As polls closed on Tuesday night, prediction markets indicated President Trump’s endorsed Attorney General Ken Paxton had a 97% chance of defeating incumbent Senator John Cornyn. The actual vote count quickly revealed an even more decisive outcome.
With just 25% of votes reported, Paxton had already built a staggering 22-point lead over the three-term senator. That margin widened significantly as election-day returns continued to pour in.
One late batch of results showed Paxton securing 87% of the vote in a single county—a figure that stunned even MSNBC’s Steve Kornacki during his broadcast. Kornacki noted the margin as “74 points” and highlighted it came with “full knowledge of President Trump’s endorsement.”
Cornyn’s campaign confirmed he would not attend his own watch party, opting instead to monitor results “privately” in a venue reportedly limited to about 50 chairs.
The race had become a direct test of whether Republican incumbents could withstand the political pressure of a Trump-endorsed challenger. President Trump’s late endorsement gave Paxton an immediate boost among Republicans who prioritize loyalty to the former president over established Senate figures.
Cornyn entered the contest with significant campaign resources, having spent roughly $90 million on advertising since last year, much of it aimed at countering Paxton’s rise. However, the timing of Trump’s endorsement proved pivotal as conservative voters questioned whether Cornyn adequately represented the MAGA base.
Early returns suggest that concern became a major factor in the election night outcome. With numbers like these, John Cornyn’s chances have been effectively eliminated.
This runoff carries national implications beyond Texas. The winner will face Democratic state representative James Talarico in November, and the result could shift Senate dynamics as Democrats seek to regain control of the chamber.
For the Republican Party, this race represents a critical test: whether primary voters continue to reward Trump’s picks over Washington incumbents.