VOTEWIRE AI ANALYSIS
Seven Republican-held seats flip, all thirteen Democratic seats hold, and Georgia leads the way — VoteWire's AI analysis calls from final polls, live markets, social tone, a record early vote, and tonight's exit-poll mood. Projections, not official results.
Update, 7:45 PM ET: VoteWire's AI analysis now calls these races — Georgia first, for Ossoff. These are VoteWire's projections, not official results.
First, what this is and what it is not. This is VoteWire's own AI analysis call, made from pre-results data as of 7:30 PM ET on Election Night: final VoteHub poll averages, live Polymarket prices from the 6:53 PM ET refresh, Bluesky social tone, early-vote totals and party-registration mix from the University of Florida Election Lab, and tonight's national exit-poll mood. These are VoteWire's projections, not official results. They are also distinct from the Signal Predictor on the Signal page, which uses published weights only and no AI judgment — that ensemble independently points the same way, calling 20 of the 35 seats Democratic, which is why the two agree on the headline.
The call: 54 Democrats, 46 Republicans. Seven flips, zero Democratic losses.
The national mood in tonight's exit polls is the backdrop, not the whole story. CNN's Voter Poll finds roughly 6 in 10 voters disapproving of President Trump's job performance, about half of them strongly, and roughly three-quarters dissatisfied with the direction of the country and calling economic conditions bad. About 9 in 10 say the cost of living was crucial or a major factor in their vote, and the economy is the top issue named, followed by the state of democracy. NBC and CBS report Trump approval at 38% among voters, the lowest for a president in a midterm in their series, with NBC finding 43% saying part of their vote was meant to express opposition to Trump. Exit polls are early estimates with margins of error, not results — but the direction is consistent across networks, and it matches the final pre-election polls, which had Democrats ahead on the generic ballot by 7 to 9 points.
The early vote is the second leg. A record 57.5 million ballots were already cast before Election Day, according to the University of Florida Election Lab's November 2 evening count, against 45.9 million early votes in the entire 2022 midterm. In states that report party registration, registered Democrats made up 43.4% of early voters against 31.6% registered Republicans. Registration is not a vote — early voters can split, and Election Day voters are a different electorate — but the turnout surge is real in the battlegrounds: more than 7 million early votes in Texas, early voting up about 37% on 2022 in North Carolina, South Carolina passing its entire 2022 early total days early, Kansas advance ballots running about two-thirds ahead of 2022, and Georgia's early in-person voting breaking the state's midterm record.
The markets are the third leg, and they moved toward Democrats all day. The Polymarket Senate-control market reached 87.5% Democratic by the evening refresh, up from 77.5% at the morning briefing, with the House market at 98.2%. Per-race prices are emphatic where the polls are close: North Carolina 98.3% Democratic, Georgia 97.8%, Michigan 85.5%, Texas and Ohio both 81.5%, Maine 73.5%, Kansas 72.5%. Markets are prices, not polls — they move on money and momentum as well as information, and they have been wrong before — which is why no call here rests on a market alone.
The seven flips, weakest first. Alaska is the softest call in the table: a 50.9% Democratic poll average across just three polls, no market price, and ranked-choice voting that makes the final round hard to model — but social tone runs against the incumbent, and the polls point the same way. Kansas is nearly as soft: a 52.3% Democratic poll average and a 72.5% market price, against a 94-year drought for Democratic Senate wins in the state and a social signal that leans the other way. Iowa is the call that changed since our November 2 analysis, which held the open seat Republican: the market has since moved to 65.5% Democratic, and no recent poll average in our data contradicts it — but it is the least poll-anchored flip, and we flag it as such. Maine is firmer — a 52.3% Democratic poll average and a 73.5% market price — with the standing caveat that Senator Collins overperformed her polls in 2020. Ohio's special election has the deepest polling of any flip: a 52.1% Democratic average across twelve polls, an 81.5% market price, and even the closest final poll with Brown ahead or tied. Texas rests on a 50–43 New York Times/Siena result and an 81.5% market price. North Carolina is the strongest flip in the country: an open seat, a 98.3% market price, and final polling with Cooper ahead by wide margins.
