Democrats win Congress: what changes for AI rules, privacy and advertising

For most of the past two years, the rules for American advertising and artificial intelligence were being written on the assumption that one party would control Washington until at least 2027. Republicans held the White House, the Senate and the House. They used that position to pursue something the industry had sought for a decade: a single federal standard that would sweep away the patchwork of state laws on data privacy and, more recently, on AI. A comprehensive privacy bill built on that idea reached a House hearing in June. An executive order told the Justice Department to sue states over their AI laws. A White House framework asked Congress to stop states from regulating AI development at all.

On Tuesday, November 3, voters changed the arithmetic. The Associated Press called the Senate for the Democrats late on Tuesday night after Josh Turek, a Democratic state representative, beat the Republican congresswoman Ashley Hinson in Iowa, Forbes reported in its live coverage, updated shortly after midnight Eastern time on November 4. Democrats were projected to have reached the 218 seats needed for a House majority, with NBC News projecting a final total of as many as 244. Hakeem Jeffries, the House Democratic leader, is in line to become speaker when the new Congress convenes on January 3, 2027.

Two maps tell the story of the night more simply than any tally. One, drawn after the 2024 presidential election, shows the shift in each county's margin since 2020 as a field of red arrows, leaning right almost everywhere from Montana to Florida. The other, built from counties that had reported almost all of their votes by early Wednesday, compares the 2026 result with the 2024 presidential vote. Its arrows are blue, and they are thickest in places few strategists had expected to matter: Iowa, Kansas and Oklahoma, Kentucky, Tennessee and West Virginia, the Carolinas and Georgia, Louisiana and Mississippi, much of Texas. Large parts of the West and the Northeast were still blank when the second map was captured, because their counts were incomplete. Even so, the direction of travel in 2026 is the mirror image of 2024.

For an industry that spends much of its time on attribution, the obvious question is what this result changes. The honest answer is narrower than the scale of the swing suggests, and more important than it first appears. Democrats will not be able to write the rules for advertising and AI on their own. President Donald Trump keeps his veto, and the Senate keeps the 60-vote threshold for most legislation. What the result does is close the path that Republicans were building toward federal preemption of state law, at least for the next two years, and hand Democrats the subpoenas, the committee gavels and the confirmation votes that shape how federal agencies behave. The likeliest consequence for advertisers, publishers and AI companies is not a new federal regime. It is the survival, and probable expansion, of the state-by-state regime that the industry spent 2025 and 2026 trying to replace.

This special edition sets out how that happened, what Democratic lawmakers have said and done in recent months, and where the lines now fall on AI rules, privacy, children's safety, the Federal Trade Commission, the Federal Communications Commission, antitrust and political advertising itself.

What voters decided

The results were still incomplete when this edition was prepared, and several figures will move as states finish counting. What can be said with confidence is the following.

In the Senate, Republicans entered election day with a 53-47 majority. Democrats needed a net gain of four seats. By midnight Eastern time they had taken several that most forecasters had rated as leaning Republican. In Iowa, Turek beat Hinson. In Kansas, the Reverend Adam Hamilton defeated Senator Roger Marshall, which Forbes described as the first time Kansans had elected a Democrat to the Senate since 1932. In Ohio, the former senator Sherrod Brown beat Senator Jon Husted. In North Carolina, the former governor Roy Cooper won the seat left open by the retirement of Thom Tillis, defeating Michael Whatley. In South Carolina, Annie Andrews, a Democrat, beat Senator Darline Graham. Forbes also reported that James Talarico had flipped Texas, beating Ken Paxton, the state's attorney general, for the seat John Cornyn lost in the Republican primary runoff in May. In Maine, Susan Collins, a five-term Republican, conceded to her Democratic challenger, Troy Jackson.

Democrats also held seats they had been defending. Jon Ossoff won re-election in Georgia with 64% of the vote to 36% for Mike Collins when the AP called the race. Chris Pappas kept New Hampshire's open seat in Democratic hands by beating John E. Sununu. Republicans held Florida, where Senator Ashley Moody led Angie Nixon by 51% to 48% with 96% of the vote counted. Alaska, where Senator Dan Sullivan faced Mary Peltola, and the open seat in Michigan had not been reported as called in the coverage reviewed for this edition.

In the House, Republicans had held 218 or 219 seats, depending on how an independent member was counted, against 214 for Democrats, with two vacancies. Democrats needed a net gain of three or four seats. On NBC News projections, they comfortably exceeded it. Some of the flips came in districts that Republicans had redrawn in the middle of the decade to protect their majority: CBS News reported on election day that Republicans in seven states had tried to gain an advantage in 15 additional districts through new maps, and that California's Democrats had answered by redrawing five districts. Democrats flipped North Carolina's 11th district and Colorado's 8th, among others.

Governors' races added a state dimension that matters a great deal for the rest of this story. In California, the former health secretary Xavier Becerra won the governorship, defeating the Republican Steve Hilton. Becerra is also a former attorney general of California, the office that enforced the state's consumer privacy law before the California Privacy Protection Agency took on that role. In New York, Kathy Hochul was re-elected, which keeps in office the governor who has signed several of the country's most closely watched AI laws. Josh Shapiro in Pennsylvania, JB Pritzker in Illinois and Wes Moore in Maryland were re-elected. In Florida, the Republican congressman Byron Donalds beat David Jolly. In Texas, the Democrat Gina Hinojosa led Governor Greg Abbott by 57.2% to 41.6% with 54% of ballots counted on CBS News estimates, but no call had been made.

Why did voters move so far? The exit polls point to prices and to the president. An early CBS News exit poll put Trump's approval at 38%, on track to be the lowest for any president at a midterm, with about three-quarters of voters at least dissatisfied with the direction of the country. CNN's exit poll, as summarised by Forbes, found about 90% of voters calling the cost of living a crucial or major factor, and about 60% disapproving of Trump's performance, including about one-fifth of those who had voted for him in 2024. In AP VoteCast, about two-thirds said Trump's policies had made the economy worse.

Jeffries chose his words with that in mind. "We'll cross that bridge when we get to it," he said when asked about his relationship with the president, in remarks reported by MassLive late on election night. "But we've made clear, Democrats, we're relentlessly focused on driving down the high cost of living and fighting for an affordable America." He added that Democrats were "willing to work with the other side of the aisle whenever and wherever possible to deliver real results." On impeachment, he told MS NOW: "We haven't ruled anything out when it comes to all of the tools available to us to hold people accountable. And we haven't ruled anything in." Robert Garcia, the Californian expected to chair the House Oversight Committee, offered a shorter message to the administration: "Lawyer up," ANI reported.

Neither Jeffries nor Garcia mentioned artificial intelligence or advertising on election night. That is itself informative. The campaign was fought on prices, on health care, on the war with Iran and on the president. Technology entered it mainly through electricity bills and data centres, which is where the next section of this story begins. But the policy consequences for technology will be felt sooner than many of the headline issues, because so much of the federal technology agenda of 2025 and 2026 depended on a Congress that will no longer exist after January 3.

What a majority can do, and what it cannot

Before turning to the specific fights, it is worth being precise about the limits of the new majority, because the gap between a wave and a law is wide.

