Will Abdul El-Sayed win the 2026 US Senate election in Michigan?

TL;DR

I estimate a 61.7% chance that Abdul El-Sayed wins. He is a narrow favorite because four of the five post-primary polls put him ahead and the national generic ballot is Democratic by 6.6 points (polls; Silver Bulletin). Rogers remains highly competitive because recent Michigan Senate polls have overstated Democrats, while older and more habitual voters currently favor him (AARP).

Context

El-Sayed won the August 4 Democratic primary by about one point, 48.5% to Haley Stevens’s 47.5%, after late polls had suggested a much larger victory (primary results). He now faces Rogers on November 3 for the open seat of retiring Sen. Gary Peters.

Rogers supplies an unusually strong reference point because he was also the Republican nominee in 2024. He lost that Senate race by 19,006 votes, or 0.34 percentage points, even as Donald Trump carried Michigan by 80,103 votes, or 1.41 points (Michigan Department of State).

Evidence

The recent statewide federal-election history shows both why El-Sayed is favored and why the race is close. Democrats have kept winning Senate races, but Michigan has become almost perfectly balanced in presidential years.

The 2024 comparison is the backbone of the forecast. Rogers proved that he can get close to 49%, but he also ran roughly 1.75 points behind Trump on the same ballot. This makes him credible, not dominant.

Five polls have measured the confirmed matchup since the primary. All were fielded between August 6 and August 17, with samples of 800 to 1,215 voters for the relevant ballot test.

The simple mean is El-Sayed +2.6, but that gives too much weight to the Democratic-sponsored GBAO poll and Susquehanna’s high-end result. Published averages are much tighter: El-Sayed +0.6 at Hedge, +0.7 at RealClearPolitics, and +0.8 at 270toWin. My post-primary aggregation, which excludes most pre-primary information but discounts sponsorship and outliers, centers near El-Sayed +1.6.

Michigan’s recent Senate polling record calls for a correction. The final RealClearPolitics averages overstated the Democratic margin by about 1.8 points in 2018, 3.7 in 2020 and 2.0 in 2024 (2018 polling; 2020 polling; 2024 polling). I apply only a 0.6-point Republican correction. Three elections are a small sample, pollsters change methods, and Trump will not be on the 2026 ballot.

The national environment offsets that concern. Silver Bulletin’s August 26 generic-ballot average was Democratic +6.6, with its likely-voter adjustment at +7.6 (Silver Bulletin). Its freshest approval page put Trump’s net approval at −19.5, near his second-term low—not the materially less negative figure shown by some secondary summaries (Silver Bulletin approval tracker). These measures support a Democratic fundamentals advantage, but much of that information is already embedded in Michigan polling.

The coalition evidence is mixed. Fox found El-Sayed at net −13 favorability, compared with Rogers at −6, and found 54% concerned that El-Sayed was too extreme. Yet El-Sayed led by four among definite voters, seven among the extremely motivated and eleven among the extremely interested (Fox News). AARP found the reverse turnout risk: Rogers led voters aged 50 and older by 52% to 43%, and 86% of that group described themselves as extremely motivated, against 71% of younger voters (AARP).

Most major Democrats, including Stevens and Gov. Gretchen Whitmer, backed El-Sayed after the primary, which should improve consolidation (Washington Post). But the Hasan Piker controversy and continuing criticism from some Michigan Democrats arose on August 24–27, after every current poll had finished fieldwork (Associated Press). I treat the endorsements and controversy as roughly offsetting unmeasured developments.

Money favors Rogers at present. FEC reports covering January 1, 2025 through July 15, 2026 showed Rogers with $4.47 million in cash and El-Sayed with $2.55 million, although El-Sayed had raised more overall (Rogers FEC filing; El-Sayed FEC filing). The Senate Leadership Fund reserved $45 million for Michigan and added a $6 million August campaign (SLF; Axios). Still, Rogers allies asking AIPAC not to advertise for him suggests that some ostensibly helpful spending may revive El-Sayed’s anti-establishment message instead (Axios).

My ensemble gives most weight to a poll-based election-day margin near Democratic +1, then blends in a fundamentals estimate around Democratic +2 and a more cautious candidate-and-turnout component. The resulting expected margin is El-Sayed +1.25 with a 4.2-point standard deviation:

P(El-Sayed wins) = Φ(1.25 / 4.2) = 61.7%.

The middle 80% of the margin distribution runs from roughly Rogers +4.1 to El-Sayed +6.6. After forming this estimate, I checked markets: the two Polymarket contracts normalized to about 63% Democratic on roughly $597,000 in combined volume, while PredictIt’s bid-ask prices implied about 59% (Polymarket; PredictIt). That range validates rather than changes the independent estimate. Cook and Inside Elections both classify the race as a toss-up.

What's non-obvious

The macro environment is more Democratic than several widely circulated summaries suggest. Trump’s current Silver Bulletin net approval is −19.5, while the generic ballot remains D+6.6. That prevents a small Michigan polling lead from being read as evidence that El-Sayed is simply a bad favorite. A conventional Democrat would probably lead this race by more.

The turnout evidence does not point cleanly toward either party. Fox’s intensity screens favor El-Sayed, while AARP’s age composition favors Rogers. The race turns on whether Democratic enthusiasm can compensate for El-Sayed’s dependence on younger voters and weakness with older swing voters. Rogers’s financial edge matters less if the advertising merely reinforces opinions voters already hold.

Uncertainties

• No poll was fielded after August 17. The data does not measure the Piker controversy, late Democratic criticism or subsequent consolidation (poll archive).
• Late primary polling put El-Sayed about 7.4 points ahead, but he won by about one point. That raises the risk that likely-voter models overestimate his younger and lower-propensity coalition, though primary errors do not transfer mechanically to general elections (primary polling and result).
• The latest comparable campaign financial statements end July 15. Announced advertising reservations are not the same as ads ultimately aired or voters persuaded (FEC).
• Reliable general-election absentee and early-vote evidence will not exist until much closer to November 3. September likely-voter polls, especially from high-quality firms with full crosstabs, would close the largest current gap.