This story was originally published in On Background with Mark Stenberg, a free, weekly newsletter that explores the key themes shaping the media industry. You can sign up for it here.
Palantir built its original business on government work, including contracts with the Defense Department and Immigration and Customs Enforcement. But in recent years, the data analytics company has also begun selling to media and marketing firms.
In December 2025, Fox News Digital struck a deal with Palantir to build the internal platform that now supports the production of the vast majority of its articles. In June, Zeta Global signed a seven-year partnership with Palantir, and Stagwell has built two marketing products on its platform. And in August, USA Today Co. announced a deal that would let it analyze and monetize its audience data.
Some of the deals have drawn scrutiny. More than 800 USA Today Co. employees across 31 unionized newsrooms urged the company to end its partnership, citing concerns over reader trust, data security, and editorial independence.
But Palantir’s polarizing reputation—supporters view its defense tech and AI capabilities as essential, while detractors take issue with its surveillance tech, among others—did not weigh heavily on the executives behind these deals.
“There was no concern about reputation risk,” said Stagwell chairman and CEO Mark Penn. “We believe in Palantir and what it’s doing.”
For executives at the media and marketing firms, what Palantir could do with their data, how quickly it could do it, and, in some cases, which customers it could bring along ultimately made the case for the partnerships.
For Palantir, the tie-ups have enabled the company to learn, enter, and ally itself with key figures in the advertising space, part of a larger strategy for the intelligence firm to expand its business into “commercial data” more broadly, per Penn.
Palantir did not respond to a request for comment.
Making sense of massive data
Media and marketing companies sit on enormous amounts of fragmented data.
Zeta works with 550 million opted-in individuals and, at peak season, handles more than 7,500 computations per millisecond, according to CEO David Steinberg. Stagwell’s data sets span 1 billion touchpoints and 140 million U.S. households, according to John Kahan, its chief AI officer, who previously served as the chief data and analytics officer at Microsoft.
To make sense of these vast troves of data, Palantir pitches what it calls an “ontology,” according to Kahan, a single, organized model of a company’s data and operations that people and AI systems can query and act on. That capability became far more valuable once large language models arrived, according to Kahan.