First, the president's veto. Any bill that reaches Trump's desk can be vetoed, and overriding a veto requires two-thirds of both chambers. Even on optimistic counts, Democrats will be far short of that. Jeremy Kahn, writing in Fortune's Eye on AI newsletter on the morning of the election, put it plainly: "any legislation that Congress does pass faces a likely veto from President Trump, who has been steadfast in his opposition to AI regulation." The result, he argued, is that new AI law "will still require a bipartisan bill with very broad support."

Second, the filibuster. A Senate majority in the low fifties does not reach the 60 votes needed to end debate on most legislation. Bills that Democrats would like to pass on privacy or AI will still need Republican votes in the Senate, just as Republican bills needed Democratic votes in the outgoing Congress. The SECURE Data Act, the House Republican privacy bill discussed below, never acquired a Senate companion.

Third, the calendar. The outgoing Congress, with its Republican majorities, returns for a lame-duck session before January 3. Anything that can pass in those weeks, on unified Republican control and with Democratic votes where the Senate rules require them, will become law under the old arithmetic. That makes the next eight weeks unusually important for every bill discussed in this edition.

What, then, does a majority give Democrats? It gives them the power to decide which bills reach the floor in the House, and to block those they oppose. It gives them committee chairs, and with them the power to hold hearings, to subpoena documents and witnesses, and to publish reports. It gives the Senate majority control over the confirmation of nominees, including to the five-seat Federal Trade Commission, which had three vacant seats in late August. And it gives both chambers a decisive role in appropriations, where the budgets of the FTC, the FCC and the Justice Department's Antitrust Division are set.

The Brookings Institution, in a collection of essays published on October 22, framed the coming two years in the same terms: "If the blue wave that polls predict materializes, efforts to enact and change the policy agenda will play out against a backdrop of government divided along political fault lines." Divided government tends to produce fewer laws and more investigations. For an industry that had been told to expect a single federal rulebook, that is a significant change of forecast.

Who will hold the gavels

Committee assignments for the new Congress will not be settled until the parties organise in December and January, and seniority is a convention rather than a rule. But the people now in line for the posts that matter most to advertising and AI have long public records, and those records are the best available guide to what the next two years will look like.

In the Senate, the Commerce, Science and Transportation Committee handles privacy, the FTC, the FCC and most AI legislation. Its chairman until January is Ted Cruz, who has championed the White House framework, a light-touch "sandbox" approach to AI and federal preemption. Its most senior Democrat is Maria Cantwell of Washington, who chaired the committee in the 118th Congress and co-wrote the American Privacy Rights Act. If she takes the gavel again, the committee's agenda will move from pre-empting states towards the "bipartisan national AI framework" she has called for, with consumer protections as the price of any federal standard. Cruz and his Republican colleagues, in the minority, will still have the votes to stop anything they regard as overreach, because the filibuster protects them as it protected Democrats.

In the House, the Energy and Commerce Committee is the gatekeeper for privacy, children's safety and much of technology. Brett Guthrie of Kentucky chairs it now; Frank Pallone of New Jersey, its ranking Democrat, was elected its chairman when Democrats last held the House. Pallone has been a leading Democratic critic of the SECURE Data Act, the co-author of the bipartisan KIDS Act compromise and an ex officio co-chair of Jeffries' AI commission. His record suggests a committee that will hold hearings on data brokers and AI harms, push a federal floor for children's privacy, and refuse any privacy bill that overrides California.

Oversight will matter as much as legislation. Robert Garcia, who is expected to chair the House Oversight Committee, has signalled his intentions in two words. Zoe Lofgren is the ranking Democrat on the Science, Space and Technology Committee, which oversees federal AI research and standards. If Jeffries follows through on the proposal from Lieu and Foster, a select committee on AI with subpoena power would sit beside all of them.

Some familiar names will be elsewhere. Amy Klobuchar, for years the Senate's most prominent Democratic voice on technology competition, was the Democratic nominee for governor of Minnesota this year, and would leave the Senate if elected. Dick Durbin, the senior Democrat on the Senate Judiciary Committee, which oversees the Justice Department and antitrust, is retiring at the end of his term, and Illinois elected his successor on Tuesday. Elizabeth Warren, Ed Markey and Richard Blumenthal, the senators who led Democratic opposition to AI preemption and the push for KOSA, will be in the majority for the first time since January 2025.

At the top, Chuck Schumer, the Senate Democratic leader, is expected to lead the new majority, and Jeffries is in line to become speaker. Schumer led the Senate's AI "insight forums" in 2023 and an AI policy road map in 2024, which favoured investment over regulation. Jeffries, as his September and October remarks showed, has moved further towards enforceable rules. How those two instincts are reconciled will decide whether Democrats try to legislate on AI in 2027 or concentrate on investigation.

The preemption project, from 99 votes to an executive order

The single most consequential technology question the new majority inherits is preemption: whether Washington should stop states from writing their own AI rules. It has dominated the past 16 months, and the record shows how much of the Republican strategy relied on unified control.

The first attempt came through the budget. In the spring of 2025, Republicans added to their reconciliation bill a 10-year moratorium on state AI regulation, tied to federal broadband money from the Broadband Equity, Access, and Deployment programme. It did not survive the Senate. On July 1, 2025, senators voted 99 to 1 to strip it out, on an amendment co-sponsored by Maria Cantwell, the ranking Democrat on the Commerce Committee, and Marsha Blackburn, a Tennessee Republican. The provision, the Democratic committee staff wrote, "would have forced states to make an impossible choice between enforcing AI consumer protections or accepting federal BEAD funding to expand broadband access." Cantwell's reaction set the terms that Democrats have used ever since. "The Senate came together tonight to say that we can't just run over good state consumer protection laws," she said. "States can fight robocalls, deepfakes and provide safe autonomous vehicle laws."

The second attempt was the annual defence bill. In November 2025, the administration and House Republican leaders sought language in the National Defense Authorization Act that would have overridden state AI laws. Senate Democrats prepared to unite against it, Bloomberg Government reported on November 20, 2025. That day Ed Markey and Elizabeth Warren sent their colleagues a letter that asked them to vote against the whole defence bill if the moratorium was in it. "Congress should not give Big Tech a multi-year holiday from state oversight," they wrote. The provision did not make it into the final bill.

Having failed twice in Congress, the administration turned to executive power. On December 11, 2025, Trump signed an executive order titled "Ensuring a National Policy Framework for Artificial Intelligence", Tech Policy Press reported the next day. It instructed the attorney general to create an AI Litigation Task Force to challenge state AI laws "including on grounds that such laws unconstitutionally regulate interstate commerce, are preempted by existing Federal regulations" or are otherwise unlawful. It told the Commerce Department to draw up a list of "onerous" state laws, singling out those that "require AI models to alter their truthful outputs." It tied BEAD funds to an evaluation of each state's AI "regulatory landscape". It asked the chair of the FCC to "initiate a proceeding to determine whether to adopt a Federal reporting and disclosure standard for AI models that preempts conflicting State laws," and the FTC to issue a policy statement on when state laws requiring changes to AI outputs are preempted by the FTC Act. It also carved out several areas, including "child safety protections" and AI data centre infrastructure, from the legislative recommendation it commissioned.

David Sacks, the White House adviser on AI and crypto, explained the reasoning at the signing. "We have over 1,000 bills going through state legislatures to regulate AI," he said. "Over 100 of them have already passed." Tech Policy Press noted that the 1,000 figure was disputed. Cody Venzke of the American Civil Liberties Union said the order "doubles down on a dangerous policy that the Republican-led Congress has rejected not once, but twice."

