Most voters meet a primary candidate for the first time in an advertisement they did not ask for, paid for by a committee whose name tells them nothing. So the plain question is who paid, and what the paying is recorded as. In the 2023 to 2024 cycle, the AIPAC-affiliated super PAC United Democracy Project reported $37,860,200 in independent expenditures, spending it is barred by law from coordinating with any campaign, on OpenSecrets’ tabulation of Federal Election Commission filings released in 2025.
The second channel is smaller and works differently. AIPAC’s connected political action committee gave $3,037,900 directly to federal candidates in the same cycle, capped at $5,000 per candidate per election, 63.02% of it to Republicans and 36.62% to Democrats. The two channels leave two different kinds of filing, and the larger one records something the direct channel cannot: money spent against a candidate, $20,786,780 of it opposing Democrats.
Key Findings
- United Democracy Project, the AIPAC-affiliated super PAC, reported $37,860,200 in independent expenditures in the 2023 to 2024 cycle, uncoordinated with any campaign by the statutory definition that permits such spending (OpenSecrets, 2025, tabulating FEC filings).
- AIPAC’s connected PAC gave $3,037,900 directly to federal candidates in the same cycle, reaching 347 House candidates and 20 Senate candidates, with a largest single recipient total of $20,000 (OpenSecrets, 2025).
- The single largest direction of independent spending was against Democratic candidates: $20,786,780 opposing Democrats, against $14,067,863 supporting Democrats and $3,043,662 opposing Republicans, with no expenditure recorded supporting a Republican (OpenSecrets, 2025).
- More than $14.6 million of that independent spending went into one House race, New York’s 16th District, in 2024 (OpenSecrets, 2025).
- On the assistance itself the recorded votes run near-unanimous: the House passed the Iron Dome Supplemental Appropriations Act 420 to 9 on 23 September 2021 (House Roll Call 275, 117th Congress).
How does AIPAC-affiliated money actually reach a congressional race?
AIPAC’s connected political action committee gave $3,037,900 directly to federal candidates in the 2023 to 2024 cycle, on OpenSecrets’ tabulation of Federal Election Commission filings released 6 February 2025. That is the smaller of two channels. The second, a separately registered super PAC, files independent expenditures rather than contributions, and it carries the larger sums.
The direct channel is granular and capped. FEC committee C00797670 recorded contributions to 347 House candidates and 20 Senate candidates in the cycle, with the largest single recipient total at $20,000, to Representative Ronny Jackson of Texas. Party direction on that channel is majority Republican, 63.02% against 36.62% to Democrats.
The independent channel is FEC committee C00799031, United Democracy Project. The Guardian reported in 2024 that AIPAC established it in 2021 to engage directly in electoral campaigns. Its $37,860,200 in the 2023 to 2024 cycle is money spent about races rather than given to campaigns, and it is filed as such.
Neither figure is an AIPAC publication. OpenSecrets is an aggregator that tabulates filings the committees themselves submit to the Federal Election Commission, and totals shift as filings are amended, which is why every figure here carries its cycle and its release year. AIPAC’s own site describes its political programme in general terms and publishes no race-by-race spending ledger.
Is it often said the spending decides Democratic primaries?
It is often said that AIPAC-affiliated money decides Democratic primaries. The filed scale is real and it is concentrated: United Democracy Project reported $37,860,200 in independent expenditures across the 2023 to 2024 cycle (OpenSecrets, 2025, tabulating FEC filings). What those filings record is amount, direction and race. They do not record what any race would have done without the money.
The Intercept and The New York Times have both reported at length on the reach of that spending and on its concentration in a small number of contests, including the Illinois congressional primaries of March 2026, where the outlet reported AIPAC-affiliated committees backing candidates who won and candidates who lost in the same round of voting.
Two things are true at once and are frequently collapsed into one. The spending is large in the races it enters, and a defeat in a primary that drew heavy outside spending is not, on its own, a measurement of that spending’s effect. The filings carry no comparison group. The design that would produce an effect estimate, a set of comparable races differing only in whether the money arrived, does not exist in the public record, and no publisher has constructed one.
What the filings do establish is direction and venue. Of the itemized totals OpenSecrets tabulates for the cycle, $14,067,863 was recorded supporting Democratic candidates and $20,786,780 opposing them, while nothing was recorded supporting a Republican. The contests carrying that money are overwhelmingly contests between Democrats, which is the checkable core of the claim: the network’s independent spending is directed at Democratic races, in both directions at once.
