comes from the “seeing stones” in The Lord of the Rings, a nod to a technology for “seeing from afar.” Thiel funded the early days with $30 million and secured $2 million from the CIA’s investment arm (In-Q-Tel). Palantir’s original mission was to build data analysis software inspired by PayPal’s anti-fraud tools, in order to detect hidden patterns in large datasets to fight terrorism and fraud after 9/11 while preserving civil liberties. Alex Karp took on the CEO role early on, with Thiel as chairman. The first years focused on developing the platform alongside intelligence analysts through pilots with government agencies.

1. 2003 — Founding
   Former PayPal employees and Stanford students led by Peter Thiel, together with Alex Karp, Joe Lonsdale, Stephen Cohen and Nathan Gettings.
2. 2005-2012 — Early customers
   The growth years under the wing of US intelligence and defense agencies.
3. 2013-2016 — Expansion and diversification
   From government into the private sector, with Foundry as the enterprise platform.
4. 2020 — Going public
   Direct listing on the NYSE and a customer base broadened into healthcare, industry and energy.
5. 2022-2025 — Ukraine and AI
   Prominence for its technology support to Ukraine, the bid for the UK NHS health data contract and the launch of AIP.

Twenty years of Palantir, from the back rooms of intelligence to the stock market spotlight.

Early customers and growth (2005-2012)

Palantir’s first customers were US intelligence agencies (CIA, FBI, NSA) and Department of Defense units. Between 2005 and 2008 the company refined its Palantir Gotham product (formerly called Palantir Government), aimed at the intelligence and counterterrorism community. The FBI and CIA adopted Palantir to connect previously siloed databases, enabling integrated searches that used to require querying separate systems. A notable early case came in 2009, when the Recovery Accountability and Transparency Board (RATB) used Palantir to uncover fraud in economic stimulus programs: then Vice President Joe Biden credited Palantir’s software with the success of those investigations. Around 2010, Palantir was also helping uncover cyber-espionage networks such as GhostNet and the Shadow Network through its partnership with the Information Warfare Monitor. During this period the company began expanding beyond government into the private financial sector: in 2010 Palantir launched Palantir Metropolis (also called Palantir Finance), focused on quantitative financial data, and partnered with Thomson Reuters to sell Metropolis to financial analytics customers.

Expansion and diversification (2013-2016)

By 2013, Palantir already counted multiple US agencies among its customers (CIA, NSA, FBI, the Department of Homeland Security, the Marine Corps, the Air Force, Special Operations Command, the West Point military academy, and others). That same year, CEO Alex Karp said Palantir had no plans to go public in the near term because “being public would make it very hard to run a company like ours.” The company nevertheless kept attracting massive private funding: funding rounds lifted its valuation to $9 billion in 2013 and $15 billion by the end of 2014, making it one of Silicon Valley’s most valuable startups. Its investors included In-Q-Tel (CIA), Founders Fund (Thiel’s fund) and tycoons such as Kenneth Langone and Stanley Druckenmiller. In 2014-2016 the company raised hundreds of millions more (including $880M in 2015), reaching valuations close to $20 billion. During these years Palantir diversified its applications: it introduced its Palantir Foundry platform (publicly launched in 2016), aimed at commercial enterprises and designed to integrate business data for operational use. Foundry in fact ended up succeeding Palantir Metropolis as the main platform for financial-sector and corporate customers. Palantir also acquired complementary technologies, such as the startup Kimono Labs in 2016 (web scraping), to strengthen its data integration capabilities.

Going public and the present day (2017-2025)

After years of speculation, Palantir finally decided to go public in 2020. It opted for a direct listing (DPO) on the NYSE on September 30, 2020. In its last year as a private company (2019), Palantir had revenue of ~$740M but net losses of ~$580M, reflecting heavy investment in growth. Once public (ticker PLTR), Palantir kept expanding its customer base beyond government into healthcare, industry and energy. During the 2020 pandemic its tools were used to manage health data (e.g. COVID-19 tracking and vaccine distribution), which raised its profile in public health. In 2021 Palantir formed strategic partnerships, including one with IBM to integrate Foundry with IBM’s AI tools (Watson) and broaden its enterprise reach. In 2022-2023 the company gained prominence for its technology support to Ukraine during the Russian invasion (CEO Karp himself traveled to Kyiv and opened an office there) and for competing for a major health data platform contract with the UK’s NHS. Palantir also kept innovating on the product side with an emphasis on artificial intelligence: in 2023 it launched Palantir AIP (Artificial Intelligence Platform) to integrate LLMs (large language models) into private networks. In early 2023 the company reported its first quarters of positive net income, and by 2024 it had reached ~$2.87 billion in revenue with ~3,900 employees. Palantir has become a key data infrastructure provider for governments and businesses alike, though not without controversy over the privacy implications of its technologies.

~$740M

revenue in 2019, its last year as a private company

~−$580M

net losses that same year, driven by investment in growth

09/30/2020

NYSE debut via direct listing, ticker PLTR

Key figures

Since its beginnings Palantir has been led by Alex Karp (CEO) and Peter Thiel (co-founder and chairman). Thiel, a well-known tech investor and PayPal co-founder, was the largest early shareholder and brought the “mission-driven” vision of applying software to national security. Karp, who holds a PhD in philosophy, has been Palantir’s public face, defending its role in Western security while criticizing other tech giants for not working with the defense establishment. Other co-founders include Stephen Cohen (current president), Joe Lonsdale (who left the company years ago) and Nathan Gettings. Its advisory board has included figures such as former Secretary of State Condoleezza Rice and former CIA director George Tenet, reflecting Palantir’s close ties to the intelligence community.

Watch the video on YouTube

Palantir’s products: current and discontinued platforms

Palantir has built specialized software platforms for different types of users and sectors. Its four main current offerings are Gotham, Foundry, Apollo and AIP, plus the now-discontinued Palantir Metropolis (Palantir Finance). Each is described below, including its purpose, general architecture, target sectors and evolution:

Government Gotham

The original flagship product, for intelligence, defense and security agencies: it integrates massive datasets and uncovers hidden connections between entities.

Enterprise Foundry

Launched in 2016, the “operating system for the modern enterprise”: it integrates disparate data (databases, spreadsheets, IoT sensors) into a single model.

DevOps Apollo

The continuous deployment platform, born around 2019 to orchestrate Gotham and Foundry updates in isolated environments, including classified networks.

AI AIP

Launched in April 2023 to integrate large language models and other generative AI into operations, inside private networks.

Retired Metropolis

The former Palantir Finance, focused on quantitative analysis in finance and inspired by PayPal’s anti-fraud tools. Discontinued.

The four current platforms and the one that fell by the wayside; each is covered in detail below.

Palantir Gotham

Palantir Gotham is Palantir’s original flagship product, designed primarily for government intelligence, defense and security agencies. It is a data integration and analysis platform focused on uncovering hidden connections between entities in massive datasets, both structured and unstructured. Gotham lets counterterrorism analysts, military personnel or police officers unify scattered databases (person records, transactions, calls, license plates, reports, etc.) and run advanced searches, visualizations and analyses to support operational decisions. For example, an analyst can enter a vehicle’s license plate and Gotham will instantly show every location, date and relationship tied to that vehicle across different police databases. Gotham’s architecture emphasizes human-machine collaboration (“augmented intelligence”), on the premise that algorithms alone cannot defeat adaptive adversaries without interaction with human analysts. That is why Gotham provides tools for the analyst to filter and steer the analysis (e.g. link networks, timelines, maps). It also includes privacy and audit controls (granular tagging of sensitive data, role-based access management, etc.) to meet legal requirements and protect civil liberties.

Target sectors

Gotham is aimed at intelligence and defense. Its users include agencies of the US Intelligence Community (USIC) such as the CIA, NSA and FBI, the Department of Defense (Army, Air Force, special operations commands) and national and allied security forces. It has also been adopted by police departments for criminal analysis (fusion centers, crime prediction programs). Gotham has been crucial in military settings: it was deployed in Afghanistan to help predict the location of IEDs (improvised explosive devices), with results that soldiers considered superior to the Army’s own analysis system (DCGS-A). Likewise, it has supported operations against terrorist networks (it is credited, according to unofficial reports, with helping locate Osama bin Laden) and cyber-espionage investigations (the GhostNet and Shadow Network cases).