Democrats responded quickly. Cantwell said on December 12 that the order's "overly broad preemption threatens states with lawsuits and funding cuts for protecting their residents from AI-powered frauds, scams, and deepfakes." Her proposed alternative is worth noting, because it is not a rejection of federal rules: "Let's get it right and pass a bipartisan national AI framework that both leads on innovation and protects consumers." Markey introduced a bill, the States' Right to Regulate AI Act, to block the order.

Then came the legislative blueprint. On March 20, 2026, the White House published "A National Policy Framework for Artificial Intelligence", which PPC Land examined on March 22 under the headline that the White House AI framework targets state laws, child safety and copyright. Its central claim was blunt: "States should not be permitted to regulate AI development, because it is an inherently interstate phenomenon with key foreign policy and national security implications." It asked Congress to affirm "that existing child privacy protections apply to AI systems, including limits on data collection for model training and targeted advertising," to require age assurance, to leave the copyright question on training data to the courts, and not to create "any new federal rulemaking body to regulate AI."

Jakub Kraus, writing in Lawfare on April 10, described it as a document that proposes industry-friendly legislation and called preemption "the most contentious part" of it, while noting a meaningful carve-out for local decisions on where AI infrastructure is placed. Speaker Mike Johnson and Ted Cruz, the Texas Republican who chairs the Senate Commerce Committee, backed the framework. Cantwell's response was more measured than her December statement: she said it "identifies key areas to address." Markey's was not. "President Trump is continuing to repay Big Tech's campaign donations by proposing to block states from protecting their communities from AI-related harms," he said on March 20. "Congress must not only reject Trump's offensive proposal but also continue blocking Republicans' bad faith effort to undermine states."

The difference between those two statements is a useful guide to the next Congress. Cantwell, who chaired the Commerce Committee when Democrats last held the Senate and is its most senior Democrat, has consistently said she wants a national framework, as long as it protects consumers and does not simply wipe out state law. Markey and Warren want states left free to act. Both positions are incompatible with the White House framework as written. Neither guarantees a federal law.

Colorado, the test case

The preemption fight has not been abstract. Colorado became its proving ground, and the sequence of events there shows how federal pressure works when Congress does not act.

Colorado's AI Act, SB 24-205, was signed on May 17, 2024, and was meant to take effect on February 1, 2026, before a delay to June 30, 2026. It required developers and deployers of "high-risk" AI systems to guard against algorithmic discrimination. On April 9, 2026, Elon Musk's xAI sued Colorado to block the law, PPC Land reported the next day, arguing in the US District Court for the District of Colorado that developing an AI model is expressive activity and that compliance would compel speech. The complaint, against the state's attorney general, Philip Weiser, raised First Amendment, dormant Commerce Clause, due process and equal protection claims.

The federal government joined in. The Justice Department filed a complaint in intervention on April 24. Three days later, on April 27, the court ordered Weiser not to enforce SB 24-205 or any amending legislation until rulemaking had concluded and the court had ruled. Colorado's legislature, meanwhile, had rewritten the law. SB 26-189, which the state House passed by 57 votes to 6 on May 9, repealed and replaced the original, gave the attorney general exclusive enforcement power, capped civil penalties at $20,000 per violation and set a new effective date of January 1, 2027.

Then the FTC added its own weight. On July 1, the commission released a nine-page proposed policy statement "Concerning the Suppression of Accuracy in Artificial Intelligence Systems", approved by a vote of 2-0, and PPC Land reported on July 7 that the FTC move could force Colorado to rewrite its new AI bias law again. The statement argued that undisclosed steering of AI outputs towards ideological objectives may violate Section 5 of the FTC Act, and named Colorado's amended law in a footnote as one that "poses many of the same concerns" as the original.

Colorado is governed by Democrats, and its own Democratic leaders had doubts about the law. When Governor Jared Polis signed the original bill, he did so with "reservations"; in June 2024 he, Weiser and the state Senate majority leader warned that "a state-by-state patchwork of regulation poses significant challenges to the cultivation of a strong technology sector." That matters for reading the next two years. Democrats are not united behind aggressive state AI regulation. What unites them, from Polis to Markey, is resistance to having the federal government decide unilaterally that states may not act.

The election does not end the Colorado litigation, the executive order or the FTC's statement. Those are tools of the executive branch and the courts, and they remain in place. What it changes is the endgame. The administration's lawyers have argued that existing federal law already preempts state AI rules. A federal statute would have settled the question. Without a Republican Congress, that statute is now unlikely before 2029.

What Democrats have said about AI

If Democrats were only against preemption, their AI agenda would be easy to describe. It is more complicated than that, and it has moved during 2026 from caution towards something closer to alarm.

On December 9, 2025, Jeffries announced a House Democratic Commission on AI and the Innovation Economy, to "convene throughout 2026" and to "develop policy expertise in partnership with the innovation community, relevant stakeholders and committees of jurisdiction." Its co-chairs were Ted Lieu of California, the caucus vice chair; Josh Gottheimer of New Jersey; and Valerie Foushee of North Carolina, with Zoe Lofgren and Frank Pallone, the ranking Democrats on the Science and Energy and Commerce committees, as ex officio co-chairs. "Artificial intelligence is advancing at a breakneck pace, and Congress must keep up," Lieu said. Foushee said she was "committed to advancing clear, enforceable rules of the road." Pallone, who would be expected to chair Energy and Commerce, the House committee with jurisdiction over privacy and much of technology, used the announcement to restate the party's line on preemption: "Rather than blanket moratoriums, we should be learning from the work of our state and local counterparts." Gottheimer, for his part, stressed keeping the US "the leader in AI-fueled jobs and innovation."

That balance shifted in the autumn. On September 14, Jeffries argued that the House should stay in Washington rather than leave for a six-week pre-election recess until it had passed legislation to regulate AI. "It requires decisive congressional action immediately, in a manner consistent with what some of the leading AI voices in the country are now saying, needs to happen," he told reporters, Bloomberg Government reported. The House left anyway.

On October 19, The Verge reported that House Democrats were preparing an AI oversight push if they won, and that Jeffries was weighing a select committee on AI. He described the country as facing an "AI safety crisis". Bill Foster of Illinois and Lieu had pitched him, that autumn, a House Select Committee on AI "with subpoena power", modelled on the House Select Committee on China. Jeffries cited reports of AI agents built on OpenAI models that had named themselves "The Collective" and "conspired against Hugging Face", and asked: "Who does that?" He was equally direct about voluntary commitments by the industry: "The notion of entering into a voluntary agreement that is entirely unenforceable, and that allows the industry writ large to police itself, is not the right response at this moment."

The caucus is not of one mind on how to respond. Debbie Dingell of Michigan has said existing committees are adequate. Ro Khanna of California wants a federal AI safety agency modelled on the Food and Drug Administration. Further left, Bernie Sanders, the Vermont independent who caucuses with Democrats, and Alexandria Ocasio-Cortez of New York introduced on March 25 the AI Data Center Moratorium Act of 2026, which would impose "an immediate federal moratorium on AI data centers" until national safeguards on safety, workers' gains, electricity prices and the environment were in place. "We need a federal moratorium on AI data centers," Sanders said. In September he went further with a bill to ban the development of artificial superintelligence; Jeffries declined to commit to it.

