ColdSpark

2026 Senate Battleground Monitor

Enter your email to view the live dashboard.

Your email is used only to track dashboard access and won't be shared.
ColdSpark
Public Affairs Intelligence

2026 Senate Battleground Monitor

Public polling and campaign ad spending across the twelve most competitive U.S. Senate races, drawn from AdImpact via the Wesleyan Media Project and Punchbowl News (spent-to-date and booked-through-Election-Day), live FEC disclosure (independent expenditures, candidate media buys, party-coordinated spending), Google/YouTube's political ad transparency data, and the latest published polls.

Data as of Oct 20, 2026
Betting Markets

Who controls the Senate after November?

At A Glance

Where each side is winning

Shaded for the twelve tracked races only — every other state is shown for context and isn't scored. Hover a shaded state for the exact margin.

TV Ad Spend Advantage

US States Map (Albers USA)
Strong D
Strong R

Polling Advantage

US States Map (Albers USA)
Strong D
Strong R
Republican candidate Democratic candidate (Nebraska: independent Dan Osborn) Undecided / other Ratings: Cook Political Report, Sabato's Crystal Ball, Inside Elections · Markets: Polymarket
Race by Race

Twelve races will decide the majority

Reading this dashboard

  • Polling bars show each race's most recent public poll or polling average, not a projection. A stale tag flags any race without an independent poll in the last several weeks.
  • Betting markets show Polymarket's live implied win probability, checked once at the "as of" time above — these move continuously and thinner-volume markets (noted per race) can swing on small trades, so treat them as a sentiment check, not a forecast.
  • Inside Elections notes call out the race's current IE rating and, where IE has published a rating change or standalone analysis in roughly the last 45 days, what changed and why. Where no recent single-race piece exists, that's stated explicitly rather than substituting another outlet's take.
  • Booked Aug 1 → Election Day (the line under each spend bar) is a different, forward-looking measure: AdImpact's tally of ads already aired plus future reservations from Aug 1 through Nov 3, by party side, as reported by Punchbowl News on Sept 22, 2026. It shows where each side has committed money for the stretch run — Republicans lead in eight of the nine races it covered (a $244M combined edge), Georgia being the lone Democratic exception; Kansas, Nebraska and South Carolina weren't in that tally, so their cards show the best available substitute or say none exists — and it is not additive to the spend bar above it, which counts only money already spent in a fixed window.
  • TV ad spend bars now use AdImpact all-media spend (broadcast, cable, CTV, digital, radio, satellite) via the Wesleyan Media Project's tracked window (now Sept 28–Oct 11, 2026, from WMP's Oct 15 release; previously Sept 8–27 and Aug 10–Sept 13) for every race — replacing the old one-time AdImpact snapshot with a figure that refreshes as WMP's releases roll forward, so the R/D split and the "Tracked Ad Spend" headline stat move with each update rather than staying fixed at a single August reservation date. Each card's source line adds two further layers, both pulled via API rather than press summaries: FEC cycle-to-date disclosure — independent expenditures (Schedule E, deduped for dual-listed and amended filings), candidates' own ad-like media/mail disbursements (Schedule B, lagged to each committee's last filed report — Q3 reports are due Oct 15 and the pre-general report Oct 22), and, post-NRSC v. FEC, party-coordinated spending (Schedule F/Form 9 electioneering) where present — and a Google/YouTube political-ad footnote for the most recent week, joined to campaigns via FEC committee ID. The Google figure is a live digital-only slice and is not additive to the WMP total above, since WMP's all-media figure already includes Google/digital spend. Combined tracked spend is a floor, not a ceiling — FEC disclosure in particular lags real-time spending by weeks and understates groups that haven't yet filed.