Which committee may coordinate with a campaign, and which may not?
A multicandidate committee may give a candidate no more than $5,000 per election under 52 U.S.C. 30116, the cap governing every one of the 367 candidate recipients AIPAC’s connected PAC recorded in the 2023 to 2024 cycle. An independent-expenditure committee faces no such ceiling, and, by the statutory definition that permits its spending, may not coordinate with the campaign it helps.
That distinction is the mechanism, not a technicality. A capped contribution buys a candidate’s own campaign a fixed and small amount of capacity, and the candidate controls what it buys. An independent expenditure buys advertising the candidate does not control, cannot direct and, in the case of spending against an opponent, may not want. The two committees therefore exert different kinds of pressure on a race even when the same network stands behind both.
The filing categories follow the same split. Direct contributions appear on the recipient’s own disclosure as receipts, which is why a candidate’s donor list is a public document. Independent expenditures are filed by the spender, itemized by race and by whether the expenditure supports or opposes a named candidate, which is why the direction of the spending is knowable in a way the motive behind a contribution is not.
One consequence is worth stating plainly, because it is where the accounting of these networks usually goes wrong. The two channels are not additive descriptions of the same activity. $3,037,900 given to candidates and $37,860,200 spent about races are different legal acts with different disclosure paths, different limits and different relationships to the campaigns involved. A single cycle total that merges them describes a network, not a candidate’s funding.
Why does the money concentrate in primaries rather than generals?
More than $14.6 million of United Democracy Project’s independent spending went into a single House race, New York’s 16th District, in 2024 (OpenSecrets, 2025). Concentration on that scale is possible only in the uncapped channel. The direct-contribution channel tops out at $5,000 per candidate per election, so no amount of interest in a race can move a large sum through it.
The filings record where the money went and against or for whom. They do not record why a race was selected, and no committee publishes that reasoning. What the record supports is a description of the pattern rather than an explanation of it: the sums are not spread evenly across the 435 House contests, they land in a small set of districts, and the largest single-race figure in the cycle exceeds four times the entire direct-contribution total for all 367 candidate recipients combined.
The venue matters for a structural reason that needs no attribution. A party primary is decided by the subset of a district’s voters who turn out for it, and the advertising market in a single House district is finite. The same sum therefore represents a different share of everything a voter sees depending on which contest it enters. That is arithmetic about districts, not a finding about any committee’s intent, and the filings are silent on intent.
It is worth separating a second claim that travels with this one: that heavy spending in a primary signals to other members what criticism will cost them. The filings cannot test that. They record expenditures, not the deliberations of members who were never the target of one, and no survey of members on the question has been published.
How much of the outside spending opposes candidates rather than supports them?
The largest single direction in United Democracy Project’s 2023 to 2024 filings is spending against Democratic candidates: $20,786,780 in independent expenditures opposing Democrats, on OpenSecrets’ 2025 tabulation of FEC records. Direct contributions cannot carry a negative sign at all, since a contribution goes to a campaign, so the connected PAC’s channel is support-only by construction.
The remaining directions, as filed and tabulated, are $14,067,863 supporting Democrats, $3,043,662 opposing Republicans, and no expenditure recorded supporting a Republican. These are separately itemized categories rather than shares of a partition, and they are reported alongside the cycle total rather than as its components.
The pattern in that chart is the checkable version of a claim usually made in characterisation. The network’s independent spending in the cycle is directed almost entirely at Democratic contests, and within them the larger sum is negative. It is also the reason the two channels give opposite impressions of party direction: the contribution channel ran 63.02% Republican, while the independent channel recorded nothing at all supporting a Republican. Both are the same network in the same cycle, filed under two different rules.
What the support-or-oppose field does not carry is content. An expenditure filed as opposing a candidate may be an advertisement about foreign policy, about a domestic vote, or about neither. The FEC form records the direction and the amount, and the advertisement itself sits outside the filing.
Do the backed candidates win their races?
AIPAC’s connected PAC contributed to 347 House candidates and 20 Senate candidates in the 2023 to 2024 cycle (OpenSecrets, 2025). No published ledger matches that roster to outcomes race by race, and the recipients include non-incumbents, so the number is a count of candidates rather than of sitting members.
For the prior cycle, OpenSecrets reported in November 2022 that candidates backed by the AIPAC-affiliated super PAC won their general-election races following heavy spending in the primaries that preceded them. For the most recent round, The New York Times reported that AIPAC-affiliated committees backed both winning and losing candidates across the Illinois congressional primaries of March 2026.