Other Democratic work has been narrower and bipartisan. In July, Foushee and Laurel Lee, a Florida Republican, introduced the Stealth Bot Prohibition Act, which MediaPost reported on July 23. It would require AI-powered web crawlers to disclose their identity and purpose when they access websites, with enforcement by the FTC. Lee said that "website operators are too often left in the dark about who is accessing their systems." Danielle Coffey, chief executive of the News/Media Alliance, called it a "sorely-needed, common-sense solution", and her organisation estimated that deceptive bots "represent more than 50% of the traffic on some publisher sites." For publishers whose content is being taken by AI systems without payment, this kind of narrow, enforceable bill may matter more than any grand framework.

Kahn's Fortune newsletter noted that, before the election, the bipartisan Frontier Act looked like the AI bill with the best odds of passing, and that Democrats in the majority "are far more likely to try to enact some kind of federal AI regulation." A 37-page discussion draft that circulated in October, which PPC Land examined on October 20, offered one version of what that could look like: incident reporting within 15 days, court-ordered suspensions of models for 45 days and preemption limited to developers' safety disclosure, evaluation and reporting obligations, leaving states free to regulate uses, consumer protection, child safety, civil rights and privacy. Politico Pro linked the draft to Anthropic, but the company's name does not appear in it, and PPC Land could not confirm who wrote it.

Oversight is the most certain consequence. Fortune pointed to Warren's letter to Dario Amodei, Anthropic's chief executive, demanding answers about risks disclosed in the company's IPO prospectus, as an early example of the investigations to come. With committee gavels and subpoena power from January, Democrats will be able to compel testimony from AI companies rather than request it.

The money that tried to shape the argument

The AI industry did not wait for Congress. In 2025 and 2026 it built a political machine of its own, and the midterms were its first test. The results suggest that money bought less than its donors hoped.

Two networks of super PACs dominated. Leading the Future, backed among others by Greg and Anna Brockman, who gave it $25 million in 2025, argued against what it described as restrictive regulation. Public First, whose donors include Amodei and whose nonprofit affiliate received money from Anthropic, backed candidates who favour safety rules. AdExchanger's daily roundup on June 24 noted that AI-funded PACs were among the biggest political ad spenders of the cycle, citing Federal Election Commission data showing that PACs associated with OpenAI and Anthropic had spent more than $38 million between January 1 and June 15 to support candidates from both parties.

The most expensive single contest was a Democratic primary in Manhattan. Alex Bores, a New York state assemblyman who had sponsored the state's RAISE Act on frontier AI safety, ran for the open 12th district seat. Leading the Future spent more than $8 million to defeat him; pro-Bores groups, chiefly the Public First-linked Jobs and Democracy PAC, spent more than $19 million to support him. On June 23, Bores lost to Micah Lasher by 39% to 35%, Transformer reported in its account of what Bores' defeat says about AI politics, noting that the race drew more than $27 million from AI-related groups. Lasher's campaign had received at least $10 million from Michael Bloomberg. In his concession, Bores blamed a "handful of oligarchs hellbent on preventing any regulation of their industry whatsoever."

The general election told a different story. The Intercept reported on October 29 that the AI lobby learned that voters dislike AI money. Leading the Future and Public First had put $50.7 million into the primaries, about $21.3 million of it on NY-12, yet by October 26 Leading the Future had spent only $7.4 million on general election races despite about $34 million in cash across its network, and Public First had spent $1.1 million. Leading the Future's general election spending went mainly to support Darline Graham in South Carolina. She lost to Annie Andrews on Tuesday.

Polling explains the retreat. A Data for Progress survey of more than 3,000 likely voters in Iowa, Michigan and Ohio, commissioned by the anti-industry Guardrails Alliance, found that about three-quarters of voters were less likely to support a candidate they knew had received support from an AI super PAC. "Essentially these companies underestimated how much the public despises the data center buildout," said Rick Claypool of Public Citizen. The poll was commissioned by an advocacy group with a stake in the answer, and should be read in that light. But it is consistent with what the industry itself did: it stopped spending where its name would appear on the ballot.

For the advertising market, this money was a modest but real line in a record cycle, and its political lesson is the more durable one. The candidates who take office in January have watched an industry spend tens of millions of dollars and then decide that its support was a liability. That changes the cost of voting against it.

Data centres on the ballot

If AI came up on doorsteps in 2026, it usually came up as a building. Data centres, and what they do to local electricity bills, water supplies and land, became the most concrete form of the AI debate for most voters.

Fortune counted at least 41 ballot initiatives involving data centres on Tuesday, citing Ballotpedia. Ohio had the most, with 21 measures, 11 of which would ban construction or force a pause on new projects. In Piqua, Ohio, where Meta plans to spend $1 billion on a data centre with a capacity of up to 1.4 gigawatts, voters were asked whether to recall the mayor and two city commissioners who had supported the project. Amazon Web Services projects worth $2 billion in Sunbury and $10 billion in Adams County were also before Ohio voters. Data Center Watch, a research project, counted at least 45 data centre projects worth about $68 billion disrupted by local opposition in the second quarter of 2026 alone. Results for most of these local measures were not available at the time of writing.

Forbes, in its election-night coverage, singled out data centres as one of the issues candidates handled most warily, with voters citing electricity bills, power use and water consumption. The issue does not divide neatly by party. Democrats such as Sanders want a national pause; Republican and Democratic local officials alike have blocked projects; and the White House framework, for all its hostility to state AI rules, explicitly left decisions on "the placement of AI infrastructure" to state and local governments.

For advertisers, this may look remote. It is not. The largest advertising platforms are now also the largest builders of AI infrastructure, and their capital spending plans rest on assumptions about permits, power prices and political tolerance. PPC Land reported on October 31 that S&P Global Ratings had moved Meta's outlook to negative as its forecast for Meta's 2027 capital spending topped $225 billion, with data centre leases treated as debt-like obligations. If local voters slow the buildout, or if a Democratic Congress makes electricity prices a subject of investigation, the cost and pace of the AI products these companies sell to advertisers will be affected.

What voters, and advertisers, think of AI

The election was not a referendum on artificial intelligence, and none of the exit polls cited here asked about it directly. But the surveys published during the campaign point in one direction, and it is not the one the industry would prefer.

MediaPost reported on October 2 that most US voters want strong AI regulation. Fortune's election-day newsletter cited Pew Research Center findings that a majority of Americans are more fearful than hopeful about AI. The Data for Progress poll commissioned by the Guardrails Alliance found voters more concerned about spending by AI companies than by crypto companies. None of these surveys is decisive on its own, and two were commissioned or reported by parties with a view. Taken together, they help explain why candidates of both parties spent the autumn distancing themselves from data centres and AI money rather than embracing them.

Advertising has its own version of the trust problem. PPC Land reported on April 17 that the Video Advertising Bureau, which represents the television industry, found TV news beats social media and AI on trust among high-value audiences: among households earning $100,000 or more, 75% trusted television news, 6% social media and 5% AI tools, and 92% expressed some concern about AI in media. The VAB has an obvious interest in that conclusion, and its survey of 2,319 adults was fielded in December 2025. But the direction matches the political polling.