Evolution

Gotham keeps being updated for dynamic operational environments. Palantir has adapted Gotham to run on-premises and in classified environments (e.g. air-gapped military networks) and also as a service in the US government’s private cloud, earning high-level security authorizations (DoD Impact Level 5). Gotham was initially offered only to US government entities, but over time it was also deployed with allies (Europe, the Middle East, etc.). Notably, Gotham replaced traditional link analysis tools such as IBM i2 Analyst’s Notebook in many cases, offering greater data integration and real-time collaboration. Today, Gotham remains focused on national security missions, tactical intelligence and fighting crime, increasingly integrating artificial intelligence capabilities (computer vision, facial recognition, etc.) subject to human oversight for tasks such as target identification on the battlefield.

Palantir Metropolis (Palantir Finance) – discontinued

Palantir Metropolis was the platform built for quantitative analysis in finance. Formerly known as Palantir Finance, it was inspired by PayPal’s anti-fraud tools and was one of the first products to bring Palantir’s technology to the private sector. Its purpose was to integrate and analyze massive financial datasets (stock markets, transactions, financial instruments, etc.) to uncover trends, relationships and anomalies in real time. Metropolis connected commercial, proprietary and public data for uses such as financial fraud detection, risk management at banks and funds, and predictive market analysis. Architecturally, it offered capabilities similar to Gotham’s (search, alerts, network visualization) but adapted to the dynamics of numerical and market data.

Target sectors

Metropolis focused on financial services. It was used by hedge funds, investment banks and insurers from the late 2000s onward. For example, J.P. Morgan used Palantir Metropolis around 2009 to monitor employees’ internal communications and flag anomalous behavior after the “London Whale” scandal. A team of 120 Palantir engineers helped the bank track employees’ emails, browsing history, calls and GPS locations to detect potential leaks or insider risks. Other banks such as Credit Suisse and Morgan Stanley were also early Palantir customers for regulatory compliance and risk analysis applications.

Evolution

Palantir Metropolis was officially discontinued as Palantir consolidated its platforms. Around 2016-2017, Metropolis’s functionality was absorbed into Palantir Foundry, the new, more versatile enterprise platform. Foundry inherited Metropolis’s financial integration capabilities while adding more modern data modeling tools. As a result, Metropolis was no longer offered as a separate product. Even so, the financial expertise Palantir built through Metropolis laid the groundwork for many Foundry deployments at banks and insurers in later years. Metropolis is an example of how Palantir moved from vertical solutions (government vs. finance) toward a single, flexible platform (Foundry) that spans multiple sectors.

Palantir Foundry

Palantir Foundry is Palantir’s enterprise platform, often described as an “operating system for the modern enterprise.” Launched in 2016, Foundry lets private organizations integrate all their disparate data (databases, spreadsheets, IoT sensors, etc.) into a unified environment for analysis, collaboration and decision-making. Its architecture is modular: it provides a central Ontology layer, or unified data model, on top of which different users can interact in different ways (from data engineers writing pipelines to business analysts using dashboards, and even low-code workflow applications). Foundry combines ETL integration tools, a central repository (a hybrid data lake/warehouse) and powerful analytics features (SQL, Python/R, machine learning) alongside no-code interfaces. A distinctive trait is how quickly it lets teams build operational applications: for example, an internal app for plant managers to see maintenance alerts, or for logistics staff to visualize supply chain bottlenecks. All of this runs on the same integrated data foundation, ensuring consistency and governance (Foundry offers object- and data-level access control, data lineage tracking and full auditing).

Target sectors

Foundry is aimed at companies in every industry looking to accelerate their digital transformation through data. It has been adopted in manufacturing, energy, healthcare, finance, transportation, telecommunications, retail and more. For example, the pharmaceutical company Merck KGaA uses Foundry to integrate research data and speed up clinical development. Aerospace manufacturer Airbus uses Foundry as the core of its Skywise platform to analyze aircraft data and improve maintenance (see the use case below). Automakers such as Fiat Chrysler (FCA) and the Ferrari racing team have implemented Foundry to optimize everything from vehicle design to on-track performance. In energy, BP has used it to monitor its oil operations and, more recently, renewable energy projects. Foundry has also been deployed at government agencies (e.g. the US Department of Health and Human Services and the Internal Revenue Service/IRS) to integrate administrative data and improve operational efficiency.

Evolution

Foundry has become Palantir’s flagship commercial product, evolving to be more interoperable with existing ecosystems. For example, it offers connectors to clouds such as AWS, Azure and Oracle, and integrates open APIs and standards so as not to “lock in” data. Palantir has launched specialized editions such as Foundry for Manufacturing (on AWS), with predefined models for industry. With the rise of AI, Foundry added a notebook workspace for data scientists and AutoML tools, and connects to the new AIP platform to bring generative AI agents to its data. One point worth highlighting is Palantir Apollo (below), which was created to deploy and update Foundry and Gotham continuously, even in on-premises or edge environments. Foundry has proven its scalability on huge projects: according to Palantir, some deployments handle petabytes of data and thousands of concurrent users. Foundry does, however, usually require a significant upfront investment in configuration and customization, which pays off in environments where the operational results (savings, optimization) offset that cost.

Palantir Apollo

Palantir Apollo is Palantir’s DevOps and continuous deployment platform. Launched around 2019, it started as an internal tool to orchestrate the delivery of Gotham/Foundry updates to customers with isolated environments (such as servers on classified government networks). Apollo has since grown into a product for managing, monitoring and updating software across multiple environments in an automated way. In essence, Apollo acts as a centralized “mission control” for software: from Palantir’s cloud, it can deploy new versions, patches and configurations of Gotham/Foundry to customers’ on-premises installations, making sure they are all up to date. It also monitors performance and can apply autonomous remediation when it detects problems (restarts, rollbacks, scaling). This is crucial for military or industrial customers running Palantir at sites with no internet connection: Apollo “pushes” updates when it detects safe windows.

Features

Apollo supports non-linear CI/CD workflows, making it possible, for example, to keep several versions running in different environments depending on regulatory needs. It ships with container-based orchestration tools (Kubernetes) and meets high security standards (cryptographic package validation, etc.). Its design unifies Dev, Sec & Ops concerns, allowing small cells of Palantir engineers to manage hundreds of customer deployments efficiently. Apollo is extensible to third-party software; although it was built for Palantir, it can manage other systems on the same principles.

Target sectors

Apollo is aimed at enterprise or government customers with hybrid or highly regulated environments. For example, defense organizations running Palantir at remote bases or on ships, or companies with private clouds, can use Apollo to keep their software up to date without exposing those environments. It is also attractive to companies that want to speed up their release cycles without building a large DevOps team. In practice, Apollo is what lets Gotham and Foundry run as software as a service even on customer premises. This has been key to Palantir’s strategy of delivering continuous (often weekly) improvements to its platforms without disrupting its users’ critical operations.

Evolution

Palantir has gradually opened Apollo to more users, even offering it on cloud marketplaces (Azure, AWS) as a standalone deployment solution. It continues to be refined to handle “edge” environments: for example, updates on field devices, sensors, drones and so on, complementing initiatives such as MetaConstellation and Skykit (see below). Apollo reflects Palantir’s bet on software autonomy, cutting administration costs and letting its engineers focus on building features rather than on-site support. In short, Apollo is the least visible but critical component that ensures Gotham and Foundry run 24/7 with the latest capabilities anywhere in the world.

Palantir AIP (Artificial Intelligence Platform)

Palantir AIP is the company’s new artificial intelligence platform, launched in April 2023 to securely integrate large language models (LLMs) and other generative AI into organizations’ operations. AIP emerged amid the explosion of tools such as GPT-4, but focuses on enterprise and government environments where these AI systems need to be used with confidential data and under human control. The AIP platform lets organizations connect LLMs (third-party or in-house) to the data “ontology” they already have in Foundry/Gotham. In other words, it creates a layer where an AI model can access real operational data (documents, internal databases) in a governed way to provide answers or automate tasks, while keeping logs and enforcing access policies.

