CNN Data Expert Says Trump Would Lose to Kamala Harris by Eight Points if 2024 US Elections Repeated This Year

Donald Trump won the presidency five months ago by about one point. An NBC News/SurveyMonkey poll now puts him eight points behind Kamala Harris in a hypothetical rematch. Harry Enten, CNN's chief data correspondent, walked through the figures on CNN News Central this Wednesday and none of them were kind.

Nine points. Gone. In five months.

The poll is a rerun of an earlier survey weighted to match 2024 turnout. First time around, Harris trailed by one, a near-perfect echo of November. Latest version: Harris plus eight. And the movement is not coming from people who skipped the election discovering her for the first time. The movement is coming from Trump's own voters flipping.

Enten quoted Sinatra. 'Regrets of some folks, had a few.' The panel laughed. But the underlying finding is serious. Within the weighted sample of confirmed 2024 Trump voters, Harris now leads by what Enten described as a meaningful margin. People who showed up in November and ticked the box for Trump are now telling pollsters they got it wrong. Voters do not do this lightly and they do not do it often, especially not before a president has finished his first hundred days in office.

Something broke.

Where, exactly, is harder to pin down. The poll measures movement, not motivation. But the context is not subtle. Grocery prices have not come down. Rent has not come down. Petrol has not come down. The federal immigration crackdown has produced enforcement operations ending in civilian deaths, Savannah last week being the most covered, and stories like those lose moderates in a hurry even when the base approves. Then the Epstein files landed and kept landing, week after week, tying Trump to headlines no press team on earth wants to manage. Younger voters noticed. Moderates noticed.

Net approval: minus 22 per cent. His own first-term midpoint was minus 12. Biden at the same stage sat at minus 13. Those numbers, in any normal presidency, would have strategists on both sides drawing midterm maps with the opposition gaining seats. Trump has defied those patterns before. He has also lost a midterm badly before, in 2018, when his approval was better than it is now.

Where the Base Is Cracking

Base loyalty is splitting unevenly. Around 89 per cent of 2024 Harris voters say they would back Democrats in the midterms. Only 83 per cent of Trump voters say the same about Republicans. Six points. Does not sound dramatic. In a House race decided by 8,000 votes in a suburban district outside Philadelphia or Phoenix, six points is the whole election.

Non-voters lean Democratic by 16 points. Same group leaned Trump by seven in 2020. A 23-point swing. These people are unreliable and frequently do not show up. But they showed up in 2018, when they were angry, and even a partial showing in 2026 rearranges the board.

The 'Middle Is Walking Away'

Independents are the worst news for the White House. They delivered the margin in November. They made the one-point win possible. Enten's data shows them drifting leftward, worn out (his phrasing) by the noise. Losing your base is survivable if the middle holds. Losing the middle while your base softens is how you end up defending seats you expected to win comfortably. Prediction markets now give Democrats a 40 per cent chance at the Senate, up from 19 per cent a year ago. Still under coin-flip odds. The trendline is steep enough to notice, though, and Republican operatives have noticed. Nobody confirmed this on the record. Mind you, nobody ever does.

Enten presented the data on air without commentary, without predictions, without telling the audience what to conclude. Smart approach. The numbers did not need a narrator. Democratic voters: unified. Republican voters: wobbling at the edges. Non-voters: leaning left. Independents: restless and drifting. Put all of it on a midterm map and the picture shifts toward Democrats in a way consultants on both sides are now describing, privately, as significant. One Republican strategist, speaking to a different outlet last week and not for attribution, used a word stronger than significant. The word was 'bad.'

A Snapshot, Not a Verdict

This is a snapshot. Enten said so. Polls taken this far from an election measure where voters are, not where they will be. Moods shift. Events intervene. Trump has come back from worse positions in public opinion, though not from this particular combination of base erosion, negative approval and independent drift all at once, not this early, and not while already holding the office. The snapshot shows a sitting president eight points behind the woman he beat half a year ago with his coalition visibly thinning at every edge. Enten laid out the numbers on a Wednesday afternoon and let them sit on the screen. The audience did the maths themselves.