Palantir highlights Foundry growth as Karp defends contrarian product strategy

Traders Union
All News
Business & Wealth
Billionaires
Palantir Foundry growth strategy

Highlights

Palantir's U.S. commercial revenue grew 149% year over year to $764 million in Q2, with full-year guidance raised to over $3.4 billion, implying at least 134% growth.
Palantir and Zeta announced a partnership in June to rebuild Zeta's Data Cloud on Foundry, expected to generate more than $100 million annually within the next year.
CEO Alex Karp emphasized Palantir's contrarian strategy and forward-deployed engineering model, arguing that AI is increasing the competitive cost of execution across industries.

Foundry growth and leadership message

As reported by Fortune, Karp says one of Palantir's top executives once urged him to stop repeatedly telling employees how large Foundry would become, arguing the message felt disconnected from day-to-day realities inside the company.

Speaking on Thursday with Zeta Global Chief Executive David Steinberg at Zeta's annual AI marketing conference, Karp recalls Ted Mabrey telling him to stop talking to "Palantirians" about Foundry's future scale because it was depressing and made him seem out of touch. Karp says he stopped addressing staff for about six months, but adds that his decision-making style still depends on pursuing ideas that colleagues initially think cannot work.

Karp says he relies on long-standing internal relationships to get blunt feedback on his ideas. He describes those exchanges as harsh, while arguing that some of the company's most important initiatives begin as proposals that others reject.

Foundry, Palantir's platform for consolidating company data and controlling access to it, is now central to the company's fastest-growing business. In the second quarter, U.S. commercial revenue grows 149% year over year to $764 million, according to Palantir's earnings release, and the company raises its full-year U.S. commercial revenue guidance to more than $3.4 billion, implying growth of at least 134%.

That expansion is also feeding customer partnerships. In June, Palantir and Zeta announce a deal to rebuild Zeta's Data Cloud on Foundry, and Steinberg tells Adweek the seven-year partnership is expected to generate more than $100 million a year in sales within the next year or so.

Investor skepticism and wider AI implications

Karp says Palantir spends years convincing investors because many did not understand its software model in the company's early period. He argues the business resists the conventional approach of making products so embedded that customers cannot leave, describing that model as parasitic and saying venture capital investors repeatedly looked for that kind of lock-in.

He also says Palantir was mocked for sending engineers to work inside client organizations through its forward-deployed engineering model. Karp says that approach has since become one of Silicon Valley's fastest-growing roles, while Steinberg says Zeta copies the practice because of its efficiency.

Karp adds that Palantir has about 4,500 employees and acknowledges the company is not for everyone. He says some of its strongest endorsements come from former employees who later seek to return, echoing Mabrey's public appeal in March for "former hobbits" to come back to the "Shire."

During the AI-focused discussion, Karp argues artificial intelligence is raising the cost of weak execution across industries. Using defense as an example, he says AI application layers can determine whether an otherwise average system outperforms a superior product, and he tells executives the same principle applies to their own companies.

His advice to business leaders is to identify what differentiates their operations, build full visibility from the factory floor to cost of goods, and apply AI without allowing rivals or model providers to learn how the business works. He says companies that fail to do that risk being overtaken by competitors that execute better, even if those rivals are otherwise inferior.

In our earlier article on workflow-based defensibility in the AI era, we explained that AI tools become hardest to replace when they are embedded in the real systems, approval rules, and exception handling that run day-to-day operations. We also noted how dealmaking and strategy increasingly focus on owning these operational workflows—because even better models struggle to dislodge software that is already wired into how work gets done.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.