Who will win the Democratic nomination for President in 2028?

TL;DR

Other is the largest category at 25%, because it combines several credible contenders. Pete Buttigieg leads the named candidates at 18%, followed by Gavin Newsom and Kamala Harris at 17% each (forecast calculation). The forecast turns on who runs, not just who leads the polls.

Context

I forecast the person formally nominated by the Democratic Party, using information available by October 5, 2026, at 11:34 UTC. Withdrawal before formal nomination can change the answer; replacement afterward cannot. Earlier polls and interviews are background evidence, not events that resolve the question (forecast scope).

The DNC approved its sanctioned early-state calendar on August 15, 2026: South Carolina on January 22, 2028; Nevada on February 1; New Hampshire on February 8; New Mexico on February 15; Michigan on February 22; and Virginia on February 29. This gives candidates a sequence of different electorates to navigate, rather than a single opening-state test (DNC announcement).

Evidence

The historical backbone favors a dispersed forecast. A Gallup retrospective published July 30, 2002, found that early Democratic front-runners failed to win in seven of ten examined nomination contests. Its historical discussion spans elections from 1952 through 2000. This is a small, old, institutionally mixed reference class, not a fitted probability for the current polling leader (Gallup retrospective).

Entry uncertainty has a close historical analogue. Al Gore led hypothetical nomination polling before declining a presidential bid on December 15, 2002. His decision then reordered the field. A former nominee can have substantial support without deciding to seek the nomination again (Gallup contemporary account; CNN announcement coverage).

The verified national polling establishes Harris as the current plurality leader, but also shows how much survey design changes the answer. These are selected cross-sectional observations, not a comparable time series. All support figures below are voter preferences, not nomination probabilities.

Harvard uses an opt-in online sample, a short candidate menu, and no published undecided category. Echelon offers a broader randomized menu and includes Democratic leaners. Emerson screens for intended Democratic-primary participation rather than simply Democratic identification. I therefore reject both a mechanical average and the claim that the contrary Emerson result identifies the uniquely correct electorate (Harvard methodology; Echelon questionnaire; Emerson methodology).

Harris’s lead is not adequately explained by name recognition alone. Harvard reports her at 56% among Black Democrats and 47% among Hispanic Democrats, although subgroup sample sizes are missing. I read that as a real constituency and a potential advantage in the sanctioned opening contests, not proof of a South Carolina polling lead (Harvard demographic results).

There is weaker evidence against assuming that Harris will easily consolidate the field. CSP/Liberal Currents surveyed 683 Democratic voters on September 3 and September 8–10, 2026, using SMS-to-web and online panels. Harris received 25% first choice and 42% among the weighted Black subgroup of 116. On the fallback question, she received 8%, versus Newsom’s 15%, Kelly’s 13%, and Buttigieg’s 12%. First-choice supporters cannot choose themselves as a fallback, so this is not a complete transfer matrix or a direct measure of anyone’s ceiling. The original September-release PDF was available by the cutoff but does not display an exact publication date (original crosstabs).

Early-state evidence gives Buttigieg and Ocasio-Cortez a concrete opportunity. Saint Anselm’s August 17–18, 2026, online survey reported Buttigieg at 25%, Ocasio-Cortez at 17%, Newsom at 7%, and Harris at 3% in its weighted Democratic-primary tab of 733. The PDF lacks an exact release timestamp but was available before the cutoff. This measures New Hampshire, not the national coalition needed to win (original survey).

The entry evidence changes the ranking more than small polling differences:

• Harris and Newsom: Harris remains undecided; five of eight former advisers expected non-entry in a September 20 report (Axios). Newsom’s September interview contains an explicit conditional pledge: “I wouldn’t run if she ran.” His explanation concerned overlapping networks and the cost of competing against her. I take that seriously without treating it as irrevocable (CNN transcript, September 15).
• Buttigieg: His August interview described him as “more inclined than not” to run. October 4 reporting documents completed New Hampshire campaigning and a further multistate itinerary. These are affirmative preparation signals, not a declaration (August interview; October 4 reporting).
• Ossoff: His June and August denials expressly reject running in the coming presidential cycle; they are stronger than a generic refusal to discuss future plans. Reelection gives him a reason to suppress speculation, and a victory could change his incentives. It does not justify ignoring the denials today (June 28 CNN transcript; August 17 interview, published August 18).
• Ocasio-Cortez and Whitmer: Ocasio-Cortez is weighing the presidency against a Senate campaign, while upstate New York activity preserves both paths (September 27 reporting). Whitmer softened her rejection of a run with “Never say never” later on May 28. That preserves an option, not a commitment (Bridge Michigan).

