AI Momentum 107.6 +0.8%GitHub Pulse 5216 -10.2%Capital Flow $224.8B +3206.4%Hiring Heatmap 0.0 new
Advertisement
Regulation & Policy

Democrats plan a wave of AI bills. The bigger near-term risk for AI companies is sworn testimony

Democrats who take control of both chambers of Congress in January are lining up a wave of AI bills, Reuters reported on Friday. For AI companies and their investors, the more immediate change is not new law. It is the power to summon executives and demand documents.

Three priorities have emerged. The first is shielding households from electricity costs tied to data centers. The second is requiring independent testing of the most advanced AI systems. The third is spelling out when AI makers are legally liable for harm their products cause.

The obstacle is plain. President Trump holds a veto until January 2029 and has repeatedly dismissed calls for mandatory AI rules. Overriding a veto takes two-thirds of both chambers. Most Senate bills also need 60 votes to get past a filibuster.

The track record is thin too. About 150 AI bills were proposed during Trump's second term, according to IBTimes. Only one became law: a ban on posting nonconsensual sexual images, including AI-generated ones.

What could actually pass

Bills that already have members of both parties behind them stand the best chance. Common Sense Media, a children's advocacy group, says the new Congress could revisit two. The GUARD Act targets risks to children from AI companion chatbots. The AI LEAD Act would set a liability standard for AI systems and let individuals and state attorneys general sue AI companies.

Independent testing has a Senate champion in Maria Cantwell, the top Democrat on the Commerce Committee, which oversees tech. She is pushing for mandatory testing of AI models before release, Semafor reported. In the House, Rep. Lori Trahan plans bills requiring outside experts to check leading models against safety standards, her deputy chief of staff told Reuters.

IBTimes reports that a Trahan draft would let people sue developers when AI causes harm that would be illegal if a person had done it. It would leave state laws in place.

One policy analyst warns against writing all of this off. "The probability of legislation is much higher than people anticipate," said Ed Mills of Raymond James. He told Reuters that Democrats could attach AI rules to routine annual bills. That forces a president to weigh a veto against unrelated priorities. Mills added that a crisis could build enough support to overcome White House opposition.

Oversight is the nearer-term cost

Committee control brings subpoena power, and Democrats are signaling they will use it. Rep. Greg Casar of Texas said AI chief executives "should be called to testify under oath." Rep. Ted Lieu of California promised public hearings and "reasonable guardrails" on AI.

The groundwork is already laid. In October, Rep. Zoe Lofgren, who is seeking to chair the House Science Committee, joined other Democratic committee leaders in questioning OpenAI, Anthropic, Meta, Google and SpaceX. They asked about security practices and about AI tools that behaved in ways their designers did not intend. IBTimes reports that House Democratic committee leaders also asked the five companies to report safety incidents to Congress within 48 hours of detection.

Sworn testimony and document demands carry real costs even without a bill. They can surface internal safety records, set headlines and feed state attorneys general and private lawsuits. Companies raising large amounts of capital will be disclosing risks while those hearings unfold.

The states, and the power bill

The industry's main ask under the Republican Congress was federal preemption: one national rule that overrides state AI laws. That now looks less likely, according to a Foley & Lardner analysis cited by IBTimes. The firm expects California, New York, Colorado, Texas and Illinois to set the practical compliance standard.

Dallas Dolen, who leads PwC's technology, media and telecom practice, told Fox Business that companies face a patchwork of state rules. He said they cannot pause major AI investments while those rules take shape.

Data centers are where federal and state pressure meet. Chuck Schumer, expected to become Senate majority leader, named protecting consumers from data-center energy costs as a priority for next year. Democratic tax writers are weighing a rollback of bonus depreciation for AI infrastructure, IBTimes reports. That tax break lets companies write off equipment costs faster.

UBS strategists expect the states, more than Congress, to act on data centers, Fox Business reported. Possible construction pauses in states such as Texas, where Democrats won the governorship and the state House, could put about 30% of expected 2028 US data-center additions at risk, they said. UBS also expects data centers to pay more of the cost of grid upgrades.

Investors are already pricing some of this in. AI chip and data-center stocks fell sharply on Thursday, with analysts pointing in part to political resistance to new AI infrastructure. Reuters also cited doubts about whether heavy AI spending will pay off, so the election was not the only factor.

The new majority is not united. Semafor reports a split between progressives and business-friendly moderates, and between the House and Senate. Rep. Greg Stanton of the centrist New Democrat Coalition said his caucus has not agreed on a specific policy. A Democratic staffer doubted both chambers would back Casar's call to pause advanced AI development.

Investors should watch three things: who wins the key committee chairs, who is called to the first hearings and whether AI provisions ride on must-pass bills. State Street Investment Management still expects AI investment to stay strong in 2027, though it flagged tighter credit and supply-chain strain.