On the other side, all thirteen Democratic-held seats are predicted to hold, including Michigan, where El-Sayed's market price is 85.5%, and Georgia, where Ossoff's case is the strongest of any Democratic hold: a 54.7% poll average, a 97.8% market price, social tone running his way, and a state early-vote record. Republicans are predicted to hold fifteen seats, including Florida, where the ensemble leans Republican despite early counted votes leaning the other way — the Associated Press rates Florida too early to call, and counted votes so far are not the electorate. South Carolina is the closest Republican hold, a genuine toss-up leaning Republican on polls and markets even as social tone leans Democratic. Nebraska leans Republican. Virginia was already called for Warner by the Associated Press at research time, matching this prediction.
What would prove this wrong is also visible in advance. If Alaska and Kansas both stay Republican, the total is 52 Democratic seats, not 54. If South Carolina flips Democratic, it is 55. If Election Day turnout breaks sharply differently from the early vote — and early voters are not a random sample of the electorate — several of these calls move at once. And if the markets are mostly reacting to the same polls rather than new information, their confidence overstates what is known. The range, not the point estimate, is the honest headline: Democrats 52 to 55, with 54 as the central prediction.
| State | Call | Confidence | Poll avg (Dem %) | Market (Dem %) | Social (Dem) |
|---|---|---|---|---|---|
| Alabama | Republican | Likely | — | 2.2 | 45.0 |
| Alaska | Democrat | Toss-up | 50.9 | — | 75.0 |
| Arkansas | Republican | Likely | 43.3 | 3.5 | 55.0 |
| Colorado | Democrat | Likely | — | 99.2 | 75.0 |
| Delaware | Democrat | Likely | — | 99.0 | 75.0 |
| Florida | Republican | Likely | 47.2 | 15.5 | 43.2 |
| Georgia | Democrat | Likely | 54.7 | 97.8 | 66.7 |
| Idaho | Republican | Likely | — | 2.5 | 25.0 |
| Illinois | Democrat | Likely | — | 99.2 | — |
| Iowa | Democrat | Likely | — | 65.5 | — |
| Kansas | Democrat | Toss-up | 52.3 | 72.5 | 25.0 |
| Kentucky | Republican | Likely | — | 9.1 | — |
| Louisiana | Republican | Likely | — | 16.5 | 25.0 |
| Maine | Democrat | Leans | 52.3 | 73.5 | 51.3 |
| Massachusetts | Democrat | Likely | 65.3 | 99.1 | 67.9 |
| Michigan | Democrat | Likely | — | 85.5 | — |
| Minnesota | Democrat | Likely | — | 93.5 | — |
| Mississippi | Republican | Likely | 48.0 | 18.5 | 36.1 |
| Montana | Republican | Likely | — | 8.0 | 25.0 |
| Nebraska | Republican | Leans | 49.0 | — | 25.0 |
| New Hampshire | Democrat | Likely | — | 96.0 | — |
| New Jersey | Democrat | Likely | — | 99.2 | 75.0 |
| New Mexico | Democrat | Likely | — | 98.9 | 58.3 |
| North Carolina | Democrat | Likely | — | 98.3 | — |
| Ohio | Democrat | Leans | 52.1 | 81.5 | 65.0 |
| Oklahoma | Republican | Likely | — | 1.2 | 25.0 |
| Oregon | Democrat | Likely | — | 99.3 | 62.5 |
| Rhode Island | Democrat | Likely | 64.8 | 99.0 | 75.0 |
| South Carolina | Republican | Toss-up | 49.4 | 37.0 | 75.0 |
| South Dakota | Republican | Likely | 43.8 | 5.0 | 25.0 |
| Tennessee | Republican | Likely | — | 1.1 | 25.0 |
| Texas | Democrat | Likely | — | 81.5 | 50.0 |
| Virginia | Democrat | Likely | — | 99.1 | 54.2 |
| West Virginia | Republican | Likely | — | 1.5 | 25.0 |
| Wyoming | Republican | Likely | — | 0.9 | 53.6 |
Poll avg = two-party Democratic share from VoteHub, latest poll per pollster, 30-day window. Market = Polymarket Democratic price at the 6:53 PM ET refresh. Social = Bluesky tone on the incumbent's name, converted to a Democratic share; open seats have no social signal. — = no signal. Confidence bands: Likely means the Democratic win probability is more than 15 points from 50, Leans 5–15 points, Toss-up within 5 points. Totals: 34 Democratic seats not up this year plus 20 called here = 54; 31 Republican seats not up plus 15 called here = 46.