Platform antitrust is back in the Senate, and this version of AICOA comes with an explicit AI design exception. Senators Chuck Grassley (R-Iowa) and Amy Klobuchar (D-Minn.) reintroduced the American Innovation and Choice Online Act on June 11, targeting what the bill’s sponsors call “systemically important” digital platforms, and reportedly drawing a line around AI development that earlier iterations of the bill didn’t include.
The threshold language matters. As introduced, the bill would reportedly designate platforms above $175 billion in annual revenue and 34% US user reach as “systemically important,” according to reporting on the bill’s provisions. That framing targets a narrow set of companies, effectively the largest US technology platforms, and subjects them to prohibitions on self-preferencing, misuse of competitor data, and blocking data portability. These are the core AICOA provisions from prior iterations. They’re not new.
AICOA 2026: Positions
What’s new is the reported carve-out. The reintroduced bill reportedly includes language stating that neither the FTC nor the DOJ would be authorized to dictate AI development, model design, or ranking policy, according to reporting on the legislation. This provision hasn’t been independently verified against the bill text for this brief, it’s the item’s primary news value, and it’s sourced to industry reporting rather than the bill itself. Don’t build compliance plans around the carve-out’s specifics until the bill text is directly consulted.
The catch is that a carve-out in an introduced bill is a negotiating position, not a guaranteed outcome. The same platforms that would face AICOA’s self-preferencing restrictions are the companies most likely to lobby hard against the bill’s passage, and Apple has already pushed back publicly, arguing that restrictions on product integration would compromise user security and privacy, according to reporting on the company’s response. That’s Apple’s standard objection to platform regulation, but it carries lobbying weight regardless of its merit as a policy argument.
Three things to watch. First, whether the bill reaches committee markup at all, AICOA has been introduced and stalled multiple times. Second, whether the AI carve-out survives committee pressure, particularly from senators who may view it as an unacceptable loophole. Third, how the FTC and DOJ respond to a bill that would explicitly limit their authority over AI model design, that’s a jurisdictional question with implications for how federal agencies think about AI oversight more broadly.
What to Watch
The real question is whether the 2026 legislative calendar gives AICOA a path it didn’t have in prior sessions. The bill’s bipartisan structure is the same as before. The AI carve-out is new, and it’s designed to neutralize one of the strongest industry objections to prior versions. Whether that’s enough to move the bill or just enough to restart the same stalling pattern is what compliance and policy teams should be tracking.