Features

AIP offers an interface that lets non-expert users create AI “agents” tailored to their needs. For example, a business analyst can, without writing code, spin up an agent that answers questions about the company’s inventories or generates report summaries, because the LLM is linked to the relevant internal sources. The platform lets users define templates, test cases and output schemas to validate that the AI’s answers are correct and safe. It also implements human controls: according to Palantir, AIP will not carry out critical actions autonomously without human approval, to mitigate the risk of runaway AI. In a military context, it was demonstrated that an operator could ask the AI to select targets or recommend tactics, but the final decision rests with the human (there is no “autonomous Terminator”). AIP also includes a library of prebuilt templates for different domains (e.g. infrastructure planning, logistics network analysis, resource allocation) to speed up adoption. Palantir has even run 5-day training bootcamps (AIP Bootcamps) for customers and an annual AIPCon event with demos, given how new the technology is.

Target sectors

AIP has applications in both defense and commercial industries. In defense, it was pitched as a force multiplier: in simulations, a military analyst could ask an AIP chatbot about enemy movements and get tactical recommendations in seconds. In fact, Palantir says AIP is already used in combat scenarios to deploy AI models that assist in the field (under supervision). In the commercial sector, infrastructure, telecommunications and finance companies, among others, are exploring AIP to summarize huge documents, optimize plans or detect risks in complex operations. The promise is to bring LLMs “behind the firewall”: companies can tap ChatGPT and similar tools with their private data without exposing it externally. Palantir has stressed that AIP is especially useful for organizations with sensitive data and regulated environments, where using public AI is not viable because of the risk of leaks. In short, AIP extends Palantir’s reach into the wave of generative AI, competing with offerings such as Microsoft’s Copilots and IBM watsonx, but with the differentiator of native integration with Palantir’s data platforms and a focus on security.

Evolution

Young as it is, AIP has been adopted quickly in pilots and contracts. In 2023 Palantir stressed that it can take AI models from prototype to real production “like no other company,” in Karp’s words. Palantir frequently demos AIP to prospective customers and has developed specialized hardware in parallel (such as the Titan and Skykit systems, below) to run AI in the theater of military operations. Going forward, AIP can be expected to integrate even more tightly with Foundry (e.g. to generate code automatically or explain insights) and to support more types of models (vision, audio). AIP embodies Palantir’s strategy of leading the enterprise artificial intelligence platform market, something confirmed by the analyst firm Forrester, which in 2024 named Palantir a leader in AI/ML platforms, ahead even of Google and Microsoft.

Other specialized initiatives: TITAN, MetaConstellation and Skykit

Beyond its main platforms, Palantir has developed specialized solutions, mostly for military customers, that combine software and hardware for advanced use cases:

• Palantir TITAN (Tactical Intelligence Targeting Access Node): a mobile tactical intelligence system. Essentially, it is a vehicle (truck) fitted out as a battlefield data station. TITAN integrates sensors, satellite communications and Palantir’s software suite (Gotham/Foundry/AIP) to give deployed units on-site AI analysis capabilities. It was initially developed with Palantir’s internal funds (IRAD) and is now built in partnership with defense companies such as Anduril and Northrop Grumman. In 2022-2023 the US Army contracted Palantir to deliver at least 10 TITAN units, with the promise of improving long-range target identification and the coordination of precision strikes from hostile terrain. In short, it brings “the Palantir cloud” to the mobile theater of operations.
• MetaConstellation: a network and service that connects satellites and AI models for real-time surveillance and response. With MetaConstellation, a user (a military one, for example) can request information about a specific geographic coordinate; the platform then orchestrates multiple satellites and drones to collect imagery, which is analyzed with AI (object and change detection) and sent to the user with minimal delay. This ability to coordinate satellite constellations gives analysts a near-live view of different areas. Commands such as US Northern Command have reportedly used MetaConstellation to improve situational awareness in homeland defense. It is an example of how Palantir extends Gotham/Foundry into the space domain.
• Skykit: a portable intelligence kit for operating in harsh environments. It consists of a case or backpack (there are Skykit Backpack and Skykit Maritime versions for boats) that includes a ruggedized laptop with Palantir’s software preinstalled, long-life batteries and a mini quadcopter drone for capturing video. The goal is for field teams (e.g. special forces, rescue workers) to be able to set up a mini Palantir station in minutes, with satellite connectivity (via MetaConstellation) and computer vision capabilities through the drone. In 2023 it emerged that the Ukrainian army had received Skykit units for its officers on the front line. With Skykit, Palantir offers intelligence “in a backpack”, so that even in areas with no telecommunications infrastructure, data (imagery, sensor readings) can be collected and analyzed locally with advanced algorithms.

These initiatives show Palantir’s push toward end-to-end defense solutions, combining its software with specialized hardware to ensure that the “last mile” of data (whether the soldier in combat or the engineer on the factory floor) has access to the same advanced analysis tools.

Palantir’s customers by sector

Palantir serves both governments and private companies. Below are some of its most prominent customers, grouped by sector, with the solutions they use, since when and in what context:

Defense and security (military and intelligence)

• US Department of Defense: Various US military branches have used Palantir Gotham for field intelligence since the late 2000s. Units of the Marine Corps, Air Force, Special Operations Command and others used Palantir in Afghanistan and Iraq to analyze threats (for example, predicting IEDs). The US Army was initially reluctant to adopt it, favoring its own DCGS-A system, but after a 2016 lawsuit and a court order, Palantir was able to compete and supply its solution as part of the Army’s tactical intelligence equipment. In 2019 Palantir won the Pentagon’s Project Maven contract (AI for analyzing drone video) after Google pulled out. More recently, in 2023, Palantir secured a ~$480M contract with the DoD’s Chief Digital and Artificial Intelligence Office for an AI-enabled operational platform.
• US Intelligence Community: The CIA, NSA, FBI, DHS and other agencies were Palantir’s first customers. From 2005 onward, Gotham was used to connect their classified databases and enable cross-agency analysis for the first time. The FBI applied Palantir to uncover domestic terrorist networks and financial fraud cases; the NSA to correlate signals and human intelligence; and the CIA to integrate field intelligence with public data. These deployments cemented Palantir’s reputation as an essential “cyber-surveillance” tool after 9/11. They also drew criticism over opacity and privacy, however, by enabling the mass aggregation of data on individuals.
• UK Ministry of Defence: The British MoD turned to Palantir after struggling to analyze large volumes of intelligence in the age of terrorism. Palantir was hired in 2016 to equip French intelligence (DGSI), through NATO, after the Paris attacks, and the UK similarly integrated it into its forces. In 2023 the MoD signed a contract worth £75 million (~€91M) to deploy Palantir Gotham/Foundry as a permanent military intelligence platform. The UK has used Palantir both in joint military operations with the US and in cybersecurity missions.
• Israel Defense Forces: In January 2023, Palantir announced a partnership with Israel’s Ministry of Defense to provide technology support for its operations. Although the details are confidential, it is understood to involve the use of Gotham and possibly AIP for tactical intelligence, given Israel’s interest in military AI. Palantir already had earlier ties to the country through Unit 8200 (Israeli military intelligence).
• NATO and Western allies: Palantir has been steadily building a presence in international coalitions. In 2024 NATO acquired a system called the Maven Smart System (MSS NATO), developed by Palantir, which incorporates generative AI, ML and LLMs for allied forces’ joint strategic operations. This suggests Palantir is providing NATO with an adapted version of its platform (probably AIP + Gotham) to improve intelligence coordination among member states. Other countries, such as France, Germany, Canada and Spain, have also used or are evaluating Palantir in their armed forces. For example, in 2023 Spain’s Ministry of Defense awarded Palantir an €8.5M contract for a proof of concept of a military intelligence platform.

