Year-round
Software Engineer Intern
Infrastructure
Palantir
5,001-10,000 employees
Enterprise software enabling data-driven transformation
Compensation Overview
$60.58/hr
No H1B Sponsorship
New York, NY, USA
In Person
US Top Secret Clearance Required
 Software Engineering |
|---|
Get referred to Palantir
See people who can refer or advise you
- Engineering background in fields such as Computer Science, Mathematics, Software Engineering, or Physics.
- Familiarity with data structures, storage systems, cloud infrastructure, front-end frameworks, and other technical tools.
- Experience coding in programming languages, such as Java, C++, Python, TypeScript, JavaScript, or similar languages.
- Must be planning on graduating in 2028. This should be your final internship before graduating.
- As a Software Engineering Intern in Foundations, your projects may include building features to improve the developer experience for other Palantir engineers, or improving the scalability and reliability of Palantir’s platforms.
- You’ll be involved throughout the product lifecycle, from idea generation and design, to execution and rollout, all while being paired with a mentor dedicated to your growth and success.
- You'll collaborate closely with technical and non-technical counterparts to understand our developers' and customers' problems and build infrastructure to tackle them.
- Passion for helping other developers build better applications.
- Empathy for the impact your changes will have on the workflows and productivity of developers and end users.
- Ability to communicate and collaborate with a variety of individuals, including engineers, users and non-technical team members.
- Willingness to learn and make decisions independently, and the ability to ask questions when needed.
- Active US Security clearance, or eligibility and willingness to obtain a US Security clearance.
Palantir builds software that helps large organizations run their digital transformation by giving them tools to access, connect, and analyze all of their data. Its platforms pull data from many sources, clean and link it, and then let users explore dashboards, reports, and AI-powered insights to make informed decisions. Unlike many analytics tools that focus on one data source or a single function, Palantir emphasizes an integrated, enterprise-wide data foundation with governance and security to support complex environments. The goal is to turn raw data into actionable intelligence that guides strategy and operations, helping clients deploy and scale transformative programs.
Company Size
5,001-10,000
Company Stage
IPO
Headquarters
Aventura, Florida
Founded
2003
Get referred to Palantir
See people who can refer or advise you
Simplify's Take
What believers are saying
- Q2 2026 revenue rose 93% to $1.94 billion; U.S. commercial revenue grew 149%.
- August 2026 guidance raised full-year revenue to $8.15 billion-$8.158 billion, Palantir’s largest increase.
- Metropolitan Police litigation targets a £50 million contract, showing demand persists despite ethics objections.
What critics are saying
- Percepta AI litigation and 2025 noncompete fights expose aggressive employee-retention tactics and trade-secret disputes.
- Zurich’s June 2026 ruling gutted Palantir’s Republik lawsuit, reinforcing reputational and legal backlash abroad.
- At $421 billion valuation, a guidance miss or AIP slowdown triggers severe multiple compression.
What makes Palantir unique
- Palantir’s AIP keeps customer data inside client environments, not shared with frontier labs.
- Its software spans defense, government, and enterprise workflows, enabling cross-domain operational integration.
- Management markets outcomes-based contracts, aligning fees with mission results instead of token consumption.
Help us improve and share your feedback! Did you find this helpful?
Benefits
Transparency
Take-What-You-Need Time Off Policy
Family Support
Community
Equity
Mental Health and Wellbeing
Healthcare
Growth & Insights and Company News
6 month growth
↑ 0%1 year growth
↑ 0%2 year growth
↑ 0%Palantir Technologies' stock rose 34% through late August, but the factors driving the surge were publicly disclosed months earlier. Key contract wins, including deals with Stellantis, the USDA (worth up to $300 million), and Cleveland-Cliffs, were all announced by early May. The company's net dollar retention jumped to 150% in fiscal Q1 2026, up from 139% the previous quarter, indicating existing customers were spending significantly more. This growth came with improved profitability, as trailing operating margin reached 38.1% compared to a three-year average of 14.2%. In August, Palantir reported US commercial revenue up 149% year-over-year to $764 million and raised full-year 2026 revenue guidance to $8.154 billion, representing 82% growth. Management attributed the momentum to "AI sovereignty" demand, as companies sought greater control over their data and security.
