Palantir Gotham vs Foundry vs Apollo vs AIP: What Each Platform Does and Which One You Need

Last updated: October 9, 2026

By Omar J, Director, PAL4C Labs

TL;DR Palantir sells four platforms that stack in layers. Apollo deploys and updates everything underneath. Foundry is the data operations platform where the Ontology lives, and Gotham is the defense and intelligence platform. AIP adds generative AI across both. Commercial companies almost always need Foundry plus AIP, with Apollo running underneath. Our Palantir Foundry and AIP implementation team starts by mapping your use case to that stack.

Ask five people what Palantir sells and you’ll get five answers. A spy software company. An AI company. A data platform. A defense contractor. A consultancy, which it isn’t.

They’re all partly right, because Palantir runs four platforms that the press regularly mixes up, each with very different buyers. The confusion matters when you’re the one writing the business case. Pitch Gotham to a manufacturer’s CFO and you’ll lose the room. Assume Apollo is an add-on line item and you’ll misread the proposal.

This guide sorts it out using Palantir’s own descriptions from its SEC filings and documentation, not marketing pages or investor folklore. It covers what each platform does, how they fit together, where programs like Maven and TITAN sit, what the 2026 revenue numbers say about who buys what, and which combination a commercial team actually needs.

How Do Palantir’s Four Platforms Fit Together?

Think three layers. Apollo is the delivery layer that deploys and updates software. Foundry and Gotham are the operating platforms where data and decisions live. AIP is the AI layer that runs across them.

Palantir’s architecture documentation describes Apollo as the continuous delivery platform that manages the infrastructure hosting Foundry and AIP, and calls the integrated AIP, Foundry, and Apollo stack an enterprise operating system. Gotham sits beside Foundry rather than inside it, built for defense and intelligence missions.

Simple enough on paper. The trouble starts with naming. Press coverage regularly says Palantir when it means Gotham, or AIP when it means a Foundry workflow with an LLM step. Ask which platform before you believe a headline.

What Is Palantir Gotham?

Gotham is Palantir’s platform for defense and intelligence work. It helps analysts find patterns across sensitive data sources and supports missions across allied defense and intelligence operations. It’s sold mainly to government agencies.

Palantir’s FY2024 annual report says Gotham lets users identify patterns hidden deep within datasets, “ranging from signals intelligence sources to reports from confidential informants.” The Q2 2026 quarterly report adds that it integrates with Palantir’s other platforms and broader defense offerings “to power a wide array of missions across allied defense and intelligence operations.”

Read that source list again. Signals intelligence. Confidential informants. That’s not a supply chain tool with a scary name.

Gotham was Palantir’s first product, built for the intelligence community, and it’s where the company’s reputation (good and bad) comes from. You’ll see it described as geospatial software in some explainers. Maps are part of it. They’re not the point. The point is connecting people, places, events, and signals so an analyst can hand something actionable to an operator.

Can a private company buy Gotham? Palantir’s filings describe it in government terms and don’t spell out commercial availability, so we won’t pretend to know Palantir’s sales policy. In practice, commercial teams get pointed to Foundry.

What Is Palantir Foundry?

Foundry is Palantir’s data operations platform. It integrates data from source systems, models it in the Ontology, and runs analytics and operational workflows on top. Every Palantir commercial customer uses it, and so do several government customers.

The 10-Q calls Foundry the “foundational data operations platform,” providing data management, logic authoring, systemic mapping through the Ontology, analytics, and workflow development. The FY2024 10-K is blunter. “All of our commercial customers now use it, as do several of our government customers.”

That second line is the most useful sentence in this whole post for a commercial buyer. If you’re a manufacturer, a hospital, a bank, or a logistics operator, your Palantir program is a Foundry program.

What makes Foundry different from a data warehouse is the Ontology. Instead of tables, teams work with objects like Shipment, Supplier, Patient Encounter, or Aircraft, linked to each other, with actions that can write decisions back to source systems. Our Palantir Ontology explainer covers how that works in practice, and it’s where most of the real implementation effort goes.

One proof point from our own work. For a global logistics operator, we connected 18 disconnected warehouse and inventory systems into a single Foundry Ontology, then added AIP forecasting agents. Procurement delays fell 34%, and the client saves $12M a year, per our case studies page. No Gotham anywhere in that stack. Wasn’t needed.

What Is Palantir Apollo, and Do You Buy It Separately?

Apollo is Palantir’s continuous delivery platform. It deploys and upgrades Palantir software and manages the infrastructure underneath it, across cloud, on-premises, and air-gapped environments. Commercial buyers usually get it as part of the platform rather than as a separate purchase.

The 10-Q describes Apollo as enabling “the orchestration of upgrades of services and assets every day to manage the underlying infrastructure that hosts our other platforms.” Palantir’s Apollo documentation says it supports all major cloud, on-premises, and disconnected (air-gapped) environments.

Why does a delivery tool get platform status? Because shipping software to a submarine, a classified network, or a factory with no internet is genuinely hard. Apollo is how Palantir keeps those deployments current without someone flying out with a USB stick. Unglamorous. Essential.