Healthcare

• UK National Health Service (NHS): The NHS is one of the most high-profile cases. In March 2020, during the COVID-19 pandemic, the NHS awarded Palantir an emergency contract to integrate data on patients, tests, beds and more into a single platform (the COVID-19 Data Store) using Palantir Foundry. That platform helped visualize the real-time situation on dashboards and coordinate the distribution of medical resources. The initial contract (£23.5M) was extended for two years, until 2022. Then, in 2023, the NHS put out to tender a long-term federated health data platform (£480M), seen as a “must-win” for Palantir. Palantir, regarded as the favorite given its existing presence, reportedly (according to press sources) secured a deal worth close to £400M to embed its software permanently in the NHS. This sparked controversy over privacy and Palantir’s possible competitive advantage. Even so, the NHS has said Palantir was “critical to the success” of programs such as the mass vaccination rollout and PPE (personal protective equipment) management. Eleven British hospital networks came to use Foundry, although some paused their use in 2023 over “operational issues” during the transition.
• Centers for Disease Control and Prevention (CDC, US): The US public health agency has used Palantir since at least 2010 for biosurveillance. In 2012, Palantir was among the CDC’s tools for tracking epidemic outbreaks and bioterrorism threats, integrating epidemiological data with other information. During the Ebola crisis (2014) and later with COVID-19, the CDC also used Palantir Gotham/Foundry to consolidate data on cases, contacts, vaccine inventory and more, supporting a coordinated response. Foundry was also used in 2020 for COVID vaccine distribution in the US, through a federal system that tracked vaccine manufacturing, allocation and administration nationwide.
• National Institutes of Health (NIH) and HHS (US): The NIH (the main US medical research body) has been a Palantir customer for biomedical data analysis projects. For example, it integrated clinical research data into Palantir to speed up discoveries, although the details are not public. The Department of Health and Human Services (HHS), which oversees the NIH and other agencies (FDA, etc.), uses Palantir Foundry in at least 4 of its agencies today. Applications include tracking health crises, managing public health programs and disaster response (e.g. coordinating medicine supplies during hurricanes).
• Pharmaceutical and life sciences companies: Palantir has won major customers in this sector. Germany’s Merck KGaA uses Foundry to unify drug discovery, clinical trial and manufacturing data, ensuring regulatory compliance while identifying opportunities to accelerate time to market. French drugmaker Sanofi has also worked with Palantir on data projects (for example, in 2016 to make use of diabetes patient data). NHIS (Korea’s national health service) hired Palantir in 2022 to build its health Big Data platform on Foundry, focused on medical research based on anonymized records. These partnerships show how far Palantir has penetrated private-sector healthcare, helping with tasks such as pharmacovigilance, drug supply chain optimization and genomic data analysis at scale.

Finance

• JPMorgan Chase & Co.: The banking giant was one of the first to trial Palantir for internal surveillance. In 2009 its security team used Palantir Metropolis to identify employees at risk of leaking information or committing fraud, after suffering losses from unauthorized trading. Palantir helped monitor communications (emails, calls), file access and after-hours behavior, detecting suspicious patterns. Although the project stirred internal debate (over workplace privacy implications), it showed how banks could use big data for compliance. JPMorgan later used Palantir for market risk analysis, although reports indicate that in recent years it has built more in-house capabilities.
• Morgan Stanley: Another major Wall Street firm, Morgan Stanley has been cited as a user of Palantir Foundry for financial data integration. Foundry would allow it to consolidate investment, trading and research data on a common platform for analysis by its quants and analysts. Although the details are confidential, Morgan Stanley is known to have invested in Palantir and to have applied its tools in areas such as card fraud detection and portfolio optimization.
• Credit Suisse: The Swiss bank is cited as a Palantir customer for risk analysis and anti-fraud. In 2016 it emerged that Credit Suisse was using Palantir to strengthen its anti-money laundering controls, linking transactions with databases of individuals and regulatory alerts. Its use in private banking to analyze client networks and detect unusual flows was also explored.
• Hedge funds and asset managers: Numerous hedge funds have used Palantir (especially the former Metropolis platform) to gain an information edge in the markets. For example, funds could combine real-time market data with alternative datasets (e.g. social media, weather data) in Palantir to feed quantitative strategies. Steven Cohen’s investment firm Point72 was a Palantir customer for its analyses, as were some London hedge funds, according to reports. In private equity, Palantir Foundry has been used for acquisition due diligence, making it easier to scan large volumes of information on target companies.
• Insurance and fintech: Global insurers such as AXA and Sompo Japan have tested Palantir Foundry to model risk, combining claims history with geospatial and demographic data. In fintech, the payments company Stripe was publicly named as a Palantir user for detecting fraud in online transactions (fitting, given the PayPal heritage). The World Bank and development institutions have also used Palantir to analyze financial risk in global projects; although they are not “commercial customers,” they fall into this category because of the financial nature of their data.

Industrial and commercial

• Airbus (aerospace): The European aircraft manufacturer is a flagship customer. In 2017 Airbus launched the Skywise platform, in collaboration with Palantir, as part of its digital transformation. Skywise, built on Palantir Foundry, centralizes design, production, flight operations and maintenance data for the entire Airbus fleet and partner airlines. Its purpose is to enable predictive maintenance (predicting component failures before they happen) and optimize flight operations. More than 140 airlines connected to Skywise, including EasyJet, AirAsia and Delta TechOps, benefiting from fewer delays and efficiency gains thanks to integrated analysis of trillions of aircraft sensor data points. Airbus credits Skywise with significant improvements in fleet availability and savings in maintenance costs. Palantir remains a strategic partner to Airbus, expanding Skywise into digital twins, IoT and advanced analytics in new projects.
• Fiat Chrysler Automobiles (automotive): The automaker (now part of Stellantis) adopted Palantir Foundry to integrate its manufacturing, quality and supply chain data globally. From 2018-2019, FCA used Foundry at various plants to detect production bottlenecks, manage parts logistics in real time and improve component traceability. Since the Stellantis merger, Palantir has reportedly continued to support Industry 4.0 use cases at the company.
• Scuderia Ferrari (motorsport): Ferrari’s Formula 1 team is another notable user. Since 2016, Ferrari has used Palantir Foundry to help its technical teams make data-driven decisions during development and on race weekends. Foundry integrates huge volumes of telemetry data from the cars, test bench results, information on every part of the car and more, allowing engineers at the Maranello factory and at the track to quickly analyze the car’s performance and reliability. This has drastically cut the time needed for calculations or adjustments: tasks that used to take minutes now take seconds, feeding into decisions during a race (engine strategies, aerodynamic adjustments, etc.). Ferrari and Palantir renewed and expanded their partnership in 2022, highlighting the benefits of real-time analytics in a sport where milliseconds count.
• BP (energy): Oil company BP was one of Palantir’s first large corporate customers, in a partnership that began in 2014. BP integrated data from hundreds of thousands of sensors on its wells, refineries and pipelines into Palantir, creating dashboards that helped engineers detect equipment anomalies and optimize crude production. BP has reportedly even taken an equity stake in Palantir, given how important the collaboration is. In 2020 BP deepened the partnership to use Palantir Foundry in its transition to clean energy, applying data analysis to optimize wind farms and EV charging networks and to reduce emissions. This partnership illustrates the value Palantir brings to the energy industry, handling large streams of industrial IoT sensor data for efficiency and safety.
• Other industrial sectors: Palantir Foundry is used in advanced manufacturing, for example at semiconductor companies to improve production yields (integrating data from fab equipment); in logistics and retail to fine-tune supply chains (Walmart used Palantir in pilot projects to optimize inventory); and in telecommunications to plan infrastructure (Palantir’s customers have included mobile operators that combine demographic and network data to decide on optimal rollouts). On the public sector side, infrastructure operators (ports, airports) use Palantir to coordinate operations. One concrete case: Airbus not only applied Palantir internally but also offers Skywise to airlines, which means third-party developers can build value-added applications on the platform. This has created an ecosystem in which Palantir is the enabler of industrial data marketplaces.