Palantir CEO Alex Karp pushes AI sovereignty model, warns against Silicon Valley concentration of power. August 26, 2026 Palantir Technologies CEO Alex Karp is making the case that companies should own their own AI tools and data outright, a direct challenge to the handful of frontier labs that dominate the industry. Karp laid out his argument at a recent Palantir boot camp, warning that the current trajectory of artificial intelligence development funnels too much wealth and control to a tiny sliver of the tech world. The Miami-based software company has spent the past several months rolling out what it calls an "AI sovereignty" model, pitching businesses and governments on an alternative to renting intelligence from the big labs. The pitch lands at a moment when public confidence in AI companies is remarkably low. A December 2025 poll conducted by Anthropic and YouGov found that just 15 percent of Americans trust AI companies to make decisions about the technology's development and use. Palantir is betting that distrust can become a selling point, if the company can persuade customers that ownership, not dependency, is the smarter path forward. Karp targets "2,500 people sitting in Silicon Valley" At the boot camp, Karp framed the stakes in blunt economic terms. As the New York Post reported, the Palantir CEO warned that the current model concentrates gains among a vanishingly small group of insiders. "We cannot have a society where all the value goes to 2,500 people sitting in Silicon Valley. That just will not work, and no one's going to put up with it." That figure is rhetorical, Karp's own shorthand for the engineers, executives, and investors who stand to capture most of the upside from frontier AI. But the complaint tracks with a broader unease, shared across the political spectrum, about how much power a few companies now hold over a technology that touches nearly every industry. Palantir's answer is a model where the customer, whether a hospital system, a manufacturer, or a government agency, keeps its own data and runs its own AI tools. Kevin Kawasaki, Palantir's head of business development, put it plainly in an interview: "Your data, your model, not ours, not someone else's, you're not handing it over to anyone else." Kawasaki went further, drawing a sharp line between ownership and what he described as a rental arrangement under the current lab-dominated system: "Sovereignty means you own the right to run your business. If an AI lab can access your data, train on it, and sell that model to your competitors, you're not running your business. You're renting it." That framing positions companies like OpenAI and Google DeepMind not just as service providers but as potential competitive threats to their own customers. Whether that characterization holds up across every use case is an open question. But the underlying concern, that handing proprietary data to a third-party AI lab creates real risk, is one that corporate boards and government procurement officers are already wrestling with. Palantir's nine-point manifesto and the "token-maxing" critique. Over the past few months, Palantir has escalated its sovereignty campaign. The company posted a nine-point manifesto on X, its official account laying out the case against what it calls dependency on frontier labs. Palantir also launched the boot camp series for customers, designed to walk them through the alternative model in practice. A central piece of the argument is financial. Chad Wahlquist, Palantir's lead architect, drew a direct contrast between how Palantir gets paid and how the major AI labs create revenue. Companies like OpenAI make more money when customers buy more tokens, the units of data processing that power AI queries. Palantir calls this "token-maxing," a term the company uses to describe an incentive structure it considers misaligned with customer interests. Wahlquist was blunt about the difference: "Everyone else is focused on consumption. That's how they get paid. We get paid on outcomes. That is the core differentiator about us to everyone else." The distinction matters if it holds. A consumption-based model rewards volume, more queries, more data processed, more billing. An outcomes-based model, at least in theory, ties the vendor's revenue to whether the customer actually gets results. Palantir has not publicly detailed the specific contractual mechanics of its outcomes-based pricing, so the claim is worth watching but not yet independently verified. Frontline workers versus lab-driven fear, according to Palantir. Wahlquist also pushed back on the narrative coming from some of the most prominent voices in AI, the warnings about mass unemployment, civilizational risk, and the need for government-administered safety nets to cushion the blow of automation. The article references Anthropic CEO Dario Amodei as an example of AI leaders who have promoted what Palantir characterizes as a "doomer narrative" about the technology's trajectory. No direct Amodei quote appears in the source, and the specific language attributed to that camp, talk of extinction events and handouts, may reflect Palantir's own characterization rather than Amodei's precise words. What is clear is that Palantir sees an opening in the gap between the apocalyptic tone of some lab leaders and the everyday experience of people who already use AI on the job. Wahlquist framed the contrast this way: "You actually talk to the nurses, the factory floor workers, the people that actually use AI to do their job that are