# AIPAC, AI, and Crypto-Funded Democrats Underperform Democrats Who Reject Corporate PAC Money - Data For Progress

*Источник: Data For Progress*
*Дата: 2026-07-28*
*Язык: en*

**Кратко:** AIPAC, AI, and Crypto-Funded Democrats Underperform Democrats Who Reject Corporate PAC Money
With the midterm elections less than four months away, some analysts are already predicting it will be the most expensive midterm cycle in U.S. history, with spending expected to reach $10.8 billion over the cycle to support or oppose candidates across the ballot.

AIPAC, AI, and Crypto-Funded Democrats Underperform Democrats Who Reject Corporate PAC Money
With the midterm elections less than four months away, some analysts are already predicting it will be the most expensive midterm cycle in U.S. history, with spending expected to reach $10.8 billion over the cycle to support or oppose candidates across the ballot.
This year, outside spending money has broken through in Democratic primaries like never before. In New York's 12th Congressional District, artificial intelligence (AI) super PACs poured more than $27 million into the June 23 primary. And in Michigan’s Senate primary, the American Israel Public Affairs Committee (AIPAC) and its affiliates have spent close to $30 million backing Representative Haley Stevens against Abdul El-Sayed, their largest ever investment in a single race.
Already, cryptocurrency, AI, Big Tech, and online betting corporations have collectively spent $294 million in political races this cycle, more than half of the $517 million spent by corporations overall.
While financial support from an outside group can help fill the coffers for a general election campaign, taking money from an unpopular source, such as AIPAC or AI companies, can also be costly to a candidate’s reputation.
To take a deeper look at how sources of outside funding affect candidate viability, Data for Progress conducted a conjoint survey experiment analyzing how funding sources impact support for candidates in a general election matchup.
We find that, compared with a Democratic candidate who rejects outside PAC money, Democratic candidates who take money from AIPAC, AI companies, cryptocurrency companies, and pharmaceutical companies perform significantly worse against a Republican candidate averaged across all funding sources.
In our experiment, each survey respondent was shown a “Candidate A” and a “Candidate B.” Each matchup included a Republican and a Democrat, with the candidates’ source of outside funding randomized across 10 different groups. In addition, we tested a candidate who “rejects all financial support from outside industry groups and PACs.”
Each respondent evaluated seven hypothetical matchups. Using that data, we first calculated the average marginal component effect (AMCE) of each funding source, using a Democratic candidate who rejects all outside PAC money as the baseline — allowing us to see how each individual funding source affects Democratic candidate performance against a Republican across a randomized set of the same funders, relative to a candidate with no outside financial backing at all (see our survey methodology for full details).
When compared with this baseline, a Democratic candidate funded by AIPAC statistically underperforms, as do Democratic candidates funded by AI companies, cryptocurrency companies, and the pharmaceutical industry.
By contrast, when compared with the Democrat who rejects outside PAC funding, Democrats funded by the National Rifle Association, oil and gas industry, real estate development industry, police unions, nurses unions, and Planned Parenthood do not see a statistically significant difference in their performance against a Republican.
But how do these funding sources stack up against one another, not just against a candidate who rejects PAC money entirely? To answer that, we calculated the AMCE of each funding source against every other funding source as a baseline — allowing us to determine which funding sources in our list are most beneficial to a candidate's support, and which are most detrimental, relative to one another.
We find that some funding sources clearly underperform others, enough to reach statistical significance in our test. For example, a Democratic candidate financially supported by AIPAC performs significantly worse than Democratic candidates supported by Planned Parenthood, nurses unions, or those who reject PAC money entirely. The full matchups are displayed in the chart below, with matchups that reach statistical significance colored in green.
Our experiment demonstrates that voters have meaningful preferences about where a candidate gets their money, and these preferences are strong enough for some voters to switch their vote from one party to another.
However, these results may not translate directly to a real-world ballot booth, where voters are considering many different factors when selecting a candidate. Our experiment does not mean that, for example, a candidate funded by AIPAC will always perform exactly 6.2 points worse against a Republican than a candidate who rejects corporate PAC money. But these results do indicate that, all else being equal, the former Democratic candidate could risk an underperformance against a Republican in a general election.
Survey Methodology
From July 17 to 19, 2026, Data for Progress conducted a survey of 1,207 U.S. likely voters nationally using web panel respondents. The survey was conducted in English. For more information, please visit dataforprogress.org/our-methodology.
In a conjoint experiment, each respondent saw seven matchups, each displaying one Democratic candidate and one Republican candidate. This design choice was made to simulate a general election environment where voters are asked to choose between one of two major parties.
Each candidate is assigned one of 11 funding sources at random (including an attribute for a candidate who “rejects all financial support from outside industry groups and PACs”). The funding sources are fully randomized, so candidates may face off against another candidate with the same funding source. Respondents must choose between one of the options, and may not opt out.
We analyze the results using a weighted least squares regression that incorporates our default survey weights for individual respondents. Our dependent variable is a binary that represents whether the candidate chosen is a Democrat. We control for the candidate’s funding source (the conjoint attribute), and the respondents' education, gender, race, partisanship, income, and age. We use robust standard errors clustered at the respondent level to estimate the 95% confidence intervals around AMCE estimates.

[Оригинал](https://www.dataforprogress.org/blog/2026/7/28/aipac-ai-and-crypto-funded-democrats-underperform-democrats-who-reject-corporate-pac-money)