A win rate assembled from those reports would not measure influence, and the reason is structural rather than a caveat about sample size. Committees choose which races to enter, and a committee that enters winnable races will post a high win rate whether or not its spending moved a single vote. The same selection runs through the contribution channel: a $5,000 contribution to a safe-seat incumbent and a $5,000 contribution to a contested challenger appear identically on the filing.
The public record therefore supports two statements and not a third. Candidates backed by these committees have often won, as OpenSecrets reported for 2022. Candidates backed by them have also lost, including in the same primary round, as The New York Times reported for 2026. What no published analysis establishes is how many of those results would have differed had the spending not occurred, and no source has claimed to have measured it.
Does the money follow existing positions or move them?
The House passed the Iron Dome Supplemental Appropriations Act 420 to 9, with 2 voting present, on 23 September 2021 (House Roll Call 275, 117th Congress).
On a question drawing 420 yes votes out of 431 members recording a position, there is very little variation left for any spending pattern to explain, because almost no one voted the other way.
The pattern holds across the small set of recent recorded votes where Israel security assistance was the measure or its principal component. The House passed H.R. 8034 366 to 58 on 20 April 2024, Republicans 193 to 21 and Democrats 173 to 37. The Senate agreed to H.R. 815 79 to 18 on 23 April 2024. The outlier is H.R. 6126, which passed 226 to 196 on 2 November 2023, and the variable distinguishing it is not the assistance but the IRS-rescission offset attached to it.
Küçük (2025), in the peer-reviewed Turkish Journal of Middle Eastern Studies, reports that members who sponsor pro-Israel bills receive more AIPAC contributions and that members facing stronger electoral competition attract more targeted donations. The study measures bill sponsorship, a voluntary individual act, rather than roll-call votes on scheduled questions, and its design cannot separate money following a position from money shaping one. The relationship between contributions and how members vote on Israel aid is correlational in every published treatment of it.
What does the public record not capture about race-level influence?
More than $4.2 million in congressional travel to Israel sponsored by the American Israel Education Foundation, AIPAC’s 501(c)(3) charitable affiliate, has been disclosed since 7 October 2023, on The Guardian’s 2026 tabulation of House ethics filings. That channel leaves a filing, but not with the Federal Election Commission, so campaign-finance totals do not contain it.
The Guardian’s tabulation counts at least 26 Democratic and 52 Republican representatives across at least 15 delegations, averaging more than $26,600 per member. The underlying records are the pre-travel approvals held by the House Committee on Ethics and the post-travel disclosures filed with the Clerk. Under the 2007 travel rules, a lobbying organization may generally sponsor only single-day travel while a 501(c)(3) may sponsor longer trips, which is why the charitable affiliate is the sponsor of record. The foundation states that its missions are educational. For 2019 alone, the Guardian analysis relays reporting by the Intercept counting 129 sponsored trips totalling $2.32 million.
Other things leave no filing at all. The FEC form for an independent expenditure records the payee, the amount, the race and whether the spending supported or opposed a named candidate. It does not record the content of the advertisement, the donors behind the expenditure at the level of a single race, or any communication between a committee and anyone. Questions about the disclosure of political spending turn on exactly that boundary.
One quantity circulates without a published basis: the share of members backed by the organization who vote for assistance without conditions. No publisher states a denominator, a vote set or a method behind that rate. The recorded votes themselves are public and are counted above, and aggregate cycle totals in the range of nine figures appear in commentary on varying accounting bases that the committees’ own filings do not reconcile.
Does election-money framing appear in measured antisemitic content?
62% of the antisemitic tweets mentioning Jews in an expert-annotated corpus of 4,016 English-language tweets fit the working definition’s paragraph on mendacious or stereotypical allegations about Jewish collective power (Jikeli and colleagues, 2022, peer-reviewed, sampled January 2019 to August 2020). That measures the prevalence of a framing in a sampled corpus, and nothing about any campaign filing.
Two further peer-reviewed measurements sit alongside it. Becker, Ascone and Troschke (2022) coded 4,500 comments on the Facebook pages of leading British, French and German media outlets and found 26.9% of the 1,504 UK comments antisemitic, with the “evil” stereotype appearing in 39.8% of those. Zannettou and colleagues (2020), analysing over 100 million posts on 4chan’s /pol/ board and Gab, judged 42% of random samples of posts containing the word “jew” to be hateful on both platforms.