Civil government (public administration and public safety)

• Federal and local agencies in the US: Palantir has an extensive portfolio in US civil government. At the federal level, from 2009 its Gotham software was used by the Recovery Accountability and Transparency Board (RATB), an agency that oversaw post-crisis stimulus funds, to audit and track where hundreds of billions of dollars went, detecting fraud and duplication. Other confirmed agencies include the IRS (Internal Revenue Service), which in 2023 was working with Palantir on a “mega API” to search records across all its tax databases. The Department of Homeland Security (DHS) uses Foundry as an analytics platform in emergency management, having deployed it at FEMA (the Federal Emergency Management Agency) to coordinate natural disaster relief. The Department of Health and Human Services (HHS) also runs Foundry to integrate its health program data. In addition, the Social Security Administration (SSA) has explored Palantir to modernize its data systems.
• Immigration and border enforcement (ICE): One of the most controversial deployments is with US Immigration and Customs Enforcement (ICE). In 2014 ICE hired Palantir to build the Investigative Case Management (ICM) system, a platform that integrates personal data and background information on immigrants (both legal and undocumented) to support investigations by its Homeland Security Investigations division. ICM reached full operational capability in 2017, giving ICE agents access to multiple law enforcement and private databases from a Gotham interface. In 2025, under the second Trump administration, ICE expanded its relationship with Palantir with a $30M contract for a tool called “ImmigrationOS,” intended to provide “near real-time visibility” into people subject to deportation (e.g. visa overstays). This system, based on Palantir architecture developed since 2014, helps prioritize deportation targets by integrating data from multiple sources (not publicly disclosed). These collaborations with ICE have sparked protests from civil society organizations over privacy and racial profiling concerns, but they illustrate how Palantir has become a central piece of the US immigration surveillance infrastructure.
• Local police and police departments: Numerous state and municipal police forces have used Palantir to modernize their crime analysis methods. From 2011, the Los Angeles (LAPD) and New York (NYPD) police departments tested Gotham to merge data on arrests, incidents, gang networks and more, and generate crime “heat maps.” A prominent example is the New Orleans Police Department, which in 2012 deployed Palantir in a predictive policing pilot: the software analyzed crime history, suspect profiles and associations to flag likely locations or people linked to future crimes. Although the program was secret and was later criticized when it came to light (over potential bias), it was one of the first uses of big data in policing. Other cities such as Chicago, Washington D.C., Salt Lake City and Oakland, as well as regions of California, use Palantir in their fusion centers for counterterrorism and organized crime. By 2017, at least 13 police departments in the US were estimated to be running Palantir. In 2025, Wired reported that Palantir had expanded even further, backed by federal policies that encourage sharing police data.
• Multilateral bodies and NGOs: Palantir has offered its services pro bono or at low cost to humanitarian organizations. The UN World Food Programme (WFP) uses Palantir Foundry to optimize food aid distribution, integrating logistics and population data to reach crisis areas efficiently. The NGO Team Rubicon (veterans who respond to disasters) used Palantir to coordinate its deployments during natural disasters. Likewise, Polaris Project (which fights human trafficking) has used Palantir Gotham to analyze large datasets (online ads, call records) and identify sex and labor trafficking networks, working with law enforcement agencies. These cases show Palantir being applied in the civil sphere to “find hidden things” (from corruption to crime victims) through massive data analysis, which Karp has described as the essence of the company.

Overall, Palantir’s customer base has grown from “a few heavy customers” (intelligence agencies) to hundreds of organizations across many sectors. By the end of 2024, Palantir reported 711 active customers worldwide, with accelerating growth in the commercial sector (55% year over year in the US), although government contracts still accounted for ~55% of its revenue. Its use cases range from the strategic (national defense) to the everyday (optimizing inventory), but what they all share is the need to integrate disparate data sources and extract actionable information, which is Palantir’s core value proposition.

Notable real-world use cases

Below are three emblematic use cases of Palantir’s technologies in action, followed by a few other notable examples:

Palantir in the war in Ukraine (2022-2023)

Russia’s invasion of Ukraine was the first large-scale war in which data analysis and AI played a crucial role, and Palantir positioned itself at the center of that effort. From the start of the conflict in 2022, Palantir offered its technology to the Ukrainian government. CEO Alex Karp traveled to Kyiv in person in June 2022 (the first Western CEO to meet Zelensky after the invasion) to coordinate software support, and announced the opening of a Palantir office in Ukraine. Ukraine began using Palantir Gotham and intelligence tools to integrate battlefield information: enemy positions spotted by drones, satellite imagery, allied intelligence reports and more. According to Ukrainian Deputy Prime Minister Mykhailo Fedorov, Palantir’s technology allowed Ukrainian forces to “follow the evolution of the war in real time,” giving them an information edge over a larger adversary.

In 2023 this support intensified with the introduction of military AI capabilities. Palantir deployed its new AIP platform in test scenarios so that Ukrainian operators could ask an AI assistant about possible movements or tactical recommendations, always under human supervision. The Ukrainian army was also equipped with portable Skykit kits that gave it access to Palantir’s tools directly on the front line, even in areas without communications, linked to the MetaConstellation network for up-to-date satellite imagery.

Alex Karp himself has said that Palantir became a decisive factor in Ukraine’s defense. In May 2023 he stated that Palantir is “responsible for most of the targeting in Ukraine,” meaning that its software helps select and prioritize military targets. According to one report, “the company controls intelligence operations through software, ultimately the factor that tips the balance on the front line,” making it possible to locate enemy troops, alert drones and suggest attack strategies, in essence automating warfare. These statements show how deeply Palantir became embedded in the Ukrainian command system, providing an edge in coordination and precision against more analog Russian forces.

How far software reaches in a war. In May 2023 Alex Karp stated that Palantir is “responsible for most of the targeting in Ukraine”: its software helps locate enemy troops, alert drones and suggest attack strategies, always under human supervision.

Ukraine has used Palantir for specific tasks such as pinpointing the location of Russian units from thermal and radar imagery, coordinating long-range HIMARS artillery against those targets, and assessing damage within seconds after a strike by combining multiple sources. While Russia relied on traditional methods, Ukraine “outmatched Russia digitally” thanks to the “big data ammunition” supplied by companies such as Palantir. In recognition of this, President Zelensky said Palantir’s collaboration was “a positive sign that Ukraine is open for business even in the middle of a war” and key to the conflict’s “digital front.”

In short, Palantir became a key technology ally for Ukraine, showing how big data and AI can influence the outcome of modern conflicts. This use case put Palantir’s philosophy to the test under extreme conditions, validating its ability to integrate diverse sources (satellites, drones, human intelligence) and deliver actionable real-time insights that saved lives and changed the course of military operations. It also opened ethical debates about the automation of warfare and the role of private companies in conflicts, issues Karp acknowledged as complex but unavoidable in a dangerous world where “advanced software engineering is the equivalent of having tactical nuclear weapons against a conventional enemy.”

Palantir at the UK’s National Health Service (NHS)

The UK’s NHS is an emblematic case of Palantir being used in public health at scale, with successes in health emergencies but also privacy controversies. In March 2020, as the COVID-19 pandemic exploded, the NHS awarded Palantir (without a tender, given the urgency) a contract to create the COVID-19 Data Store. Using Palantir Foundry, dozens of data sources from the British health system were integrated: test results, ICU occupancy, ventilator availability, PPE inventories, outbreak locations and more. All of this was consolidated into interactive dashboards that officials and ministers used daily to make informed decisions in the pandemic response. For example, you could see in real time how many beds were still free at each hospital or how many vaccines had been distributed by region. The system, initially valued at £1 and later expanded to £23.5M, was considered very successful operationally and “critical to the vaccination and PPE program” in the UK.

However, the way the contract was awarded (as an emergency measure, without transparency) and Palantir’s track record in intelligence drew criticism from civil liberties groups. Given the possibility that Palantir might retain sensitive patient data or expand its role, there were demands that any future contract go through a public tender. The NHS promised an open tender for its next project: the Federated Data Platform (FDP), a permanent platform to unify data across the entire NHS. That tender opened in 2023 for ~£480M, and Palantir competed hard (as the incumbent supplier). This caused a stir: some organizations and members of Parliament (even within the Conservative Party) raised concerns about giving Palantir such broad access to the health data of 67 million citizens. There were allegations of a “short time window” and favoritism, since Palantir “already had its feet under the table” at the NHS.

As 2023 went on, 11 NHS trusts (hospitals) paused their use of Foundry, citing “operational issues,” which some read as caution in the face of the controversy. Meanwhile, it emerged that Palantir had hired NHS staff (such as former AI chief Indra Joshi) and had considered acquiring British companies in the sector to “buy its way in” and overcome political resistance. There were even remarks from Peter Thiel (Palantir’s co-founder) criticizing the NHS, describing Britain’s devotion to the public service as “Stockholm syndrome” and saying that “the NHS makes people sick,” statements Palantir had to distance from its official position.

Despite the controversy, in October 2023 it emerged that Palantir had won the FDP tender (for roughly £480M) for the coming years, consolidating its presence at the heart of the British health system. The FDP aims to use Foundry for everything from planning services and researching clinical efficiency to detecting fraud in the NHS. Palantir committed to hosting the data within England and maintaining strict governance standards to ease concerns. Civil society organizations, however, kept up threats of legal action unless there was transparency about which patient data is shared and how it is used.