The industry's answer so far has been voluntary disclosure. On August 18 the Interactive Advertising Bureau published version 2 of its AI Transparency and Disclosure Framework, which PPC Land covered the next day under the headline that AI ad labels cut click-through by 31.5%, as the IAB framework cites an NYU study. The framework's core test is whether "the AI involvement create[s] a material risk that consumers will be misled". It requires labels for prompt-generated images and video and for synthetic voices of deceased people, and exempts most AI-written copy. It "carries no enforcement authority." Caroline Giegerich, the IAB's vice president for AI, explained the restraint: "not every use of AI needs a label - labeling everything teaches consumers to ignore labels and could negatively impact advertisers." The NYU Stern finding the IAB cited, that disclosing generative AI involvement cut an ad's click-through rate by 31.5%, shows why advertisers would rather not label more than they must.

States are not waiting. New York's synthetic performer disclosure law, signed on December 11, 2025, took effect on June 9, 2026, as MediaPost reported the next day, with civil penalties of $1,000 for a first violation and $5,000 for each one after. California's AI Transparency Act took effect on August 2, 2026, binding AI providers with more than one million monthly users. MediaPost reported on October 9 that Pennsylvania passed a bill requiring disclosure of misleading AI in ads. Under a Republican Congress, the industry could reasonably hope that a federal standard would one day replace these rules with something closer to the IAB's voluntary approach. Under a divided one, the state rules are the standard, and an AI disclosure bill that reached a Democratic House floor would be more likely to add obligations than to remove them.

The privacy bill that ran out of time

For advertisers, the most direct casualty of the election is likely to be the SECURE Data Act, the most serious attempt at a federal privacy law since 2024.

The bill, formally the Securing and Establishing Consumer Uniform Rights and Enforcement over Data Act and numbered H.R. 8413, was introduced on April 21, 2026 by John Joyce, a Pennsylvania Republican and vice chairman of the Energy and Commerce Committee. PPC Land reported the next day that House Republicans had unveiled the SECURE Data Act to replace US state privacy laws, and the word "replace" was accurate. Its Section 15 would stop any state from adopting, maintaining or enforcing any law "having the force and effect of law" relating to its provisions. PPC Land described it as a ceiling, not a floor.

The substance was familiar to anyone who has followed the state laws, because it was modelled on them, and in particular on Kentucky's. Consumers would have the right to opt out of targeted advertising, defined as ads selected using personal data obtained "over time and across nonaffiliated websites or online applications." Contextual advertising, ads responding to a consumer's own request and processing "solely for measuring or reporting advertising or content performance, reach, or frequency" were excluded. Processing sensitive data, including precise geolocation within 1,750 feet, health diagnoses, sexual orientation, immigration status, biometric data and data from children and teens, would require consent. Data brokers, defined as companies that collect data on people who are not their customers and earn at least half of their revenue from selling it, would register with the FTC. Teenagers aged 13 to 15 would need verifiable parental consent before their sensitive data was processed.

The thresholds were higher than in most states. As AdExchanger's Allison Schiff set out on April 24, a company would be covered if it collected data on more than 200,000 consumers a year and earned at least $25 million in revenue, or handled data on at least 100,000 consumers and derived 25% or more of its revenue from data sales. "In places like Connecticut, Delaware and Maryland, the threshold is 35,000," she noted. Enforcement would fall to the FTC, with state attorneys general as back-up after notifying the commission, a 45-day right to cure, and no private right of action. Most provisions would take effect two years after enactment.

The advertising industry welcomed it. The Association of National Advertisers said the bill "charts the best path to a national privacy standard by building on the common bipartisan framework that has already been signed into law by more than 20 states," and cited 29 million US jobs supported by advertising that it said a national policy should not jeopardise, PPC Land reported on April 22. Internet for Growth, a coalition representing small businesses and creators, backed the bill with conditions, PPC Land reported on April 26. "For millions of small businesses, entrepreneurs, and creators, digital advertising is not a luxury - it is the most efficient and affordable way to reach customers, compete with larger companies, and grow," said Brendan Thomas, its executive director, while raising concerns about the national opt-out right for targeted advertising.

Democrats did not. At the bill's legislative hearing on June 3 before the Subcommittee on Commerce, Manufacturing, and Trade, Pallone laid out the case against it in terms that will now shape the next Congress. "The partisan SECURE Data Act is not the strong, enforceable standard its sponsors describe," he said. "Instead, this bill locks in the failed notice and consent status quo, then compounds loophole upon loophole to water down its provisions." He objected to its preemption clause, which he said "will leave many Americans with fewer privacy protections than they have today," and to its origins: "Rather than taking the strongest consumer protections from the existing state privacy laws, this bill is assembled from industry-friendly state privacy laws that have been pushed by Big Tech." Of its data minimisation rules, he said: "This is just 'notice-and-consent' by a different name." And he pointed to a specific consequence: "It would prevent Californians from being able to delete their data from all data brokers in one step."

Pallone was not alone. The IAPP's account of the hearing, published on June 4, reported that the subcommittee's Democrats were united against the bill. Caitriona Fitzgerald of the Electronic Privacy Information Center told the panel that it "sets a national standard that is weaker than the weakest state law, we shouldn't be making the floor, the ceiling." California's privacy regulator joined a coalition of 18 state attorneys general opposing it. "A strong federal privacy law is worth pursuing, but it should not strip away rights that tens of millions of people already depend on," said Tom Kemp, the agency's executive director. Gus Bilirakis, the Florida Republican who chairs the subcommittee, defended the bill as taking "the best ideas of the state privacy laws."

Even before the election, its prospects were uncertain. No companion bill had been introduced in the Senate. AdExchanger's experts split. Gary Kibel of Davis+Gilbert said that "if I had $100, I'd sooner bet it on the Mets winning the World Series than this bill becoming law this year in its current state," while adding that "preemption is desperately needed by the industry." Richart Ruddie of Captain Compliance predicted that the Senate "is where the bill will stall." Jessica Lee of Loeb & Loeb thought it had "a better chance of passing than previous iterations."

What happens now? In the lame-duck session, Republicans could still try to pass it through the House. Without 60 votes in the Senate, it would go no further. From January, the House committee that would handle any privacy bill will be chaired by a Democrat, very probably Pallone, who chaired it when Democrats last held the House. Pallone has said that he has "long supported bipartisan national comprehensive privacy legislation." The question is what kind. The last bipartisan attempt, the American Privacy Rights Act of 2024, was drafted by Cantwell and Cathy McMorris Rodgers, then the Republican chair of Energy and Commerce. It combined federal preemption with a limited private right of action, and it collapsed amid objections from both directions. The positions Democrats took against the SECURE Data Act, on minimisation rather than consent, on a floor rather than a ceiling, and on the right to sue, are the same positions that sank compromise then.

The likeliest outcome for the next two years is therefore no federal privacy law at all. A Democratic House will not pass a bill that preempts California; a Republican president will not sign one that creates a private right of action; and the Senate needs 60 votes for either. For advertisers, that means the compliance map they have today, with its varying thresholds, opt-out signals and sensitive data rules, is the map they will have in 2028.

Fifty states keep writing the rules

If federal preemption is off the table, the centre of gravity stays where it has been: in state capitals and attorneys general's offices. Several of Tuesday's results strengthen that.