Election periods coincide with movement in these measures. Verma, Sear and Johnson (2024), in npj Complexity, tracked a hate network of about 50 million accounts across platforms and recorded antisemitic content rising 117.57% over 7 to 11 November 2020 against 2 to 6 November, alongside a 269.5% rise in anti-immigration content. The authors report co-occurrence in time and make no causal claim, and neither does this report.
The distance between these two bodies of evidence is the point of measuring them separately. One is a set of dollar amounts, dates and named committees filed with a federal agency. The other is the measured prevalence of an allegation about hidden collective control in online corpora sampled years apart. Nothing in the corpora studies examines campaign finance records, and nothing in the filings speaks to the discourse.
Methodology
This report draws on four source types, each labelled where it is used. Campaign finance figures come from OpenSecrets, a campaign-finance aggregator that tabulates filings submitted by committees to the Federal Election Commission; the 2023 to 2024 cycle figures reflect FEC data released 6 February 2025, and totals are revised as filings are amended. Voting figures are primary records: recorded votes published at congress.gov, cited by roll call number, chamber and date. Peer-reviewed research is cited by author and year, with study design stated. The congressional travel figures are a named press tabulation of official ethics filings by The New York-based reporting of The Guardian in 2026, cited to the outlet for the aggregate.
Three limits shape what the figures can carry. First, the direct-contribution and independent-expenditure channels are distinct legal instruments with separate limits and disclosure paths, and their totals are not summed here. Second, the itemized support-and-oppose figures and the reported cycle total are tabulated separately and are not treated as a partition of one another; no percentage is derived from them. Third, election-stage breakdowns, primary against general, are not published in the tabulations used, so the concentration evidence in this report is race-level rather than stage-level.
Küçük (2025) is peer-reviewed and correlational, and it measures bill sponsorship rather than roll-call votes; a finding about sponsorship does not transfer to voting, which has a different base rate. Win-rate reporting from OpenSecrets (2022) and The New York Times (2026) is treated as a count of outcomes, not as an effect estimate, because committees select the races they enter.
One quantity sought for this report is not published anywhere located: the rate at which members receiving support from these committees vote for assistance without conditions. No source states a denominator, a vote set or a method for it, so no such rate appears here.
Conclusion
So who paid for the advertisement, and what does the paying record? For this network the answer is unusually specific. The record names two committees, one capped at $5,000 per candidate and support-only by construction, the other uncapped, uncoordinated by law, and reporting $37,860,200 in a single cycle. It names the direction, including $20,786,780 spent against Democratic candidates and nothing at all supporting a Republican. It names the districts, including one House race in New York carrying more than $14.6 million. And it names a separate channel with a separate filing office, more than $4.2 million in disclosed travel since October 2023.
What the same record does not contain is the part people most want from it. The measures themselves passed 420 to 9, 366 to 58 and 79 to 18, so the votes the spending is said to purchase were carried by chambers voting nearly as one, and the one House vote that split near party lines split over an IRS offset attached to the bill. Backed candidates have won and backed candidates have lost, in the same primary round. Sponsorship correlates with contributions in the peer-reviewed literature, and the direction of that relationship is unresolved by the designs available.
That leaves a gap between a very detailed record of money and a very thin record of effect, and it is a gap that discourse fills fast. The corpora put a number on what is already circulating: 62% of antisemitic tweets in one annotated sample alleging Jewish collective power, 42% of sampled posts containing the word “jew” judged hateful on two platforms. A filed expenditure is a fact about a committee and a district. Somewhere between the filing cabinet and the feed, it is read as something else. What happens to a specific, dated, public number when it arrives in a conversation that was waiting for it is a question this record raises and does not settle.
Frequently Asked Questions
How much US assistance has Israel received in total?
Israel is the largest cumulative recipient of US foreign assistance since the Second World War, at $174 billion in bilateral assistance and missile-defense funding in current dollars, per the Congressional Research Service report RL33222. CRS is the nonpartisan research service reporting to Congress, and the total moves by report edition.
Does the 2016 memorandum of understanding obligate the money?
No. The memorandum signed at Washington on 14 September 2016 sets $38 billion over FY2019 to FY2028, $33 billion in Foreign Military Financing and $5 billion in missile defense, and its own text states the levels are “subject to the appropriation and availability of funds.” Congress appropriated $8.7 billion in combined FMF and missile-defense funding in FY2024, per CRS.
Why are there so few standalone recorded votes on Israel assistance?