In summary, the NHS case shows both the enormous potential of Palantir Foundry in healthcare (better emergency response, a complete view of a complex health system) and the social challenges its adoption faces (fear of surveillance, privacy of sensitive data). What nobody disputes is that the tool proved its practical value during COVID-19, but it also forced a national debate in the UK about who should handle citizens’ medical data. This balance of benefits and risks makes the NHS a textbook example of Palantir’s influence beyond the military sphere.

Airbus’s Skywise platform (industrial digital transformation)

The Airbus Skywise case exemplifies the impact of Palantir Foundry on manufacturing and air transport, enabling an entire ecosystem to collaborate through data. Airbus, one of the largest commercial aircraft manufacturers, partnered with Palantir in 2017 to create Skywise, conceived as the “digital backbone” of its operations. Airbus contributed its deep aeronautical expertise and access to data from its global fleet, while Palantir provided the data integration and analysis infrastructure through Foundry.

What does Skywise do?

Essentially, it collects and analyzes data from every stage of the aircraft life cycle: design and manufacturing (engineering data, component testing), flight operations (aircraft sensor readings on every leg, tens of thousands per flight), maintenance (reported issues, repair records) and business data (parts turnover, schedules, weather). By aggregating them, Skywise delivers insights that were previously impossible. For example, it provides predictive maintenance: using AI and predictive models, it detects subtle patterns that foreshadow failures in critical systems, so airlines can repair or replace a part before it causes an unexpected breakdown. This dramatically reduces unplanned AOG (aircraft on ground) events, saving airlines millions in delays and penalties.

Skywise also offers real-time fleet performance monitoring: an airline can see on a dashboard the fuel efficiency, environmental conditions and trends of each aircraft, and compare them. This makes it possible to identify less efficient aircraft or routes and correct them (for example, by adjusting flight altitudes or fleet rotation). Another use is optimizing maintenance management: Skywise centralizes all technical records and makes it easier for engineers to spot recurring problems and plan combined maintenance for multiple aircraft, minimizing time out of service.

Industry impact

Skywise was offered as a platform to Airbus’s airline customers, and by 2025 more than 50% of the global Airbus fleet (some 10,000 aircraft) was connected to it. More than 140 operators use it, from major carriers (Delta, China Southern) to low-cost airlines (EasyJet, AirAsia). This has led to concrete improvements: EasyJet reported fewer delays and better dispatch reliability thanks to predictive maintenance through Skywise. Delta TechOps (Delta Air Lines’ maintenance arm) integrated Skywise into its technical processes, increasing aircraft availability and detecting reliability problems faster. Airbus itself reaps huge benefits: it feeds real-world aircraft usage data back into its designs to make them more robust, and identifies which parts fail most often so it can redesign them. Skywise also strengthens the relationship between Airbus and airlines by offering ongoing added value (beyond selling the aircraft, it provides an operational intelligence service).

Skywise also fosters unprecedented collaboration: component manufacturers (engines, landing gear) can access, with permission, relevant data on their equipment in service, and airlines can benchmark themselves (anonymously) against industry averages, all under a secure sharing architecture that preserves each participant’s privacy. Palantir Foundry makes this granular permissioning and the continuous ingestion of data from many companies at once possible, something very hard to build from scratch. In fact, Airbus lets third-party developers build applications on Skywise (for example, spare parts management apps for airlines), creating an innovation ecosystem.

All in all, Skywise illustrates the transformative potential of Palantir Foundry in Industry 4.0: by integrating previously siloed data (manufacturer ↔ airline ↔ supplier), it delivers efficiency, safety and cost improvements on a global scale. Airbus has called Skywise the “central nervous system” of its data ecosystem, and it is seen as key to future challenges such as adding component digital twins, real-time IoT and autonomous maintenance. For Palantir, this use case showed that its tools are not just for espionage or banking but can drive the digitalization of traditional sectors such as aviation, delivering tens of millions in operational value to their users.

Other notable cases

• COVID-19 vaccine distribution in the US: Palantir Foundry was used by the US government in the Tiberius initiative to coordinate the complex logistics of COVID vaccines in 2020-2021. The platform tracked the manufacturing, allocation and delivery of hundreds of millions of doses to every state, integrating data from suppliers, states and pharmacies, which made it possible to identify bottlenecks and make the process more transparent. This contributed to a more efficient vaccination campaign in a country of 330 million people.
• Natural disaster response: Organizations such as Team Rubicon (an NGO of veterans) and agencies such as FEMA have used Palantir to coordinate relief efforts. After hurricanes such as Harvey (2017) and Ida (2021), Foundry was used to combine data on damage, victims’ locations, available resources and volunteers, optimizing the allocation of aid in the critical first hours. The British Civil Defence Force has also tested Palantir for domestic crisis management.
• Fighting financial crime and corruption: The World Bank used Palantir to analyze patterns in infrastructure projects and detect possible cases of corruption or bid rigging, combining financial data with corporate networks. Similarly, the Australian Federal Police integrated bank records with police data in Palantir to dismantle a drug-linked money laundering network in 2015, uncovering connections between seemingly unrelated accounts.
• Retail supply chain optimization: Large retailers have used Palantir Foundry to fine-tune inventory and logistics. For example, Home Depot (the US home improvement chain) used Palantir in 2016-2018 to forecast product demand using local weather and socioeconomic data, avoiding in-store stockouts and reducing excess inventory, with significant savings. Walmart likewise ran pilots to analyze cash flows in stores and prevent losses by combining camera, sales and staffing data.

These additional examples reaffirm Palantir’s versatility: from speeding up the delivery of a lifesaving vaccine, to finding needles in financial haystacks to stop white-collar crime, to making our everyday shopping more efficient. In every case, the ability to unify disparate data and quickly extract actionable knowledge is the differentiating value Palantir brought, often becoming critical infrastructure wherever it is deployed.

Palantir versus the competition: a comparative analysis

Palantir operates at the intersection of data platforms, advanced analytics and artificial intelligence, a highly competitive space. Its rivals range from specialized startups to established tech giants. Below, Palantir is compared with three of its main competitors (Snowflake, Databricks and IBM), highlighting relative strengths and weaknesses in features, architecture, scalability, cost and sectors served. Other relevant companies in the competitive landscape are also mentioned.

Palantir vs. Snowflake

Snowflake is known for its data cloud platform, focused on storing and analyzing structured data at scale. It started as a highly scalable cloud data warehouse, giving companies a simple way to consolidate data and run fast SQL queries. In terms of functionality, Snowflake specializes in being an efficient data repository, with separate compute and storage that lets it grow almost without limit while keeping query performance high. Palantir, for its part, offers a broader stack: besides storing and integrating data (although it relies on underlying databases and is not a pure data warehouse), it provides numerous integrated tools to turn that data into operational applications, workflows and actionable intelligence. In other words, Palantir Foundry is more end-to-end (from ETL to the final application), while Snowflake integrates with other tools (external BI and ETL) to deliver value.

In terms of architecture, Snowflake was built cloud-native on AWS, Azure and GCP infrastructure, and is offered only as multi-tenant SaaS. Palantir initially deployed on-premises in government environments and later adopted a hybrid, flexible model: Foundry can run in the cloud or on the customer’s private infrastructure, even in air-gapped environments with no internet (thanks to Apollo). This gives Palantir an advantage in highly regulated sectors (defense, government healthcare) where Snowflake, being public-cloud only, cannot always be used. However, Snowflake is usually easier to implement at typical enterprises: its uniform architecture on the cloud of choice and familiar SQL make adoption easier for traditional data teams.

On scalability, both platforms shine, but in different ways. Snowflake has shown it can easily handle petabytes of data and thousands of concurrent users running analytical queries, scaling out on demand (customers pay for compute as they use it) with little intervention. Palantir Foundry has also been proven at large scale (e.g. Ferrari F1’s trillions of data points, or data for half an army), but its scalability usually requires careful design of the ontology and pipelines tailored to each customer. Foundry can operate at enterprise level with guaranteed performance, but it requires customization and tuning by engineers (often Palantir’s own), especially at the start, while Snowflake scales on a more self-service basis.

