Palantir

Magic AI

Palantir

Palantir's business model is enterprise software — specifically, large multi-year contracts with governments and corporations. Contracts typically start at $1-5 million and can scale to hundreds of millions annually for large government agencies.

The sales process is uniquely intensive. Palantir deploys "forward-deployed engineers" (FDEs) who embed directly with customers for months, configuring the platform for specific use cases.

This hands-on approach is expensive but creates deep integration that makes switching nearly impossible. Once Palantir is embedded in an organization's workflows, it's practically permanent.

Revenue split has shifted over time. Government contracts (US and allied nations) historically dominated, but commercial revenue has been growing faster.

By 2024, commercial revenue approached 45% of total. Annual revenue exceeded $2.8 billion.

The company has been profitable since 2023.

Magic AI

Magic is building an AI that works like a software engineer, not a chatbot. The goal is a model you can hand a whole code base and a task, and it writes, reviews, and debugs like a teammate.

Their edge is memory. In 2024 they showed off LTM-2-mini, a model with a 100 million token context window.

To put that in perspective, that is about 10 million lines of code the model can look at in one shot, by far the biggest of any model at the time. The plan is to sell this as an AI coworker to companies that write a lot of software.

As of the 2024 rounds, the product was still not for sale.

Palantir

Palantir was born from the aftermath of September 11, 2001. Peter Thiel — PayPal co-founder and contrarian investor — realized that the same fraud-detection algorithms PayPal used to catch financial criminals could help intelligence agencies catch terrorists.

The US government had mountains of data but terrible tools for connecting the dots.

Thiel co-founded Palantir in 2003 with Alex Karp (a Stanford law graduate who then took a PhD in social theory at Goethe University in Frankfurt, studying under Jürgen Habermas), Joe Lonsdale (a Stanford student who'd worked at Clarium Capital), Stephen Cohen (an engineer), and Nathan Gettings (a Clarium colleague). They named it after the palantíri in Tolkien's Lord of the Rings — the seeing stones that let you view distant events.

The CIA's venture arm, In-Q-Tel, was the first investor and first customer simultaneously. The initial product, Palantir Gotham, was built specifically for intelligence analysts who needed to find connections across massive, messy datasets — linking phone records, financial transactions, travel data, and classified intelligence into a single coherent picture.

The company operated in extreme secrecy for its first decade, with most employees unable to discuss what they actually built.

Magic AI

Eric Steinberger and Sebastian De Ro started Magic in 2022. Steinberger is an AI researcher who became convinced that the missing piece for a true AI software engineer was memory.

Chatbots forget what you told them a few thousand words ago. He wanted a model that could read your entire code base and remember all of it.

They raised a small seed in 2022, then a $23 million Series A in early 2023 led by Alphabet's growth fund CapitalG. From day one the pitch was not a chatbot.

It was an AI colleague that could handle real engineering work.

Palantir

Palantir's growth strategy for two decades was simple: get inside the US government, prove indispensable, and expand from there. CIA led to NSA.

NSA led to the Army. The Army led to the Air Force.

Each agency saw what the others were doing and wanted it.

The AIP launch in 2023 was the commercial growth inflection point. By integrating large language models into the platform, Palantir made its data analytics accessible to non-technical users.

A supply chain manager could ask questions in plain English and get answers from their data. This dramatically expanded the potential user base within existing customers and attracted new commercial clients.

"Boot camps" became the commercial go-to-market innovation. Palantir runs intensive multi-day workshops where potential customers bring their actual data and problems, and Palantir engineers build working prototypes on the spot.

Companies leave with tangible proof of value, which accelerates the sales cycle dramatically.

Magic AI

Magic's strategy is research first, product later, which is both unusual and expensive. Instead of shipping an app and iterating in public, they raised huge sums to chase a technical breakthrough in model memory.

The headline move was the 100 million token context claim, which got the whole AI world talking. They also cut a deal with Google Cloud in 2024 to build two supercomputers, called Magic-G4 and Magic-G5, using NVIDIA chips, to train bigger models.

Star backers like Eric Schmidt served as validation. The bet is that if the tech is far enough ahead, the product and the customers follow.

Palantir

The ethical debate follows Palantir everywhere. Privacy advocates have criticized Palantir's work with ICE (Immigration and Customs Enforcement), police departments, and intelligence agencies.

The company has been accused of enabling mass surveillance. Karp has been unapologetic — arguing that democracies need powerful analytical tools and it's better that a company with ethical guidelines builds them than the alternative.

Customer concentration was a historical risk. For years, a handful of massive government contracts drove the majority of revenue.

Losing a single contract could crater a quarter. The push into commercial has diversified the revenue base, but government still represents over 55% of revenue.

Valuation has been the market debate. Palantir trades at astronomical revenue multiples (60-80x revenue at its 2024 peaks), which assumes massive future growth that may or may not materialize.

Bears argue it's the most overvalued stock in tech. Bulls argue that AIP will drive exponential commercial growth.

The debate is loud and ongoing.

Magic AI

Magic has raised close to half a billion dollars and, years in, still has no product the public can buy. That is the whole problem in one sentence.

A giant context window is a great demo, but rivals like Cursor, GitHub Copilot, and Cognition's Devin already have real users and revenue. Magic has to prove that more memory actually makes a smarter engineer, not just a bigger number on a slide.

And it has to do that while burning enormous sums on compute against competitors who are already shipping. The clock and the cash both keep running.

Palantir

Palantir Gotham — the original intelligence platform used by government agencies for counterterrorism, military operations, and law enforcement. Integrates and analyzes data from disparate classified and unclassified sources.

Palantir Foundry — the commercial platform that lets corporations build data-driven applications without coding. Used for supply chain optimization, clinical trials, financial modeling, and manufacturing.

Palantir AIP (Artificial Intelligence Platform) — launched in 2023, this layer brings large language models and generative AI into Palantir's existing platforms, letting users query and act on their data using natural language. The product that supercharged the stock price.

Palantir Apollo — a continuous delivery system that manages software deployment across every environment: cloud, on-premise, classified networks, and even air-gapped military systems.

Magic AI

The centerpiece is Magic's LTM, or Long-Term Memory, model family. In 2024 they unveiled LTM-2-mini, a model with a 100 million token context window, pitched as the key to an AI that can understand an entire code base at once.

The broader product is an AI software engineer, or coworker, that can write and change code across a whole project. The company also announced Magic-G4 and Magic-G5, two supercomputers built with Google Cloud and NVIDIA to train these models.

Much of it is still research rather than a shipping product.

Palantir

In-Q-Tel (the CIA's venture arm) was the first investor and provided both capital and credibility. Peter Thiel's Founders Fund invested from the founding.

The company raised extensively from institutional investors including Tiger Global, Dragoneer, and Sompo Holdings. The September 2020 direct listing on the NYSE (similar to Spotify — no new shares sold) valued the company at approximately $22 billion.

The stock subsequently surged past a $200 billion market cap in early 2025.

Magic AI

Magic's backers are a murderers' row of AI insiders. It started with a $5 million seed round in 2022 and a $23 million Series A in early 2023 led by Alphabet's CapitalG.

Then in February 2024 came a $117 million round led by Nat Friedman and Daniel Gross, two of the most connected investors in AI. The blowout was August 2024.

Magic raised $320 million with checks from former Google CEO Eric Schmidt, CapitalG again, Atlassian, Jane Street, Sequoia, and Elad Gil. That pushed total funding to around $465 million, all for a company still building toward its first sellable product.