Major AI companies are rapidly increasing their political influence through lobbying and donations as they navigate conflicting regulatory priorities. While these firms are united in their need for compute, data, and engineering talent, they differ significantly in their approach to government oversight under the current administration. OpenAI and Meta have increasingly aligned with the Trump administration, while organizations like Anthropic have distanced themselves, often supporting stricter federal oversight on issues such as model transparency and AI safety.

Anthropic CEO Dario Amodei recently donated $1 million to Public First, a super-PAC that advocates for an AI governance framework, export controls on chips, and safeguards against catastrophic risks. Anthropic had previously contributed $20 million to the group’s associated action committee. In contrast, OpenAI president Greg Brockman donated $25 million to Leading the Future (LTF), an industry-backed PAC that aims to preempt strict state-level regulations. LTF seeks to establish exclusive federal jurisdiction over AI policy to prevent states from imposing independent restrictions, mirroring lobbying tactics observed in the cryptocurrency sector.

Industry disagreement over the role of state regulation remains a point of contention. Public First argues that state oversight, such as that seen in New York or California, acts as a necessary check in the absence of comprehensive federal standards. Proponents of LTF, however, advocate for federal intervention to prevent localized regulatory barriers. Despite these internal debates, the article notes that both sides share the goal of securing American dominance in AI relative to China. The effectiveness of current hardware export controls is increasingly questioned as Chinese firms, including Moonshot AI with its Kimi 3 model, continue to achieve performance competitive with American frontier models while shifting toward domestic hardware solutions.

From an external perspective, the distinctions between these political projects may be less significant than they appear from within the industry. Growing public and state-level backlash against the resource demands of data centers could bypass the influence of these PACs, potentially leading to broader political outcomes such as public firm ownership. As AI technology evolves, the window for industry-led regulatory capture may be closing, regardless of which lobbying faction holds more sway in Washington.

Major AI companies are rapidly increasing their political influence through lobbying and donations as they navigate conflicting regulatory priorities. While these firms are united in their need for compute, data, and engineering talent, they differ significantly in their approach to government oversight under the current administration. OpenAI and Meta have increasingly aligned with the Trump administration, while organizations like Anthropic have distanced themselves, often supporting stricter federal oversight on issues such as model transparency and AI safety.

Anthropic CEO Dario Amodei recently donated $1 million to Public First, a super-PAC that advocates for an AI governance framework, export controls on chips, and safeguards against catastrophic risks. Anthropic had previously contributed $20 million to the group’s associated action committee. In contrast, OpenAI president Greg Brockman donated $25 million to Leading the Future (LTF), an industry-backed PAC that aims to preempt strict state-level regulations. LTF seeks to establish exclusive federal jurisdiction over AI policy to prevent states from imposing independent restrictions, mirroring lobbying tactics observed in the cryptocurrency sector.

Industry disagreement over the role of state regulation remains a point of contention. Public First argues that state oversight, such as that seen in New York or California, acts as a necessary check in the absence of comprehensive federal standards. Proponents of LTF, however, advocate for federal intervention to prevent localized regulatory barriers. Despite these internal debates, the article notes that both sides share the goal of securing American dominance in AI relative to China. The effectiveness of current hardware export controls is increasingly questioned as Chinese firms, including Moonshot AI with its Kimi 3 model, continue to achieve performance competitive with American frontier models while shifting toward domestic hardware solutions.

From an external perspective, the distinctions between these political projects may be less significant than they appear from within the industry. Growing public and state-level backlash against the resource demands of data centers could bypass the influence of these PACs, potentially leading to broader political outcomes such as public firm ownership. As AI technology evolves, the window for industry-led regulatory capture may be closing, regardless of which lobbying faction holds more sway in Washington.