When it comes to cost, they differ significantly. Snowflake uses a consumption-based model: customers pay for the compute and storage capacity they use (per second of compute and per TB stored), which can be an advantage for starting small and scaling as needed. Palantir has traditionally meant high upfront fixed costs (multimillion-dollar contracts that include licenses and implementation consulting), although the platform then generates operational value. Put another way, Palantir requires a larger initial investment but is designed for deep transformations, while Snowflake lets you start small and grow. That is why Snowflake has thousands of customers (more than 8,000), many of them mid-sized companies, while Palantir opted for a few large customers (711 by 2024). Even so, with its Foundry for Builders offering and its growing commercial base, Palantir is trying to lower barriers to entry (it even offers startup packages).

On analytics and AI capabilities, Palantir has built machine learning tools, semantic ontologies and now generative AI (AIP) more natively into its platform. Snowflake initially focused on structured data and only recently added integrations for Python, Snowpark and the like, with which it hopes to attract data science workloads. Even so, Palantir has the edge in integrating AI directly into operational workflows: for example, a model trained in Foundry can easily be deployed as an application for an end user, and with AIP you can now build chatbots on top of private data. Snowflake usually requires moving data out to notebooks or external tools for advanced ML. In fact, Forrester ranked Palantir as a leader in AI/ML platforms, ahead of cloud competitors, reflecting its strength in this area.

As for sectors, Snowflake is widespread in financial services, retail and technology, among companies that mainly need to run data warehouses for BI. Palantir, by contrast, dominates in government, defense and heavy industry, where its granular security controls and its ability to customize around complex data are valued. In sectors such as healthcare, the two compete head to head: Palantir with an end-to-end solution approach (e.g. the NHS case), Snowflake promoting its Data Share for exchanging clinical data between institutions. One final difference is culture and approach: Snowflake sells to data teams (CTOs, data engineers) on technical power, while Palantir targets senior executives with the promise of solving hard business problems (its demos talk about “solving operational problem X” rather than technical metrics). This is why Palantir is sometimes seen more as tech consulting + software, while Snowflake is clearly a pure tech product.

Palantir vs. Databricks

Databricks is another prominent company that, like Palantir, aims to help organizations make the most of big data and AI, but with a different approach. Databricks grew out of the Apache Spark ecosystem and is optimized for processing large volumes of data and for collaborative machine learning in the cloud. It offers a unified platform called the Lakehouse, which combines data lakes with data warehouse features and is strongly geared toward data scientists and data engineers.

In terms of design and target audience, Databricks was designed by and for technical profiles, while Palantir Foundry was conceived with a more all-encompassing philosophy geared toward the business end user. Foundry provides interfaces for multiple roles (from experts who can code in Python/R against the data, to analysts using a point-and-click UI, to operators who consume apps without ever seeing raw data). Databricks historically offered mainly shared notebooks, required writing code (PySpark, Scala, SQL), and its strength lay in letting data science teams train ML models at scale. Although Databricks has added dashboards and no-code features, it still excels in environments with strong data engineering teams. Palantir, by contrast, aims to “elevate” business users by giving them power without needing to code, wrapping the complexity in easy-to-use applications.

From a functionality standpoint, Databricks stands out in massive-scale processing and ML. Its optimized Spark engine can churn through terabytes of data to train AI algorithms or run complex aggregations very quickly. Palantir Foundry can also handle big data, but it is better known for managing heterogeneous data with heavy governance: it integrates relational databases, documents, real-time streams and more, with a semantic model (ontology) that preserves the context of the data and very fine-grained access control. A bank, for example, might prefer Databricks so that its 50 data scientists can experiment with risk models in notebooks; but to then roll out an application built on those models to 5,000 loan officers, it would need to build a lot around Databricks, whereas Foundry has it built in (out-of-the-box operational applications).

On architecture, Databricks is mostly cloud SaaS (Azure, AWS, GCP), although it allows some flexibility in where storage lives (the customer’s data lake). Palantir, as mentioned, runs in the cloud, on-premises and at the edge. Because it derives from open-source Spark, Databricks fits well into an open ecosystem: it connects easily with Kafka, Delta Lake, MLflow and others. Palantir’s stack is more proprietary: although it has APIs, most of its components are closed and developed in-house (which sometimes means greater stability but less familiarity for the outside technical community). That is why the open-source and developer community revolves more around Databricks; Palantir has a smaller community, centered on its professional services.

On scalability and performance, both are strong but in different scenarios: Databricks excels at compute-intensive workloads (e.g. training a deep learning model on GPUs over a gigantic dataset, or running a SQL query over trillions of rows; it has set performance records in TPC benchmarks). Palantir can handle large workloads too, but in settings with a great variety of data and concurrent users doing different things. The Foundry platform has been proven in enterprise environments with thousands of simultaneous users running smaller but numerous operational queries, while keeping the experience consistent. One thing about Databricks is that it is notebook-centric: ideal for exploration, but low-latency operational use cases sometimes require exporting models to another application. Palantir makes it easy to package analytics and serve them in production applications within the same platform.

On cost, Databricks also has a pay-as-you-go model (based on the “Databricks Units” of compute consumed), similar to Snowflake’s. This can be economical for occasional development work, but under heavy production use, costs can climb if workloads are not optimized (much like heavy use of cloud computing). Palantir, with its high fixed cost, is expensive at the outset but then does not penalize heavy use; in fact, it encourages using it as widely as possible once purchased. For companies with an open-source culture, Databricks may be more attractive because it avoids vendor lock-in, while Palantir, by providing a more finished solution, justifies its cost through the outcomes it delivers (lower spending on inefficiencies, etc.).

On sectors: Databricks has its base in technology, internet and finance (companies that already had big data teams) and has been moving into manufacturing and healthcare as those industries hire more data scientists. Palantir, again, came in through government and industry, where in-house technical talent was sometimes lacking and organizations needed a partner to deliver the complete solution. Still, the two are converging: Palantir is adding more technical, customizable capabilities (Jupyter notebook support, integration with developer tools) to keep engineers happy, while Databricks is adding more BI features and applications to reach business users. Both are also pushing hard into generative AI: Databricks launched Dolly (an open-source LLM) and its Lakehouse AI to let customers fine-tune models on their own data; Palantir, with AIP, offers a more ready-to-use, agent-based approach.

In short, Palantir vs. Databricks can be seen as “end-to-end operational platform vs. flexible data platform.” Palantir excels when an organization needs an immediate solution to connect complex data and hand it on a plate to non-technical users, under strict governance (which is why it dominates in defense and the public sector). Databricks shines when an organization wants to get the most out of its data experts to innovate with custom models and analysis, in a freer, more programmable environment (which is why it is popular in tech and among data scientists). Many large companies even choose to use both: Palantir for certain core operational applications and Databricks for exploration and advanced research, later bringing the models into Foundry for deployment.

(Comparison in brief: Palantir Foundry focuses end to end on operational workflows, serving both technical and non-technical users; Databricks focuses on data science and massive-scale processing, serving mainly technical profiles.)

Palantir vs. IBM

IBM is an enterprise software veteran with multiple products that partly overlap with Palantir. Although IBM does not have a single platform identical to Palantir’s, it competes in key areas: data analytics, artificial intelligence and solutions for government and industry.

In intelligence and security, Palantir historically competed with IBM i2 Analyst’s Notebook, a visual network analysis tool that IBM acquired in 2011 and that was widely used by police and intelligence agencies. i2 lets users chart links between people, events and objects (a function similar to Gotham’s), but it is more manual and case-by-case, while Palantir Gotham is a broader platform for automated data mining and collaboration. Agencies that modernized their systems frequently replaced i2 (or complemented it) with Palantir to handle larger volumes of data in real time. Even so, IBM retains a presence thanks to decades-long relationships with governments and a broader portfolio (hardware, mainframes, etc.).

