Oracle Team USA ensured the regatta will stretch into another day, winning their fifth straight race to close the scoreline in the first-to-nine series to 8-6.
But with Oracle copping a two-point penalty before the regatta, today's 33-second win levels the score on the water to 8-all.
Team NZ still need that one precious race win, while Oracle now need three to successfully defend the America's Cup.
With the start of the race delayed 30 minutes while the race committee waited for the breeze to fill in, it was only possible to get in one race today in the tight broadcast window.
While Team NZ had the benefit of port entry at the start and defended the leeward end of the line well, Oracle accelerated quicker off the line, popping up on their foils more easily in the lighter breeze, and were able to roll over the top of Kiwi boat.
Playing follow the leader once again down the first run, Team NZ trailed by 13 seconds at the bottom mark as Oracle were able to foil more consistently in the patchy breeze.
While Team NZ tried to keep it close up the beat, they weren't able to engineer any passing opportunities as the confident US did not make any mistakes.
By the top mark Oracle had stretched their lead out to 19 seconds, and in complete control.
Although the official wind range for the event is 5-23 knots, regatta director Iain Murray did not want to start a race in under 10 knots after the teams failed to complete the course within the prescribed 40-minute time limit in light winds on Saturday.
That saw the race abandoned just as Team NZ were about to round the bottom mark for the final time with a 1.5km lead - much to the outrage of Kiwi fans, who thought they were about to witness a famous victory.
Murray said now they have the data modelling that shows teams cannot complete the course in time in anything under 10 knots, it would be silly to go ahead and start races if the breeze is not strong enough.
"We work hard at keeping abreast of the increasing performance of these boats, or as it might be the decreasing performance in the lighter winds. We now know that when we run the velocity prediction model over the course, we won't finish in under 40 minutes.
"So it's about the practicality here of setting these guys off on the course, knowing they're not going to finish," said Murray.
