How TVK won Tamil Nadu in 2026: what the polls said and what happened

TVK (Tamilaga Vettri Kazhagam) was fighting its first Assembly election. It finished with the most seats and the most votes, but not with a majority on its own. Vijay was sworn in as Chief Minister on 10 May 2026 with the support of the Congress, which left the DMK-led alliance after the result, and outside support from four other parties.

Three things stand out. First, the polls: only one of seven exit polls called a TVK win, and the surveys taken before voting gave TVK between 2 and 10 seats. Second, the size: TVK did not squeak through, it won 43 seats by more than 20,000 votes. Third, the source: most of its seats came from the DMK, which lost 65 seats to TVK alone.

Before the campaign’s last stretch, a number of TV channels and newspapers published seat forecasts. These are opinion polls, taken well before polling day. The table below lists the ones that were released before the silence period began on 21 April, as reported by Oneindia. We have not audited how any of them were done.

Two of these also gave vote shares. Nakkheeran put the DMK alliance at 45.5%, the AIADMK alliance at 38% and TVK at 9.5%. Lok Poll had 40.1%, 29% and 23.9%. Nobody in this group had TVK above about a quarter of the vote. The actual figure was 34.9%.

It is fair to say the direction these polls pointed in was consistent: a contest between the DMK and the AIADMK, with the newcomer a distant third. That was also how the 2024 general election had gone, when the DMK-led alliance won all 39 Lok Sabha seats in the state.

Exit polls are different. They ask people who have just voted, so they should be closer to the truth. Seven agencies published exit polls on 29 April, according to The News Minute. Only one, Axis My India, said TVK would win.

Five of the seven put the DMK alliance ahead. JVC put the AIADMK alliance ahead. Axis My India was the outlier, and it was right about the winner. Look at its vote-share line too: 35% for TVK, against an actual 34.9%.

A caution on comparing these. Pollsters may count a party’s allies inside its share, and we do not know how each agency did it. TVK’s own share is what the Commission reports for its candidates.

Vote shares here are each party’s own, not its alliance’s. The DMK and the AIADMK fought both elections with allies, so their alliance totals would be higher. Our database does not yet hold alliance-level totals for 2026, so we do not print them.

Compare raw vote counts between the two elections and the picture is stark. In 2021 the DMK polled about 1.74 crore votes. In 2026 it polled about 1.19 crore, roughly 55 lakh fewer. The AIADMK went from about 1.54 crore to about 1.05 crore, roughly 49 lakh fewer. TVK, a party that did not contest in 2021, polled about 1.72 crore.

This is arithmetic, not proof that those 1.04 crore voters switched to TVK. Turnout changes, first-time voters, new names on the roll and removed names all move these totals. Without a survey of individual voters, no one can say exactly who moved where.

One more small sign of the mood: the number of voters who pressed NOTA fell from 3.46 lakh in 2021 to 2.00 lakh in 2026. We do not know why fewer people chose “none of the above”.

Tamil Nadu has not been redrawn since 2008, so a seat in 2021 is the same seat in 2026, and we can see who held each one before. Here is where TVK’s 108 seats came from.

The DMK held only 40 of its 133 seats, lost 65 to TVK and 22 to the AIADMK. It also picked up 15 seats from the AIADMK, plus a few from smaller parties, which is how it ended on 59. The AIADMK held 22 of its 66.

TVK was competitive almost everywhere it stood. Of its 233 candidates, 108 won and 74 came second, so it finished first or second in about 182 seats. When TVK won, it usually won well: the median winning margin was 15,969 votes, 43 seats went by more than 20,000, and 13 by more than 50,000. Only 19 of its wins were by fewer than 5,000.

The high-profile seats tell the same story. In Perambur, Vijay won by 53,715 votes. In Kolathur, the seat of the outgoing Chief Minister M.K. Stalin, TVK’s V.S. Babu won 82,997 votes to Stalin’s 74,202, a margin of 8,795. Edappadi, the seat of AIADMK leader Edappadi K. Palaniswami, stayed with the AIADMK by 98,110 votes, the widest margin in the state.

Regionally, the political writer Prem Panicker counted TVK’s wins as 44 of 69 seats in the north (Chennai, Thiruvallur, Kancheepuram), 27 of 68 in the western Kongu belt, and about 25 of 51 in the south, where the DMK won 23. He called the north a “hostile takeover”. Those are his regional groupings and counts, not official divisions.

No official body explains why people vote the way they do, so the reasons below are the views of named analysts. We have only included what we could read directly in their articles.

• An anti-establishment vote. Senior journalist R. Kannan, writing in The Federal on 7 May, called the result a “political tsunami” and an anti-establishment vote. His reasoning: voters wanted jobs and systems that work, and they were tired of the two familiar parties.
• Young voters. Kannan pointed out that around 2.5 crore of the state’s 5.67 crore voters are under 40. Panicker also described an urban youth swing.
• Vijay himself. Kannan’s view was that many TVK candidates were unknown and people voted for Vijay. He also argued TVK was the first party to rise to power in the state without paying for votes, which is his claim, not something we have checked.
• Fan clubs turned into booth workers. Panicker wrote that Vijay’s fan-club network converted into a booth-level cadre, and that this worked far better in the urban north than anyone expected. He described voters as wanting “correction, not rupture”.

These fit the seat map. TVK’s strongest region on Panicker’s count was the urban north, and its losses of the AIADMK and DMK seats were spread across the state. But fitting is not proving. We did not find a published post-poll survey with breakdowns by age, gender or community for this election, which is what would settle questions like “did women vote for TVK more than men?”.

Tamil Nadu went through the Election Commission’s Special Intensive Revision of the voter list before this election. According to The News Minute, the roll went from 6,41,14,587 names in October 2025 to 5,67,07,380 in the final list published on 23 February 2026, a net drop of about 74 lakh.

Turnout was about 85%. Our database shows 85.9%; other compilations based on Commission documents give 85.1%. We are not sure whether the difference is how postal votes are counted, so we say “about 85%”. Either way it was a heavy turnout, in an election where the winner changed.

What we can not tell you is whether the roll revision helped or hurt any party. We have not seen a credible seat-by-seat analysis for Tamil Nadu that links removed names to the result, and we do not have names-removed data by seat ourselves. If one is published, we will add it here.

TVK’s 108 seats were 10 short of the 118 needed for a majority. Congress, which had contested with the DMK, moved to the new government with its five MLAs, and four other parties from the DMK-led alliance gave outside support. The government then won a confidence vote in the Assembly with 144 votes in favour. Governor R.V. Arlekar administered the oath to Vijay and nine ministers on 10 May.

A result can be challenged in the High Court after the fact. For how that works, see our explainer on election petitions.

• Who voted for whom, by age, gender, caste or region. No post-poll survey we could find covers this.
• Why most exit polls missed. Some may have under-counted young or first-time voters, but that is a guess, and none of the agencies has, to our knowledge, published a post-mortem.
• How much the Karur tragedy or the voter-list revision mattered, if at all.
• The alliance-level vote shares. We only report each party’s own share.

Seats, votes, margins and seat flips are our own analysis of the results published by the Election Commission of India, from the same data that powers the Tamil Nadu 2026 election page on this site. Vote shares are each party’s votes divided by all votes counted, including NOTA. Opinion-poll and exit-poll figures come from the outlets named, and analyst views are credited to the person who wrote them. If you spot a number that does not match the Commission’s, tell us on the contact page.