In AI and data platforms, IBM offers its IBM Watson suite (recently evolved into watsonx). Watson first became famous for winning Jeopardy! (2011) and was later steered toward cognitive AI for businesses (e.g. Watson Health). IBM has positioned watsonx as a platform for companies to train and deploy their own AI models, with data management and governance tools. Here the competition with Palantir AIP/Foundry is direct: both promise to bring AI into business applications in a governed way. IBM has the advantage of a vast ecosystem (consulting, integrations with enterprise legacy systems, etc.) and is usually the “familiar elephant” for IT departments. Palantir, on the other hand, is seen as more innovative and agile in rapidly deploying operational AI solutions. In fact, IBM and Palantir decided to collaborate: in 2021 they announced a partnership under which IBM would integrate Palantir Foundry with its Cloud Pak for Data offering and resell it to its customers, combining Watson’s AI with Palantir’s core. This suggests IBM recognized Palantir’s strength as the “glue” that binds data and operations, while Palantir leveraged IBM’s sales force and the trust it enjoys among traditional corporations.

On scalability and architecture, IBM has traditionally provided both on-premises software (e.g. DB2, Cognos, SPSS) and cloud offerings. Its hybrid cloud (Red Hat OpenShift) competes with Palantir’s flexibility to deploy anywhere. IBM has experience running gigantic systems (global banks, entire governments running on its systems) and a strong focus on open standards and enterprise-grade compliance. Palantir also operates at large scale, but IBM can point to its track record and global support ecosystem as an advantage on long-term projects.

In terms of analytics features, IBM offers modular solutions: for example, IBM Watson Studio for data science and IBM Cognos Analytics for BI, which together could cover part of what Foundry does. Palantir’s native integration and unified experience, however, are usually superior. One user comparison report noted that on ease of data connection, IBM Watson Studio scores 9.3/10 vs. 6.7/10 for Palantir Foundry, possibly because IBM has a multitude of connectors to legacy systems. But on overall satisfaction, Palantir tends to score slightly better than IBM in user reviews (e.g. 4.6/5 vs. 4.3/5 on Gartner Peer Insights).

On cost and business model, IBM usually involves heavy consulting services (often from its IBM Services division) to deliver custom projects, which resembles Palantir in terms of hands-on involvement. However, IBM bids on many public contracts with all-in hardware-software-consulting offers where Palantir does not compete (e.g. a complete public safety system with cameras, servers and i2, which IBM can supply in full). Palantir, for its part, focuses on software and works with integrators for the rest. As for sectors, IBM is present in virtually all of them (from banks running mainframes to supermarkets on IBM systems), but in recent years it has lost ground in innovative sectors. Palantir positions itself in growth sectors (defense AI, Industry 4.0, etc.) where IBM is sometimes seen as an “old guard” vendor.

A notable advantage Palantir has over IBM is its speed of delivery and evolution. Palantir ships features constantly (e.g. AIP was conceived and launched within months of ChatGPT), while IBM, being so large, has slower cycles. On the other hand, IBM has a focus on open, scalable solutions with long-term support, which gives some customers confidence.

In conclusion, Palantir and IBM compete but also complement each other: Palantir has the innovation and the specialized focus, IBM has the breadth of portfolio and the long-standing relationships. One investor summed it up by saying you would choose Palantir over IBM for its specialized focus, strong recent financial performance and competitive edge in next-generation data/AI platforms. Even so, IBM remains a formidable competitor in any large tender where it can include its software, and with watsonx and other tools it will try not to give up ground in the emerging enterprise AI revolution.

Other competitors and the overall landscape

Besides those mentioned, Palantir’s competitive ecosystem includes:

• Big Tech (Google, Microsoft, Amazon): Each cloud provider offers its own analytics suite: Google Cloud with BigQuery, Looker and more; Microsoft with Azure Synapse Analytics, Power BI and now Microsoft Fabric (a unified data platform). Microsoft Fabric in particular takes a similar approach to unifying data workflows, although it builds on tools people already know (Excel, Power BI). Palantir set itself apart by being cloud-agnostic and focusing on specific use cases, but it competes with these mega-platforms in almost every commercial account. One factor is that Palantir is often deployed on top of cloud infrastructure (e.g. Foundry running on AWS), so it sometimes cooperates with Big Tech rather than competing head-on. Even so, Google and Microsoft have rival AI offerings (Vertex AI, Azure ML) that fight for the same budget as AIP.
• Traditional analytics software vendors: This group includes SAS (with its long-established statistical analysis suite), Oracle (with Oracle Analytics Cloud and robust databases) and SAP (with solutions such as Data Warehouse Cloud and analytics built into its ERP). Many corporations already have these systems; Palantir has to convince them to add its platform or switch to it. SAS, for example, is ubiquitous in banking and public health, but Palantir can beat it on flexibility and modernity (SAS requires coding in its own proprietary language).
• Startups and specialists: Competitors are emerging that focus on niches similar to Palantir’s. One is C3.ai, which offers an enterprise AI application platform (also heavily focused on industry, energy and defense). C3.ai competes by claiming it is easier to build specific AI applications on its platform, but it has gained less traction than Palantir in large contracts. Another is DataWalk, which pitches itself as a cheaper alternative to Palantir Gotham for network analysis in policing and intelligence. Although its capabilities are more limited, DataWalk has won some mid-sized government customers by playing the lower-cost card. There are also combined open-source solutions (Elasticsearch + Kibana, for example) that some budget-constrained agencies use for certain search functions Palantir would handle, although they do not offer full integration.
• Separate business intelligence and big data tools: Sometimes Palantir competes against a set of tools rather than a single vendor. For example, a company might combine Alteryx (for data preparation), Tableau (BI visualization), a cloud database and perhaps custom development for its operational app. That “homegrown solution” may be cheaper than Palantir at first, but it requires more integration effort from the customer and may not reach the sophistication of Foundry. Palantir argues that its platform reduces the complexity of juggling many separate pieces and shortens time to value.

Palantir’s relative strengths: Superior integration of heterogeneous data (it can ingest almost any format and relate it within its ontology), strong granular security and access control (inherited from intelligence-community requirements), a focus on operational solutions (not just analysis, but turning analytics into action through apps), and specific preconfigurations for complex industries (e.g. ready-made modules for healthcare, defense and manufacturing). On top of that, its commitment to “solving hard problems” has given it the mystique of a powerful product, backed by success stories in sensitive areas (military, espionage, pandemics) where few competitors can claim experience.

Weaknesses and challenges: High cost and the perception of a proprietary “black box” (some engineers find it too closed and hard to adapt beyond its intended use). Its learning curve can also be steep: although it offers user-friendly interfaces, mastering Foundry in full requires training, and organizations can become dependent on Palantir for support and development. In contrast, solutions such as Snowflake or Databricks draw on large communities and standard skills (SQL, Python), reducing lock-in. Palantir has also polarized opinion with its controversial government contracts; companies whose staff or customers are sensitive to these issues may hesitate to partner with a firm some have nicknamed the “tech arm of surveillance,” although Palantir has tried to improve its image on transparency.

Ultimately, Palantir’s competitive landscape reflects a balance: it is unique in its end-to-end value proposition, which has won it major contracts, but on every specific front it faces very strong specialized rivals, from the best analytical database and the best data science environment to IT giants with trusted relationships. Palantir’s strategy seems to be to stay at the cutting edge by integrating the latest technologies (such as generative AI), show quick results on critical projects, and thereby justify its price premium over fragmented solutions. Its growing recognition (for example, being named a leader in Forrester’s AI platforms evaluation) suggests it has managed to differentiate itself, but it will have to keep adapting, because in the fast-moving big data and artificial intelligence market, today’s advantage can vanish tomorrow if competitors innovate faster or customers opt for more open approaches.

Sources: The claims and data provided are based on press reports, official documentation and industry analysis, including articles from Wired, Bloomberg, Reuters, Business Insider and Wikipedia, comparative technical reports, and official statements from Palantir and its customers, among others. Each citation in the text refers to the specific origin of the information presented, allowing the content to be verified transparently. All the sources used are well-regarded or directly linked to the entities mentioned, ensuring the reliability of the data and facts included in this report.

Related service

Offensive cybersecurity services

Topics

#Data #Future #PLTR #Palantir

Keep reading

• Hardware Analysis Series – Part 5: Electromagnetic Espionage – The Silent Attack
• NIS2 in Spain, two years without a law: where transposition stands and what to do this quarter
• European compliance calendar 2026-2030: CRA, AI Act, Data Act and everything still to come
• How a news story travels: anatomy of fourteen days of press
• When an executive buys their own shares

Book a call