a16z-Backed Group Turns Data Center Lobbying Into Midterm Spectator Sport

Build American AI, a group bankrolled by Andreessen Horowitz and OpenAI's Sam Brockman, has announced plans to spend millions on political advertising during the midterm elections to convince voters in select states that data centers are actually good, actually. The premise is simple: if regulatory approval moves too slowly and local communities keep asking inconvenient questions about power consumption and land use, simply bypass both and appeal directly to the electorate. It's the kind of move that makes you wonder whether venture capital has finally abandoned even the pretense of building things and simply decided to become a full-service lobbying operation with better pitch decks.

Build American AI presents itself as an infrastructure advocacy group with a straightforward mission: promote the construction of data centers across the United States. The organization's actual product is messaging—specifically, the kind of messaging designed to make voters believe that their own economic interests align perfectly with the interests of a16z's portfolio companies and their AI infrastructure requirements. There is no revenue to speak of, no customer base, no problem it solves except for the problem of convincing Americans that they should want more massive energy-intensive computing facilities in their backyards. It is pure, distilled rent-seeking dressed up in a patriotic flag pin.

This is not Andreessen Horowitz's first rodeo with political capture, though the directness is refreshing in its shamelessness. The firm has spent years cultivating relationships with policymakers and funding think tanks that align with its worldview. But those efforts typically operate behind the scenes, feeding ideas to the right people in the right rooms. Build American AI simply cuts out the middleman: if you can't convince a zoning board or a state environmental review committee, you can at least convince enough voters to make politicians nervous about opposing you. It's democracy as a cost center.

The pitch, one imagines, centers on American competitiveness, technological sovereignty, and the urgent need to build infrastructure that will keep the country competitive against China in AI. These are real concerns, articulated in good faith by serious people. Build American AI's contribution is to reframe them as synonymous with the specific interests of its funders—a rhetorical move so common in venture capital that it barely registers anymore. When a16z funds a political advocacy campaign, it is not lobbying on behalf of artificial intelligence; it is lobbying on behalf of the specific regulatory and tax environment that maximizes returns for a16z's fund.

The success of this effort depends entirely on whether voters in select states can be persuaded that data center construction serves their interests more than, say, concerns about water usage, electricity costs, or the industrial transformation of their communities. History suggests this is an uncertain proposition. Voters tend to object to massive infrastructure projects in their backyards regardless of how many television spots explain why they should be grateful for them. The money spent on ads represents a bet that sufficient messaging saturation can overcome baseline local opposition—a bet that has failed spectacularly in other contexts.

What Build American AI really represents is the moment when venture capital stops pretending to be about innovation and openly embraces its identity as an extractive industry. These firms manage billions of dollars and deploy them to shape the world according to their financial interests. Sometimes that produces useful products; often it doesn't. But even when it does, the public conversation about whether those products should exist in a particular place, in a particular way, has been short-circuited by the simple application of capital to the problem of voter persuasion. It is the opposite of the disruption that venture capital claims to champion.

Build American AI will probably succeed in moving some voters' opinions about data centers, and it will probably fail to move many others. Either way, the real victory is already won: by treating electoral opinion as just another variable to be managed through paid media, a16z has established the principle that the only constraint on venture capital's ambitions is its willingness to spend money on advertising. Democracy, from this perspective, is just another market inefficiency waiting to be arbitraged.