Shapiro has a credible route through a strong reelection showing and an electability argument. Siena’s September 24 release reported him leading Stacy Garrity by 20 percentage points. I treat that as positioning evidence, not an already completed victory or proof that Democratic primary voters will choose him (Siena release).

Fundraising establishes capacity, not presidential intent. Ossoff’s Senate committee reported $77.3 million in total receipts from January 1, 2025, through June 30, 2026, in filings available by the cutoff. That includes transfers and is not all fresh individual donations. It finances a Senate campaign; it does not establish a presidential campaign (FEC committee records).

Other is not just an allowance for an unknown outsider. AP’s September 28 report describes planned South Carolina activity by Kelly, Beshear, Pritzker, Khanna, and Emanuel. Their preparations create several separate routes to an unlisted nominee, although planned appearances are not completed events or declarations (AP reporting).

I convert this evidence into a candidate-entry and relative-strength model. The following inputs are my judgments, not reported statistics. Entry means meaningful pursuit or availability to accept nomination before the contest is settled. Strength is measured in arbitrary relative units and incorporates polling, coalition breadth, organization, experience, and room to grow (complete inputs and calculation).

I model the omitted contenders separately rather than treating Other as one candidate. For each possible entering field, an entrant’s nomination probability is proportional to their strength raised to a concentration exponent. I average exponents of 1.3, 1.6, and 1.9 with weights of 25%, 50%, and 25%, respectively. This retains both breakout potential and an advantage for established contenders. The calculation enumerates all entry combinations, explicitly links Harris and Newsom, and retains a residual for unlisted possibilities. It is not a fitted model or a state-by-state delegate simulation, and no prediction-market prices enter it (reproducible calculation).

What's non-obvious

The sanctioned calendar is not a guarantee of the order in which people vote. New Hampshire law requires its primary to precede a similar election elsewhere by at least seven days, and October 4 reporting confirms continued local resistance to the DNC order. But the DNC strengthened sanctions to include loss of delegates and candidate exclusion from debates and early-window delegates. I retain South Carolina’s strategic importance without assuming that New Hampshire disappears from the campaign’s momentum (state statute; October 4 reporting; DNC sanctions).

The other distinction is between strength and availability. The model gives Harris a 36.9% chance conditional on entry and Newsom 33.7%, versus Buttigieg’s 22%. Buttigieg narrowly leads unconditionally because he is more likely to pursue the race. Conversely, Ossoff’s low final probability reflects entry uncertainty rather than a claim that he lacks voter interest. Other wins about one-quarter collectively; no single omitted person leads the named contenders in this calculation (conditional probabilities).

Uncertainties

I lack a comparably current, fully documented South Carolina nomination poll and a complete candidate-to-candidate transfer matrix. Harvard’s missing subgroup sample sizes limit precision. Public evidence also cannot establish private entry decisions. Those gaps matter more than small differences between the leading candidates (evidence limitations).

Sensitivity checks show the scale of the uncertainty. Changing Harris’s entry probability from 30% to 60.0%, holding other assumptions fixed, moves her nomination probability from 11% to 22% and Newsom’s from 21% to 13%. Changing omitted contenders’ strength ratings by one-quarter moves Other from 18% to 31.6%. These are assumption tests, not confidence intervals. Donor coordination, additional entry correlations, campaign performance, health, scandals, and changing issues remain outside this simplified calculation. The displayed decimals preserve arithmetic; the ranking among the leading named candidates is not robust to modest changes in judgment (sensitivity results).