Assistance is ordinarily carried inside the annual State, Foreign Operations appropriations act, where a vote on the package is not a vote on any single line. That is why the usable series of recorded votes on Israel assistance as the measure itself is four votes across 2021 to 2024 at congress.gov, rather than an annual roll call.
Why can a charitable affiliate fund longer congressional trips than a lobbying organization?
Under the 2007 congressional travel rules, a lobbying organization may generally sponsor only single-day travel, while a 501(c)(3) charitable organization may sponsor longer trips. The American Israel Education Foundation is the 501(c)(3) affiliate of record on the disclosures analysed by The Guardian in 2026, and it states its missions are educational.
Has US public opinion on the Middle East changed?
Yes, on Gallup’s forced-choice sympathies question. In February 2010, 63% of US adults said their sympathies lay more with the Israelis (n=1,025, margin of error 4 points). In February 2025, 46% did, the lowest in 25 years of the trend, against 33% with the Palestinians (n=1,004, margin of error 4 points).
Where do the underlying campaign finance records live?
With the Federal Election Commission, filed by committee identification number: C00797670 for AIPAC’s connected PAC and C00799031 for United Democracy Project. Congressional travel sits elsewhere, in pre-travel approvals held by the House Committee on Ethics and post-travel disclosures filed with the Clerk of the House at disclosures-clerk.house.gov.
Sources
- Becker, M., Ascone, L., Troschke, H., 2022. Antisemitic Comments on Facebook Pages of Leading British, French, and German Media Outlets. Humanities and Social Sciences Communications. DOI 10.1057/s41599-022-01337-8. Peer-reviewed.
- Gallup, 2010 to 2025. Middle East sympathies trend series. news.gallup.com/poll/1639. Named survey organization.
- Jikeli, G., Axelrod, D., Fischer, R., Forouzesh, E., Jeong, W., Miehling, D., Soemer, K., 2022. Differences Between Antisemitic and Non-Antisemitic English Language Tweets. Computational and Mathematical Organization Theory. DOI 10.1007/s10588-022-09363-2. Peer-reviewed.
- Küçük, Ö. N., 2025. Funding in the Shadow of Conflict: AIPAC’s Financial Influence on US Congress Post-2023 Israel-Hamas War. Turkish Journal of Middle Eastern Studies, 12(1), pp. 1 to 20. DOI 10.26513/tocd.1574363. Peer-reviewed.
- OpenSecrets, 2022. American Israel Public Affairs Committee-backed candidates won midterm races following big spending by group’s super PAC. opensecrets.org. Campaign-finance aggregator.
- OpenSecrets, 2025. AIPAC PAC contributions to federal candidates, 2023-2024 cycle. FEC committee C00797670. opensecrets.org. Campaign-finance aggregator tabulating FEC filings.
- OpenSecrets, 2025. United Democracy Project independent expenditures, 2024 cycle. FEC committee C00799031. opensecrets.org. Campaign-finance aggregator tabulating FEC filings.
- The Guardian, 2024. Reporting on pro-Israel lobby spending in congressional elections. theguardian.com/us-news/article/2024/aug/16/congress-election-pro-israel-lobby-aipac. Press reporting.
- The Guardian, 2026. Aipac affiliate has funded lavish trips to Israel for dozens of Congress members since 7 October. 2 June 2026. Press analysis of congressional ethics filings.
- The Intercept, 2024. Reporting on AIPAC spending in congressional elections. theintercept.com. Press reporting.
- The New York Times, 2026. Reporting on AIPAC-affiliated spending in the 2026 congressional primaries. nytimes.com. Press reporting.
- US Congress, 2021 to 2024. Recorded votes on Israel security assistance: House Roll Calls 275 (117th Congress), 577 and 152 (118th Congress), Senate Record Vote 154 (118th Congress). congress.gov. Official record.
- US Congressional Research Service, ongoing. U.S. Foreign Aid to Israel: Overview and Developments, RL33222. congress.gov. Official record.
- US Department of State, 2016. Memorandum of Understanding Between the United States and Israel on Security Assistance, signed 14 September 2016. Official record.
- Verma, A., Sear, R., Johnson, N., 2024. How U.S. Presidential Elections Strengthen Global Hate Networks. npj Complexity. DOI 10.1038/s44260-024-00018-8. Peer-reviewed.
- Zannettou, S., Finkelstein, J., Bradlyn, B., Blackburn, J., 2020. A Quantitative Approach to Understanding Online Antisemitism. ICWSM 2020. DOI 10.1609/icwsm.v14i1.7343. Peer-reviewed.