on the ground, it is a far different story than what is being led by prominent AI labs the fear tactics." That line carries a populist edge. It positions Palantir alongside the workers and operators who use the tools daily, against a class of Silicon Valley executives whose public statements about AI risk sometimes sound more like bids for regulatory moats than genuine safety concerns. Whether a company that builds surveillance and intelligence platforms for governments worldwide qualifies as a populist champion is a fair question. But the framing is deliberate, and it resonates with a real frustration. Only 15 percent of Americans trust AI companies with the keys. Palantir's sovereignty push arrives against a backdrop of deep public skepticism. The December 2025 Anthropic-YouGov poll, notably co-sponsored by one of the very labs Palantir is criticizing, found that only 15 percent of Americans trust AI companies to make decisions about how the technology is developed and used. That number is striking on its own. It is even more striking when you consider that Anthropic itself paid for the survey and still published the results. The poll's methodology and sample size are not detailed in the available reporting, so the figure deserves the usual caveats that apply to any single survey. Still, the direction is unmistakable. Americans are not sold on the idea that a handful of companies should steer the future of artificial intelligence with minimal oversight or competition. Palantir is positioning itself as the company that agrees with that instinct, and has a product to sell because of it. Karp's language about "vassal states of the language labs" frames the current arrangement not as a market choice but as a form of subordination. Companies that rely on frontier labs for their AI, in this telling, are not customers. They are dependents. What Palantir hasn't answered yet. For all the sharp rhetoric, several questions remain open. Palantir has not disclosed the specific contents of its nine-point manifesto beyond the sovereignty theme. The company has not named specific government or corporate clients operating under its outcomes-based model. And the term "token-maxing" appears to be Palantir's own coinage, it is not clear whether independent analysts or competitors use it or accept its framing. There is also the matter of Palantir's own track record. The company built its reputation on deep integration with intelligence agencies and defense departments. Its tools process enormous volumes of sensitive data. Asking customers to trust Palantir with their information while warning them not to trust OpenAI or Google DeepMind with the same data requires a credibility case that goes beyond a manifesto and a boot camp. None of that invalidates the core argument. The question of who owns the data, who controls the model, and who profits from the output is one of the most consequential business and policy questions of the next decade. Palantir is at least asking it out loud, which is more than most of the major labs have been willing to do. When only 15 percent of the country trusts the people building the most powerful technology of the century, somebody is going to step into that vacuum. The question is whether Palantir is offering a genuine alternative, or just a different landlord.
Azonic raises pre-seed round backed by Blackbird, Antler and former CIA and FBI leaders. Crimson Education founder Jamie Beaton, Pushpay co-founder Eliot Crowther and Emerge founder Jovan Pavlicevic join the pre-seed round. Today Azonic, Inc. is announcing Azonic's pre-seed round, backed by Blackbird and Antler alongside Jamie Beaton, founder and CEO of Crimson Education; Eliot Crowther, co-founder of Pushpay; and Jovan Pavlicevic, founder of Emerge. Azonic is a unified intelligence platform for law enforcement and intelligence agencies investigating online crime: one system for the reports, case files and intelligence that today sit in separate places, with a virtual analyst working alongside every investigator. Investigators from New Zealand, Australia and the US are already working with the platform today. Azonic, Inc. come to this from inside the fight. Co-founder and CEO Caden Scott founded the International Online Crime Coordination Center (IOC3), the non-profit that connects law enforcement agencies and the world's largest online platforms against online crime. Co-founder and CTO Thomas Holdom has spent years building AI systems for large, regulated organisations. Azonic, Inc. is also introducing its advisory board: Michael Hurley, formerly of the CIA and Palantir, where he was a senior adviser to CEO Alex Karp, and Brad Brekke, founding director of the FBI's Office of the Private Sector. "Online crime is scaling faster than the systems built to investigate it, and Azonic is tackling that gap with technology designed specifically for the realities of investigative work. Caden brings rare firsthand exposure to the problem through his work building IOC3, while Thomas brings deep experience building AI systems for complex organisations. Together, they're a potent combination of industry expertise and technical capability needed to take on a problem of this scale, and we're excited to back them" said James Palmer, Blackbird. Azonic was started because the fastest-growing criminal networks in the world have no headquarters to raid and no hierarchy to dismantle. They run on the platforms millions of Azonic, Inc. use every day, organising child sexual exploitation, sextortion, fraud