AI Stories on SHORT INFO are generated & curated with AI
unverified 03 Jul, 15:10

Palantir CEO Karp says some US government customers switched to open-source AI, blasts frontier lab token pricing, Nvidia Nemotron deal

Palantir $PLTR CEO Alex Karp says some US government customers have switched to open-source AI, calling frontier labs' token pricing broken and a push toward tools enterprises can control. Palantir and Nvidia $NVDA are bringing Nvidia's open models into federal agencies. Per CNBC

Palantir CEO Alex Karp used a public appearance this week to deliver a sharp critique of the leading AI labs, and to disclose a shift already underway inside government. According to reporting, Karp said some US government customers have moved to open-source AI models, and he argued that the token-based pricing used by frontier labs is fundamentally broken. His concern is twofold: enterprises face escalating and unpredictable costs, and they risk exposing proprietary data and intellectual property to outside model providers. The disclosure lands alongside a concrete deal. Palantir $PLTR and Nvidia $NVDA have agreed to bring Nvidia's open Nemotron models into US government agencies and critical infrastructure, pairing Nvidia's compute with Palantir's data and AI stack across classified and air-gapped environments. Karp framed the alignment around customer ownership: technically sophisticated clients want sovereignty over their compute, models and data rather than a setup where those assets quietly migrate to a third party. The broader backdrop is a cost squeeze. High token prices and uneven returns have pushed several US organizations to adopt or evaluate open-weight models, including systems developed by Chinese labs. For the closed frontier providers that built their businesses on metered token consumption, a large and security-conscious buyer like the US government signaling a move toward models it can host and control is a direct challenge to the pricing assumptions underpinning that model. Per CNBC and Yahoo Finance.

#tech
Published on
ThreadsFacebookXBluesky