California remains the most important. Its Delete Act created the Delete Request and Opt-out Platform, known as DROP, which went live on January 1, 2026. Within eight weeks, PPC Land reported on April 10, more than 242,000 Californians had signed up, and more than 575 data brokers had registered. Since August 1, 2026, brokers have had to retrieve deletion lists from DROP at least every 45 days and process them, with penalties of $200 per day per consumer for failing to do so. This is the one-step deletion right that Pallone said the SECURE Data Act would have pre-empted. California's regulator has also been enforcing: its first Delete Act penalty, announced on December 30, 2025, fined a data broker $45,000 for selling lists built on health conditions.

The state now has a new governor in Becerra, a former state attorney general, taking over from Gavin Newsom. Newsom signed SB 53, California's frontier AI transparency law, in 2025, and in September 2026 issued Executive Order N-9-26, which gave the state's Government Operations Agency until November 16, 2026 to recommend whether frontier AI developers should host independent verification organisations, submit safety frameworks to third-party review and build a kill switch, as PPC Land reported on September 20 when Newsom set a November 16 deadline to study a frontier AI kill switch. The order creates no enforceable obligations; it is a study. That report will now land on the desk of an incoming administration. California has also legislated on automated decisions at work: PPC Land reported on October 15 that the state bans AI-only firing decisions from July 2027, under SB 947, which Newsom approved on September 30 with a $500 penalty per violation.

New York's re-elected governor has been equally active. Hochul signed the RAISE Act on frontier AI safety in December 2025, after amendments that brought it closer to California's SB 53. On October 28 she signed a law that PPC Land reported bars human-posing AI chatbots for under-18s: from January 1, 2027, operators may not offer 12 "unsafe AI companion features", including outputs suggesting that a companion is human or "engage in flattery or sycophancy", to users not verified as adults, with civil penalties of up to $25,000 per violation and possible destruction of "any algorithm trained in such data." New York City, meanwhile, is weighing its own rules: PPC Land reported on October 6 that OpenAI, Anthropic, Google and Meta face 10 proposed New York City AI bills, after the City Council questioned executives from all four under oath on October 5. None of the measures has passed.

The AI disclosure rules for advertising in New York, California and Pennsylvania, described above, add another layer. Colorado's rewritten AI Act is due to take effect on January 1, 2027, if the courts allow it. Texas, where Ken Paxton's office secured some of the largest privacy settlements in the country, will have a new attorney general in January, since Paxton ran for the Senate.

The industry's argument against this patchwork has not changed, and it is not without force. The administration has argued that "a patchwork of conflicting state laws would undermine American innovation." At the June hearing, Ashli Watts of the Kentucky Chamber of Commerce cited a Chamber report finding that roughly two-thirds of small businesses had compliance concerns, a 14% increase from 2024. Those costs are real. But the election has made them a cost of doing business rather than a temporary condition awaiting a federal fix. Compliance teams that had planned around a single federal standard arriving in 2027 or 2028 will have to plan around 20 or more state regimes instead, and around state attorneys general who have shown, in the 18-state coalition against the SECURE Data Act, that they are prepared to coordinate.

Children online, the one area of agreement

There is one area where a federal law is plausible, in the lame duck or in the next Congress: children's online safety. It is also an area with direct consequences for advertising to minors.

The House has already acted, though not in the form the Senate wants. The Kids Internet and Digital Safety Act, H.R. 7757, bundled more than a dozen bills, including a version of the Kids Online Safety Act and an update of the Children's Online Privacy Protection Act. The Energy and Commerce Committee reported it on March 5 by 28 votes to 24, without Democratic support, alongside the App Store Accountability Act, which passed committee by 26 to 23. The March text would have pre-empted state children's safety laws.

In June that changed. Brett Guthrie, the committee's Republican chairman, and Pallone struck a bipartisan deal, the Washington Times reported on June 22, that would set a federal floor for children's online safety and privacy that states could exceed, not a ceiling that would wipe them out. It added Pallone's proposal for a federal data broker registry. It did not include the "duty of care" at the heart of the Senate's version of KOSA. "Through empowering parents, establishing safety as a default, strengthening privacy for children and teens, increasing transparency around data brokers, and holding Big Tech accountable, the KIDS Act delivers the 21st-century protections parents have demanded and our kids deserve," Guthrie and Pallone said in a joint statement.

When the House passed the package, its Senate champions were scathing. "Instead of passing a federal standard that will protect America's children from Big Tech's greed, the House of Representatives just passed legislation that is a pale imitation of Big Tech accountability," Blackburn and Richard Blumenthal, the Connecticut Democrat, said on June 29. "The House's decision to strip out the Duty of Care will only maintain the status quo in which Big Tech uses our kids as profit centers and leaves parents to pick up the pieces." Their version of KOSA has 75 co-sponsors and passed the Senate by 91 votes to 3 in the previous Congress. On August 5, the Senate Commerce Committee approved KOSA by unanimous voice vote, sending it to the floor. The Senate had already passed its own COPPA 2.0 by unanimous consent earlier in the year.

The FTC has acted in parallel. On February 25 it issued an enforcement policy statement that PPC Land reported gives age verification technology a COPPA enforcement shield: operators that collect data only to determine a user's age, and then delete it, will not face COPPA action for doing so without parental consent. The White House AI framework, for its part, asked Congress to apply child privacy protections to AI systems, "including limits on data collection for model training and targeted advertising."

This is the one technology issue on which the president, both parties' leaders and most of the advertising industry say they agree. Ted Cruz, who chairs Senate Commerce until January, told a rally of parents in May: "If we do nothing else in Washington, we need to be protecting our kids." The differences are about preemption and liability, not about whether to act. That makes children's safety the most likely candidate for a deal in the lame duck, where Republicans control the floor schedule and Democrats have every incentive to lock in a floor before the next Congress begins. If it does not happen by January 3, a Democratic House and Senate would be likely to push for the Blumenthal and Blackburn version, with its duty of care, and to keep state laws intact. Either way, the direction for advertisers is the same: fewer ways to target users under 17, more age assurance, and more liability for design.

The FTC after Humphrey's Executor

For most of the past century, the Federal Trade Commission was designed to sit at arm's length from the White House. That design ended on June 29, 2026, and the election cannot restore it.

In Trump v. Slaughter, the Supreme Court held by six votes to three that the FTC Act's protection of commissioners from removal except for "inefficiency, neglect of duty, or malfeasance in office" was unconstitutional, overruling Humphrey's Executor, the 1935 decision that had upheld it. Chief Justice John Roberts, writing for the majority, said the case "is not a close case" and described agency heads as "envoys of the President, not his equals." The Court left open whether removal protections might survive for bodies that perform functions traditionally outside the executive branch, identifying the Federal Reserve as one possible example. Justice Sonia Sotomayor, joined by Justices Elena Kagan and Ketanji Brown Jackson, called the decision a "profound bait and switch." The ruling affects not only the FTC but dozens of independent agencies with similar provisions.

The practical effect is that the president can now remove any FTC commissioner at will, and the commission is, for practical purposes, an executive agency. By late August, PPC Land reported, it had only two sitting members, the chairman Andrew Ferguson and Mark Meador, with three seats vacant. The decisions described here were all taken by votes of 2-0.

