Palantir sues former senior engineers for allegedly using confidential source code and client data to launch Percepta AI

Palantir has filed a lawsuit in Manhattan federal court accusing two former senior engineers of taking what the company calls its “crown jewels” to help launch a competing artificial intelligence firm named Percepta AI, according to documents filed on Thursday.

The company alleges that Radha Jain and Joanna Cohen walked away with confidential information, including internal source code and customer data, and then used the info to create Percepta’s software.

Palantir also says that Percepta is structured to compete in the same market where it has spent years developing tools that allow government agencies and private companies to better use the data they already have.

Radha left the company in November 2024 after contributing to the design and construction of Palantir’s main software platform. Joanna resigned in February after working on AI solutions tailored to individual clients.

Percepta AI was publicly introduced earlier this month and is backed by investment from General Catalyst, a well-known venture capital firm. Palantir argues that the timeline and the type of work both individuals were involved in show a direct link between what they learned at Palantir and what Percepta is now offering.

Company claims contracts were violated

Palantir says Radha and Joanna signed written agreements before they left. These agreements barred them from competing with the company for one year, from approaching Palantir customers or employees for two years, and from using or sharing any confidential information outside their work for Palantir.

In the court filing, Palantir stated, “Jain’s and Cohen’s deception and violation of their agreements with Palantir are black and white.” The company argues that these terms were put in writing, acknowledged, and then ignored once the pair left.

The lawsuit also claims that Percepta hired at least ten former Palantir workers within a few months of being formed last year. Nearly half of Percepta’s current team previously worked at Palantir, including its co-founder and current CEO, Hirsh Jain, who is Radha’s relative.

Palantir claims this shows Percepta is not just a startup operating in a similar field, but a firm built with internal knowledge, culture, and processes lifted directly from inside Palantir.

Attempts to contact Percepta for comment were unsuccessful. Attempts to reach Radha and Joanna were also not returned.

Percepta is not named as a defendant in the case, but Palantir says the lawsuit is intended to force Radha and Joanna to comply with the terms they previously agreed to.

Palantir CTO pushes back on AI “doomer” fears

As Palantir deals with the lawsuit, one of its top leaders has also entered the ongoing public debate over AI and the future of humanity.

In an interview with the New York Times, Shyam Sankar, Palantir’s Chief Technology Officer, said he is “super skeptical” of fears that AI could lead to human extinction or mass loss of control. He said some technologists in Silicon Valley are motivated by belief rather than evidence.

“I think the secularists in Silicon Valley are filling the God-shaped hole in their heart with AGI,” Shyam said. He questioned what he described as a dramatic leap from improving AI models to assuming machines will dominate humans. “Why do you think this cliff is going to happen where they somehow turn us into house cats?”

Shyam said the loudest voices warning about AI are often those who identify with transhumanist ideas or want to push a narrative for financial benefit, adding that the doomer messaging has become a “fundraising shtick” as some companies claim their AI systems are powerful enough to wipe out jobs just so that investors feel compelled to fund them.

Shyam responded to concerns about Palantir and other firms developing AI systems for military use. He said fears of machines replacing military decisions misunderstand how the technology works in real environments. “I just don’t think it would work the way that people think it would work in the Terminator nightmare scenario,” he said. He explained that using AI in military settings usually means improving how people work, not removing them from the process. “When you see it intersect the front line, it’s actually empowering the worker to do more,” he said.

If you're reading this, you’re already ahead. Stay there with our newsletter.