Bradley Tusk on regulating tech startups state-by-state, mobile voting going live in Anchorage, and AI's political backlash in midterms

Summary

Bradley Tusk, founder and CEO of Tusk Ventures, makes a pointed case that founders and investors are systematically misallocating their political capital by fixating on Washington. The vast majority of tech regulation, he argues, happens at the state and local level, and that structural reality is an opportunity rather than a liability. With roughly 200 bills passing per six-month state legislative session across 50 jurisdictions, a well-run campaign has a realistic shot at establishing precedent that then propagates nationally.

The clearest proof of concept is Doctronic, for which Tusk helped make Utah the first state to legalize prescription via AI. He points to Utah, Texas, Arizona, and Kentucky as the four states that have built regulatory sandboxes, making them the natural entry points for frontier tech companies. Texas is next on the Doctronic expansion roadmap given its scale as the second-largest state market. Earlier wins using the same state-by-state playbook include legalizing Uber (starting in 2011 under Travis Kalanick), daily fantasy sports for FanDuel following a 2016 New York Times insider-trading allegation that triggered cease-and-desist orders across dozens of states, cannabis delivery for Eaze, and prescription-by-text for Ro.

The political theory underlying the strategy is blunt: every politician makes every decision based on the next election, full stop. The leverage point is mobilizing a company's own user base to flood legislators with constituent pressure, exactly what FanDuel executed when its operating licenses were threatened. Tusk draws a direct line from union political spending to the model tech companies should replicate, noting that the threat of targeted spending against an incumbent in a primary is often sufficient to shift behavior.

Mobile Voting

Tusk's most structurally ambitious bet is the Mobile Voting Project, launched in 2017. The core thesis is electoral, not technological. A New York Times analysis of the 2024 general election found that only 8% of congressional races and 7% of state legislative races were decided by five points or less, meaning primaries, not general elections, determine most outcomes. Primary turnout currently sits around 10%, dominated by ideological extremes and organized special interests. Tusk argues that raising turnout to 40% via mobile voting would force politicians to represent a meaningfully more moderate electorate and unblock legislative gridlock on issues from immigration to climate.

After funding pilot elections in seven states for voters with disabilities and deployed military, Tusk's foundation spent five years building proprietary mobile voting technology that is end-to-end encrypted, biometrically screened, multifactor-authenticated, air-gapped, and fully open-source on GitHub. The first full municipal deployment goes live in Anchorage, Alaska in April 2026. Active legislation to allow cities in Minnesota, Colorado, Vermont, New Jersey, and Maryland to offer mobile voting in local elections is currently in progress.

The principal opponents are not ordinary voters but organized interests whose power depends on low-turnout primaries. Tusk specifically calls out groups like Verified Voting, which he characterizes as categorically opposed to any non-paper, non-in-person voting method.

AI's Political Exposure in the 2026 Midterms

On the political risk facing the AI industry, Tusk's outlook is straightforwardly negative. Affordability will be the dominant word in the 2026 midterms, which include 36 gubernatorial races and most state legislatures, and AI gets pulled into that frame through two channels.

First, job losses. Tusk contends that 2025 employment data already reflected AI-driven headcount reductions at large tech companies as early adopters realized they could perform functions with fewer people.

Second, and more politically volatile, energy prices. Hyperscaler data centers were designed to draw from existing grids, and in states where they have come online, residential energy bills have risen 30 to 50%. Unlike Uber, where users could directly demand the product in their district, there is no consumer constituency that will rally for a data center. The jobs created are minimal, and the cost is visible on every monthly bill.

Tusk expects AI to be a primary attack vector for challengers against incumbents across party lines in 2026, alongside concerns about teenage mental health and AI chatbots. Multiple states are already drafting legislation to constrain how data center energy costs are passed to consumers. The implicit warning is that the tech industry's internal enthusiasm for AI does not translate to the broader electorate, and assuming otherwise is a strategic error.