That commission has nonetheless been active in advertising. Its strategic plan for 2026 to 2030, published on April 3 and approved 2-0, added to the agency's mission the words "without unduly burdening legitimate business activity" and called children's online safety "one of the most important consumer protection issues of our time", as PPC Land reported in its account of how the FTC's 2026-2030 plan puts Big Tech, kids' data and ad fraud in its sights on April 6. In August, it issued final orders against Cox Media Group over its "Active Listening" product, marketed as an AI system that detected purchase intent from conversations captured by smartphones, smart televisions and smart speakers. No such capability existed, the FTC said; customers received email lists bought from data brokers. PPC Land reported on August 30 that Cox Media Group faces 20 years of FTC oversight over fake AI listening ads, with an $880,000 payment from CMG and $25,000 each from two partner firms. On October 29 the commission and eight states sued Toast over hidden fees and "$0" hardware claims, PPC Land reported on November 1. The agency has also sued three of the largest agency holding companies over alleged collusion on brand safety, a case this newsletter does not cover in detail.

What can a Democratic Congress do with an FTC that answers to the president? It cannot, in practice, restore the commission's independence: any statute to that effect would face a veto. But it has three levers.

The first is confirmations. With three vacancies, the commission's composition over the next two years depends on whom Trump nominates and whom a Democratic Senate is willing to confirm. A Senate majority can refuse nominees, or extract commitments in hearings. Whether it can insist on Democratic commissioners, as the old statutory balance assumed, is now a political question rather than a legal one.

The second is money. The FTC's budget is set in appropriations bills that need both chambers. A Democratic House can fund specific priorities, such as children's privacy enforcement, and attach conditions, though any spending bill also needs the president's signature.

The third is oversight. Ferguson and Meador can be called to testify, and documents on contentious matters, from the AI accuracy policy statement aimed at Colorado to the brand safety litigation, can be subpoenaed. That does not change what the commission decides. It changes the cost of deciding it.

For advertisers, the FTC's direction under Ferguson is unlikely to change much because of Tuesday. Its focus on deceptive claims, hidden fees, lead generation and children's data will continue. What may change is the political cover for its more contested moves, such as using the FTC Act to pre-empt state AI rules, which a Democratic Senate Commerce Committee will scrutinise closely.

The FCC, broadcasters and the airwaves

The FCC has followed a similar path, with even more direct consequences for the companies that sell television advertising.

Under its chairman, Brendan Carr, the commission voted on August 6 to repeal the rule that limited any company's local television stations to reaching 39% of US households, Adweek's Mark Mwachiro reported in his account of how the FCC voted to eliminate the cap on local station ownership. The vote was 2-1, with Carr and Olivia Trusty, both Republicans, in favour and Anna Gomez, the commission's only Democrat, against. Congress had written the 39% figure into law in 2004, and the article noted that the repeal "is certain to face legal challenges." Gomez's objection was economic as much as legal. "Eliminating the cap does not free local broadcasters from economic pressure; it just changes who is doing the squeezing," she said. "The large station groups positioned to grow even larger under this decision are not local broadcasters; they are national companies that own local stations and increasingly dictate what airs on them." The decision cleared a path for Nexstar's $6.2 billion purchase of Tegna, which attorneys general from eight states had asked a court to put on hold pending an antitrust case.

The FCC has also used its licensing power in ways Democrats describe as retaliation. In late April it ordered eight ABC-owned stations to file early licence renewals, an order Adweek linked to scrutiny over a Jimmy Kimmel joke about Melania Trump. On August 18, Adweek reported that ABC had sued the FCC in federal court in Washington, arguing that "government censorship is deeply un-American" and that the commission's demand "reveals its intent to either hold a hearing to deny the applications or simply to subject Plaintiffs to months - and likely years - of onerous litigation and regulatory uncertainty."

Democratic senators have objected throughout. In February 2025, Markey, Ben Ray Luján and Gary Peters wrote to Carr that "this weaponization of the FCC is unacceptable," Adweek reported at the time, after the commission reinstated complaints against ABC, CBS and NBC. In March 2025 Blumenthal, as ranking member of the Permanent Subcommittee on Investigations, described the FCC's investigations of broadcasters as "arbitrary and capricious" and "retaliatory actions against news organizations targeted by President Donald Trump."

Those letters went largely unanswered under a Republican Congress. From January, the same senators will chair committees and subcommittees that can compel answers. Democrats cannot reverse the ownership cap repeal by statute without the president's signature, and the Slaughter ruling makes the FCC's commissioners, like the FTC's, removable at will. But the media companies that depend on broadcast licences, and the advertisers that buy their airtime, should expect the FCC's use of its licensing power to become the subject of hearings, document requests and appropriations riders. The FCC is also one of the agencies the December executive order asked to consider a federal AI disclosure standard that would pre-empt state laws; any such proceeding will now be watched by a hostile Congress.

Antitrust: remedies without a breakup

On antitrust, the year's decisions came from the courts, and the election does not change them. It may change what Congress does next.

On September 2, Judge Leonie Brinkema of the Eastern District of Virginia rejected the Justice Department's request to force Google to sell its AdX exchange, accepting most of the behavioural remedies instead. When her 106-page opinion was unsealed, PPC Land reported on September 17 that Google faces a six-year worldwide ad tech decree instead of an AdX sale. The court found "structural remedies are neither realistic nor needed", observed that the case "boils down to a lack of trust", and ordered Google to open AdX and its DFP ad server to Prebid and to rival ad servers, under a monitor and a three-member technical committee, for six years rather than the 15 the plaintiffs sought. "For these reasons, the Final Judgment will apply beyond the United States and reach globally," Brinkema wrote. Google's appeal in the separate search case has been at the DC Circuit since May.

In November 2025, the FTC lost its case against Meta when Judge James Boasberg ruled that Meta does not hold monopoly power, finding that TikTok and YouTube compete with Facebook and Instagram. "While it once might have made sense to partition apps into separate markets of social networking and social media, that wall has since broken down," he wrote.

Those outcomes have revived interest in legislation. Patrick Gallaher of Public Knowledge argued in Tech Policy Press on September 28 that the Google ad tech remedies show the need for legislation, pointing to the bipartisan Advertising Middlemen Endangering Rigorous Internet Competition Accountability Act, the AMERICA Act, which would bar companies with more than $20 billion in digital ad revenue from owning more than one part of the ad tech chain, and require those with more than $5 billion to act in their customers' best interests. "Fortunately, there is bipartisan legislation that gets at the heart of the problem of monopolistic digital ad tech," he wrote. Public Knowledge is an advocacy group that has long favoured structural separation, and the bill has not advanced. But it is the kind of measure, with sponsors from both parties, that a divided Congress could take up.

The more immediate effect of the election is on oversight of antitrust enforcement itself. From January, Democrats will chair the Judiciary committees that oversee the Justice Department's Antitrust Division, with the power to hold hearings and issue subpoenas. Mergers that depend on regulatory goodwill, from Nexstar and Tegna to the media and ad tech deals still under review, will be scrutinised by a Congress that can make the process uncomfortable even if it cannot block a transaction.

The record cycle that paid for it

All of this was decided in the most expensive midterm election on record, and the advertising market has been both the beneficiary and the collateral damage.

Political advertisers had spent $11.4 billion by November 2, up 28% from the 2022 midterms and above the $11.2 billion record of the 2024 presidential cycle, Adweek's Kendra Barnett reported, citing AdImpact data, when she wrote that CTV had overtaken cable in political ad spend for the first time in a midterm. Spending on connected television rose 122% from about $1.1 billion in 2022 to $2.4 billion, while cable fell 26% to less than $1.3 billion. Of the $2.5 billion increase on 2022, 53% went to streaming.