and violent extremism across dozens of countries at once. When one account falls, three more appear, and reports of these crimes reaching agencies have multiplied year after year. Behind each one is a person, too often a child. In front of each one is an investigator working to reach them in time. Justice is slower by design. It demands care, rigour and proof that holds up in court. Offending is now automated, distributed and effectively industrial, while the work of proving it remains handmade. The mathematics of the fight have changed, and the people protecting children and communities deserve a way of working built for the new mathematics. The mission is simple to state and hard to do: help them move at the speed and scale of the crime they face. "In my former role with the FBI I observed first hand the speed and scale of threats emerging from the cyber world. I also saw the limitations of traditional methods to identify and track those threats. When I was introduced to Azonic and saw what Caden and Thomas had developed, I was impressed by their innovative platform and their personal commitment to assisting law enforcement. Azonic provides a new powerful tool for the intelligence community to stay ahead of the threats." said Brad Brekke, founding Director of the FBI's Office of the Private Sector. The virtual analyst amplifies every investigator standing between offenders and their victims: it runs down leads, connects a case to the ones that came before it, and drafts the outputs that move cases forward, so no report waits unseen. It is built for the teams doing the hardest work there is: child protection units, counter-extremism teams, and the investigators tracking offending that crosses every border it meets. "Azonic's unified intelligence platform will be an essential and invaluable tool to agencies all over the world that investigate online crimes. As such it has tremendous potential for business growth in law enforcement and national security agencies. The United States alone has 19,000 law enforcement agencies. Canada, the UK, and European countries will constitute a large market for Azonic's unique capabilities." said Michael Hurley, former CIA & Palantir. This is not off-the-shelf AI pointed at policing. On their own, today's AI models do not meet the bar these agencies hold their own people to. Azonic is engineered from the ground up to meet it: a complete intelligence system built for investigative work, of which AI is only one part. Criminal networks have adopted AI with no rules at all. Leveling that field, inside the rules of evidence, is at the core of Azonic. "We backed Azonic because Caden and Thomas have an earned insight to make a significant difference to the intelligence world. Caden has spent years inside the investigative world these agencies operate in, and Thomas has the AI engineering horsepower to ship the agentic tooling their customers desperately need." - James McClure, Antler. Azonic, Inc. is building Azonic with investigators, not just for them: what they tell Azonic, Inc. shapes what Azonic, Inc. ship each week. "I first met Caden as an inquisitive student in Business Navigators. I was blown away when he described his 'side hustle' fighting cyber crime and quickly realised his profound and burning passion for the space. The unicorn creation journey requires relentless obsession and Caden has that. We are delighted to support him," said Jamie Beaton, Crimson Education. This round funds the next stage of the platform and its work with the agencies on the front line of online crime. Crime scales without limits. Investigators do not. Azonic solves that. Azonic is a unified intelligence platform for law enforcement and intelligence agencies, headquartered in Auckland, New Zealand. Agencies can request access at azonic.ai. Media enquiries: [email protected].
31 ex-lawmakers, staffers joined military firms in past 10 years. One former official is now working as a defense lobbyist even as he serves on a major Navy commission, according to a new report Aug 24, 2026 In the last decade, 31 former members of Congress and top staff have left Capitol Hill and joined the military-industrial complex as lobbyists or contractors, according to a new report from the Bulletin of the Atomic Scientists. The 21 lawmakers named in the report served on either the House or Senate Armed Services Committee, the legislative bodies that set the Pentagon's annual policies through the annual National Defense Authorization Act. In addition, the report names 10 former senior committee staffers who transitioned to working as weapons industry government liaisons or lobbyists. These 31 former officials represent a snapshot of the military-industrial complex's "revolving door," through which ex-policymakers transition from setting the Pentagon's agenda in Congress to cashing in on those same policies in the private sector. "That report we published is the most direct explanation we've ever seen of why the defense budget always goes up," said John Mecklin, the author of the report and former editor-in-chief of the Bulletin of the Atomic Scientists. The report highlighted a variety of potential ethics conflicts, focusing in particular on the growing path between Congress and Silicon Valley defense tech startups. Chris Vieson, the former staff director of the House Armed Services Committee (HASC), was appointed earlier this year to the National Commission on the Future of the Navy, established to advise Congress on everything "from shipbuilding and acquisition