Here’s the buyer’s view. If you run Foundry in Palantir’s cloud, Apollo is doing its job in the background and you’ll rarely think about it. It becomes a visible conversation when you need Palantir running inside your own environment, at the edge, or air-gapped. Some organizations also use Apollo to deliver their own software, but that’s a separate decision with its own scoping.

Where Does AIP Fit?

AIP is Palantir’s generative AI platform. It connects third-party LLMs to your data and operations, provides tools for building agents and automations, and includes evaluation tooling. It runs across Foundry and Gotham rather than replacing either.

The 10-Q describes AIP as providing “secure connectivity to third-party-provided large language models,” a development toolchain for AI agents and automations, AI-enabled applications, and “a broad evaluations framework for governing AI workflows in production.” The FY2024 10-K says AIP is bundled with existing offerings such as Foundry, Gotham, and Apollo.

So AIP isn’t a fifth silo. It’s the reasoning layer that sits on the Ontology. An AIP agent is only as useful as the objects and actions it can reach, which is why we tell clients that AIP without a decent Ontology is a chatbot with expensive taste.

Our guide to what Palantir AIP is covers the components. If you’re building, our walkthrough of Palantir AIP agents covers Logic, Chatbot Studio, tools, and evals.

Gotham vs Foundry vs Apollo vs AIP, Side by Side

Read the Typical buyer row first. It answers most questions before they’re asked.

Where Do Maven, TITAN, and Warp Speed Fit?

They’re products and programs built on Palantir technology, not additional core platforms. Maven Smart System is a Pentagon AI program, TITAN is an Army ground station prototype, and Warp Speed is a manufacturing offering for defense suppliers.

These names show up in headlines more than the platforms do, so it’s worth placing them.

• Maven Smart System. Defense News reported a $480 million, five-year contract in May 2024 to expand access across combatant commands including CENTCOM, EUCOM, and INDOPACOM. In March 2026, Reuters reported a Pentagon memo making Maven an official program of record, with oversight moving to the Chief Digital and AI Office. NATO also acquired a version for Allied Command Operations in 2025.
• TITAN. Short for Tactical Intelligence Targeting Access Node. The U.S. Army awarded Palantir $178.4 million in March 2024 to build 10 prototypes of its next-generation intelligence ground station.
• Warp Speed. Palantir announced in December 2024 as a manufacturing operating system for defense production, with Anduril, L3Harris, Panasonic Energy of North America, and Shield AI in the first cohort.
• Army enterprise agreement. The Army’s enterprise agreement, worth up to $10 billion over 10 years, consolidates 75 existing contracts into one vehicle. That’s a ceiling, not committed spend.

Is Maven part of Gotham? Public sources don’t say precisely how Maven’s components map to Palantir’s platforms, so we’re not going to guess. Treat it as a program delivered with Palantir software.

Commercial buyers can mostly skip this section. Mostly. It’s useful context when your board asks why the company’s name keeps showing up next to the Pentagon.

What Do the 2026 Numbers Say About Who Buys What?

Palantir’s revenue is now split close to evenly. In Q2 2026, the government segment brought in $990 million and the commercial segment $945 million, with U.S. commercial revenue growing 149% year over year.

Those figures come from the Q2 2026 10-Q and Palantir’s Q2 2026 results release, which reported total revenue of $1.935 billion, up 93%. U.S. government revenue grew 90% to $809 million. U.S. commercial grew 149% to $764 million.

Not long ago, government was clearly the bigger side. In FY2024, 55% of revenue came from government and 45% from commercial, per the 10-K. Commercial has closed most of that gap in under two years.

What does that mean for you? Since every commercial customer runs Foundry, commercial growth is mostly Foundry and AIP growth. More buyers, more competition for experienced builders, and a lot more first implementations happening at the same time. Plan for staffing early.

Which Palantir Platform Do You Actually Need?

For almost any commercial company, Foundry plus AIP, with Apollo running underneath. Gotham fits defense, intelligence, and some national security agencies. Apollo becomes its own conversation only for on-premises, edge, or air-gapped deployments.

Bias disclosed. We only do Palantir work, specifically Foundry and AIP, so we’d naturally lean that way. But we’d also tell a commercial team with one data warehouse and a reporting problem that it probably doesn’t need Palantir at all. Our Palantir Foundry pricing guide explains what the platform costs and when it’s worth it.

Start With the Use Case, Not the Product Name

Three takeaways. Apollo delivers, Foundry and Gotham operate, and AIP reasons on top. Every commercial Palantir customer runs Foundry. And the platform question is easy once you know what decision you’re trying to change.

PAL4C Labs is a Palantir-exclusive consultancy that implements Foundry and AIP for commercial operations, supply chain, healthcare, and regulatory teams. We build the pipelines, the Ontology, and the AIP agents, and we measure the work by the decision it moves.

If you’re trying to figure out which Palantir stack fits your organization, bring us the decision you want to change and the systems behind it. We’ll map it to the right platforms with you and tell you plainly if Palantir isn’t the answer.

Platform Questions We Hear

Gotham vs Foundry, What’s the Real Difference?

Can a Private Company Buy Palantir Gotham?

So What Is Foundry Actually Used For?

Does Foundry Run on Apollo?

Does AIP Need Foundry?

Is the Maven Smart System Part of Gotham?