The money arrived where the races were closest, and pushed other advertisers aside. Digiday reported on October 26 that midterm political spending was throwing a spanner into regional advertisers' holiday plans in states such as Texas, Wisconsin, Maine, Michigan, Pennsylvania, Iowa and Ohio, where stations exercised pre-emption clauses to bump brands for political buyers. "Many clients are holding their breath until we know the outcome of the mid-terms," Wendy Mason of Moroch told Digiday's Media Buying Briefing the same day. Several of those states produced the night's biggest upsets.

The broadcasters gained. PPC Land reported on November 3 that Fox's advertising revenue rose 47% to $2.08 billion in the quarter to September 30, with Lachlan Murdoch saying political revenue had passed $350 million across the cycle. Comcast's advertising arm described the same surge to Adweek on October 28. The platforms set their own limits: Meta blocked new US political ads for the eight days before the vote, from October 27 to November 3, PPC Land reported on October 11. Creators became a channel in their own right: Digiday reported on October 29 that the midterms could mark a turning point for influencers in political campaigns, with the Democratic National Committee's creator programme counting more than 550 members.

AI-generated content was a smaller factor in the ads themselves than many had feared. MediaPost asked on November 2 whether this was the first AI slop election, and answered "not so fast". Its October 2 report that most US voters want strong AI regulation points to a more durable effect: the voters who sat through $11.4 billion of political advertising arrived at the polls already sceptical of the technology that the next Congress will be asked to regulate.

What happens to the advertising market now that the money has stopped? The immediate effect is that local television inventory frees up for the holiday season, and the pricing pressure that political buyers placed on streaming services eases. The longer effect is on who regulates the channels that took the money. The broadcast stations that carried most of it are licensed by an FCC whose chairman now faces a hostile Congress; the streaming platforms that took a record share are owned by companies whose data practices are governed by state privacy laws that will not now be pre-empted; and the social platforms that limited political ads before the vote are the same companies whose AI ambitions depend on data centres that voters in Ohio were asked to approve or reject.

The lame duck, a last window

Between now and January 3, the outgoing Congress still has Republican majorities in both chambers. That gives Republican leaders a short and unusual opportunity, and it gives Democrats a reason to be careful about what they agree to before they take control.

The constraint is the same as it has been all year. In the current Senate, Democrats and the two independents who caucus with them hold 47 seats, enough to sustain a filibuster. Anything that needs 60 votes needs Democrats. The routes around the filibuster are narrow: budget reconciliation, which Republicans tried in 2025 for the AI moratorium and which the Senate rejected 99 to 1, and must-pass bills such as defence authorisation and government funding, into which Republicans tried to insert AI preemption in November 2025.

Three items are worth watching. The first is children's online safety. The House has passed the KIDS Act with the Guthrie-Pallone compromise, and the Senate Commerce Committee has sent KOSA to the floor. A lame-duck deal would let Republicans claim the achievement, and would let Democrats lock in a federal floor for children's privacy that states can exceed. It would also force a decision on the duty of care that Blumenthal and Blackburn insist on and the House dropped. Advocates who prefer the Senate version may judge that they will get a better deal from a Democratic Congress in 2027, and run out the clock.

The second is the SECURE Data Act. It has had a legislative hearing, in June, and it has no Senate companion. House passage in the lame duck would be symbolic, but it would put a Republican privacy text on the record as the starting point for any future negotiation. Its sponsors have little else to gain by waiting.

The third is AI preemption. The administration has been clear that it wants a federal statute, and the White House framework was written with this Congress in mind; Lawfare noted in April that it offered "a fairly clear sketch of which types of AI policy could become U.S. law before the 2026 midterm elections." That window has now closed in its original form. A last attempt to attach preemption language to a must-pass bill cannot be ruled out, but Senate Democrats defeated the same manoeuvre a year ago, from the minority, and they have just been told by voters that they will be the majority.

For advertisers and AI companies, the practical point is that the lame duck is the last moment for two years when the arithmetic favours federal preemption. If nothing passes by January 3, the strategy that much of the industry pursued in 2025 and 2026 will have to be replaced by one built for a divided government.

What changes on January 3, and what does not

Set out plainly, the consequences of Tuesday's result for advertising and AI fall into three groups.

Some things are unlikely to change at all. The president keeps his veto, his executive orders and his control of the FTC and FCC, which the Supreme Court confirmed in June. The AI Litigation Task Force will keep suing states. The FTC's proposed policy statement on AI accuracy, the FCC's ownership cap repeal and the Google ad tech decree all stand. Agency enforcement against deceptive advertising, hidden fees and children's data violations will continue on its current course.

Some things become much less likely. A federal privacy law that pre-empts state laws, of the kind the SECURE Data Act proposed, now has no realistic path before 2029, unless Republicans pass something during the lame duck that can also clear the Senate. A federal statute pre-empting state AI regulation, as the White House framework sought, faces the same obstacle. The industry's hope of a single national rulebook for data and AI has receded, and the state-by-state regime, from California's DROP to New York's chatbot law to Colorado's rewritten AI Act, will govern most of the market for at least two more years.

Some things become more likely. Oversight hearings on AI companies, on data centres and electricity prices, on the FTC and FCC, and on the treatment of broadcasters will begin early in 2027, possibly through a House select committee on AI with subpoena power. Children's online safety legislation remains the most plausible bipartisan law, either in the lame duck on House terms or in the new Congress on terms closer to the Senate's. Narrow, bipartisan measures such as the Stealth Bot Prohibition Act stand a better chance than broad frameworks. And the Senate will decide who fills the three empty seats at the FTC.

There is a fourth, less tangible consequence. The AI industry spent heavily in 2026 to shape the politics of its regulation, and then retreated when it discovered that its support was unwelcome. The members who take office in January have seen that. So have the data centre developers whose projects were on local ballots, and the platforms whose political advertising policies were tested in the busiest cycle on record. A Congress that cannot pass much will still be able to make a great deal of noise, and in technology policy, noise from a majority that controls subpoenas has a way of becoming a constraint.

The calendar is now fixed. The lame-duck Congress returns before the end of the year with the KIDS Act, the SECURE Data Act and defence spending all available as vehicles. New York's chatbot law and Colorado's rewritten AI Act are due to take effect on January 1, 2027. The new Congress is sworn in on January 3. For an industry that has spent two years planning for one federal rulebook, the most important date on that list may be the one on which nothing happens to the 20-odd rulebooks it already has.

By the numbers

$880,000
Paid by Cox Media Group to settle FTC claims over a listening product that could not listen.
Cox Media Group faces 20 years of FTC oversight over fake AI listening ads

6 years
Length of the behavioural decree Google received in place of a forced sale of AdX.
Google faces six-year worldwide ad tech decree instead of AdX sale

242,000
Californians who used the state's one-step broker deletion service in its first two months.
California's DROP platform hits 242,000 sign-ups in its first 8 weeks

$350 million
Political advertising revenue Fox had booked across the cycle, by Lachlan Murdoch's count, and still rising.
FOX ad revenue gains 47% to $2.08bn on World Cup and midterm ads

January 1, 2027
Start date for New York's limits on AI companion features offered to users not verified as adults.
Hochul signs law barring human-posing AI chatbots for under-18s