to maintenance and repair, workforce, training, and modernization." Vieson is also an active registered lobbyist for two defense startups: Saronic Inc., a manufacturer of unmanned maritime vehicles; and watercraft builder Whiskey Project Group. Previously, Vieson worked as a lobbyist for Andreessen Horowitz, the venture capital firm gaining influence in the Pentagon and pushing AI integration. According to its executive director, the National Commission on the Future of the Navy aims to modernize the Navy to prepare for conflict with China and Iran. As part of this goal, the Commission will "test emerging ideas such as a hybrid fleet and expanded use of unmanned systems." Vieson's lobbying client, Saronic Inc., is one manufacturer of such unmanned systems. While the revolving-door phenomenon is not new, the report captures the increasing influence of a new generation of Silicon Valley military firms. Former HASC member Mike Gallagher and staffer Jeff Miller have taken up jobs as an executive and a lobbyist, respectively, for Palantir. Former Republican staff director of the Senate Armed Services Committee (SASC) Christian D. Brose is now the Chief Strategy Officer for Anduril, an autonomous weapons startup valued at $61 billion in its latest round of fundraising. Subscribe now to its weekly round-up and don't miss a beat with your favorite RS contributors and reporters, as well as staff analysis, opinion, and news promoting a positive, non-partisan vision of U.S. foreign policy. According to Mecklin, Congress is eager to let Silicon Valley officials into their bureaucracy. "It takes the Pentagon many, many years to acquire a new defense system like a fighter plane. With a high-tech system, you just can't wait that long," he said. "So they found a way around the complicated acquisition processes of the Defense Department." Likewise, Silicon Valley weapons startups that wish to challenge prime contractors like Boeing and Lockheed Martin are desperate to hire former politicians. "If you're a company that wants to get into defense contracting, it's highly beneficial to hire people that understand the process of federal appropriations in the defense sector," Mecklin explained. The Bulletin's revolving-door report also details a who's-who of top congressional stock traders. Rep. Austin Scott (R-Ga.), who has served on HASC since 2015, has been trading GE Aerospace stock via his spouse since 2023. Scott's trades this year violated the STOCK Act, a bill passed in 2012 to curb insider trading. That makes Scott one of seven Armed Services Committee members to violate the STOCK Act this year, generally for failing to disclose trades within 45 days, as the law mandates. Rep. Lisa McClain (R-Mich.) is another HASC member who has benefited from military contractor stocks. In December 2025, she disclosed her husband's $100,000-$250,000 purchase of stock in Elon Musk's xAI, which merged with SpaceX soon after in February. McClain's husband's xAI shares were converted into SpaceX shares months before the company's historic IPO in June. McClain previously violated the STOCK Act for her delayed disclosure of a trade worth up to $450,000 in Palantir stock. The report also makes note of Sen. Tommy Tuberville (R-Ala.) trading stock for prime defense contractor Lockheed Martin and aerospace manufacturer Honeywell. Tuberville, one of the most active stock traders in Congress, has served on SASC since 2021. When President Donald Trump faced criticism this spring for his own portfolio of weapons stocks, Tuberville told a reporter for MeidasTouch, "I do the same thing." Mecklin believes that laws like the STOCK Act are insufficient to curb members of Congress cashing in on defense spending. "The way you really stem this kind of behavior is to have inspectors general and law enforcement agencies that are on the lookout for them. Occasionally, take up one of these egregious cases and just prosecute it," said Mecklin. "Unless you have a system that actually scares these people so they cut it out, it's going to continue." Julian Cooper is a Chicago-based journalist, an editorial intern for The Progressive, and a research intern for the Quincy Institute's Democratizing Foreign Policy program. The views expressed by authors on Responsible Statecraft do not necessarily reflect those of the Quincy Institute or its associates. Top image credit: Rep. Mike Gallagher (R-WI) walks through the U.S. Capitol, in Washington, D.C., on Tuesday, February 6, 2024. (Graeme Sloan/Sipa USA) via REUTERS Rep. Mike Gallagher (R-WI) walks through the U.S. Capitol, in Washington, D.C., on Tuesday, February 6, 2024. (Graeme Sloan/Sipa USA) via REUTERS From Your Site Articles
Palantir reported revenue of $1.94 billion, up 93% year-over-year, with US commercial revenue jumping 149% to $764 million. The company posted a Rule of 40 score of 155% and generated $1.22 billion in free cash flow. Chief executive Alex Karp attributed growth to demand for AI sovereignty. Snowflake delivered product revenue of $1.33 billion, growing 34%, its strongest sequential dollar gain. Its Cortex Code product is now used in over 7,100 accounts, whilst Snowflake Intelligence accounts more than doubled quarter-over-quarter. The companies employ contrasting strategies. Palantir pursues a closed, integrated technology stack. Snowflake follows an open ecosystem approach, recently signing a $6 billion multi-year AWS agreement and expanding partnerships with OpenAI and SAP. Snowflake reported negative GAAP operating income of $326 million.