palantir and ukraine--5/19/2026

• 0 COMMENTS

Today's selection-- from The Philosopher in the Valley by Michael Steinberger. Palantir’s data targeting is joined to warfare:

“[Ukraine] was the first time in a while that Palantir's work in the realm of national security and defense was not bound up in controversy or ambiguity. The failed war efforts in Iraq and Afghanistan, coupled with human rights violations and concerns. about privacy and civil liberties, had tainted the fight against terrorism. The Trump years had been fruitful but challenging for Palantir. With Ukraine, however, there were no complicating factors (so far)--it was good versus evil, and Palantir was unequivocally on the side of good. 

“Karp was also eager to talk about Palantir and Ukraine because it was a way of deflecting attention from the company's sagging stock price and Wall Street's deepening skepticism. After trading as high as $45 per share in early 2021, Palantir's stock had steadily declined, and by the time Russia invaded Ukraine, it was barely above $10. True, it had been swept up in a general sell-off of tech stocks, but it was also the case that the market just didn't see Palantir as a winning investment. A headline in the Financial Times took Karp's words and used them against the company: ‘Palantir: Built for Bad Times (But Maybe Not These Bad Times).' The disconnect between the importance of the work that Palantir was doing and the value that the market placed on that work was large–and, for Palantirians, not a little maddening. 

“For Karp, the war in Ukraine was an opportunity, as well, to rally his own troops. The company was still full of industrious, ambitious people, but two years of Covid lockdowns, as well as the controversies that had dogged Palantir, had sapped morale–that was how Karp felt, anyway. The challenge on this front was evident in a town hall meeting he held that afternoon in Washington. With the pandemic still causing widespread illness, the meeting took place under a tent in the courtyard of Palantir's office. As his colleagues ate lunch, Karp, using a handheld microphone, riffed on a variety of topics, including himself. ‘Everyone thinks I'm insane,' he said. This perception, he went on to explain, had contributed to Palantir's success. ‘The thing about being viewed as bonkers is that people don't want to compete with you.’ He continued in this vein for around forty minutes. His quips drew polite laughter, but he seemed to struggle to connect with his audience. True, it was early March and a little chilly for an al fresco lunch (there were heaters in the tent, but they weren't offering much warmth), which might have muted the response. Still, it felt as if there was distance now between Karp and the rank and file, a point he acknowledged in his talk. ‘We've been separated for two years,' he said. ‘It's very hard to build a culture, to keep a culture, when you don't interact with people in person.’ He added, ‘We just have to try to get back together in the office.’ Much had changed during Covid: the company had gone public, work habits had been altered, and Karp had officially become a billionaire. He was also now a lot older than most of his employees. Could he still relate to them? Could they still relate to him? 

“Three years before Russia invaded Ukraine, Palantir started working on Project Maven, the Pentagon's artificial intelligence program, taking over for Google. According to Doug Philippone, Defense Department officials approached other tech companies first and reached out to Palantir only as a last resort because they were put off by the turmoil surrounding the firm at that time. For his part, Philippone was wary of getting involved with Maven. While the future of warfare undoubtedly revolved around AI, Maven had become a costly flop–the AI technology just wasn't good. Philippone says that the algorithms that were supposed to be able to identify, say, objects being surveilled by a drone were not up to the task: a sheep standing in a field would be mistakenly flagged as a man–that sort of thing. ‘The dirty secret of Maven was that this shit didn't even work,' says Philippone.

Project Maven analyst view

“Palantir didn't solve the AI problem–that task fell to other companies that were contributing to Maven. Instead, Palantir filled other gaps. Initially, it provided the project with a sophisticated mapping tool called Gaia. Not long thereafter, it occurred to Philippone and some of the engineers he worked with that Foundry could possibly deliver the data integration and analytic capabilities that Maven still lacked. That proved to be the case, and Foundry soon became Maven's operating system. Philippone says that Palantir ‘built the infrastructure that would allow the AI to be useful’ and gave the officers who were running Maven ‘breathing room' until the AI technology was workable. 

“By the time Russia invaded Ukraine, Maven had become a formidable asset, dramatically improving the military's situational awareness in combat zones and accelerating the so-called kill chain, the time between the identification of a target and its destruction or liquidation. The Biden administration sought to aid Ukraine without putting the United States in direct conflict with Russia, and Maven was integral to this strategy. Operating out of a secret site in Germany known as ‘The Pit,' the Army's 18th Airborne Corps, under the direction of Lieutenant General Christopher Donahue (who would later oversee the evacuation of Kabul Airport), used Maven to track Russian troop movements. The system ingested and analyzed vast quantities of satellite and drone imagery to show where Russian forces were attacking or massing in preparation for an assault; where Russian munition depots were located; and other vital intelligence. The U.S. shared this information with the Ukrainians, which helped them mount unexpectedly effective resistance during the first months of the war. 

“But via Foundry, Maven was able to deliver much more than just an exquisitely detailed view of the battlefield. Palantir's technology also enabled Donahue's team to pull in cellular data. As it happened, many Russian soldiers–including, it turned out, commanding officers–were using their personal cell phones to communicate because their military-issued phones didn't work. This allowed the Americans to identify their precise locations, information then shared with the Ukrainians. ‘It was the war on terror all over again,’ Philippone says. ‘I mean, the one thing we know how to do, we know how to pinpoint you anywhere in the world and kill you. The terrorists who learned to live without their phones survived. Those who couldn't, didn't.’ A number of Russian generals were dispatched to the front to try to salvage the sputtering war effort, and Ukrainian forces killed several of them in the first months of the conflict. They were located with the help of Maven, too, but the American military was cagier about how it passed along that information. The Ukrainians were given coordinates and told, in so many words, that they would be pleased with the result. 

“Karp was not quite as circumspect. ‘Palantir is responsible for most of the targeting in Ukraine,’ he said at a conference that the company hosted in Palo Alto. Karp had grown tired of having to always be so discreet about what he considered Palantir's most important work. In the past, he had talked only in vague terms about the terrorist attacks that Palantir had supposedly helped foil. His inability to offer specifics meant that the company never got the credit that he believed it deserved. It was, he said, a ‘twenty-year frustration.’ But after deleting that passage in his letter referencing Palantir's involvement in Ukraine, he had now decided that it was time for Palantir to get the recognition that it deserved: the world would know that it was helping Ukraine. (The Pentagon was not thrilled about his outspokenness and eventually asked him to dial it back, which he did.)"

author:	Michael Steinberger
title:	The Philosopher in the Valley: Alex Karp, Palantir, and the Rise of the Surveillance State
publisher:	Simon & Schuster
pages:	215-222