There were 100,021 people inside the Melbourne Cricket Ground for the NFL’s first regular-season game in Australia. The 49ers-Rams crowd was the seventh-largest in regular-season NFL history and only the fourth since 2000 to top 100,000.

The MCG, long associated with cricket and Australian rules football, was hosting an NFL game for the first time. Netflix carried the matchup in the United States, starting an opening week that would test the league’s increasingly scattered distribution model.

The NFL had moved its traditional Thursday night season opener to Wednesday, with the defending champion Seattle Seahawks facing the New England Patriots. The following night, the Rams and 49ers played in Melbourne, with Netflix carrying the game in the U.S. Then Sunday brought another change to the familiar schedule, as CBS and Fox both had national games in the 4:25 p.m. ET window. The changes showed up in the numbers.

The shift to Wednesday also immediately created a comparison with the previous season’s opener. In 2025, NBC and Peacock’s Thursday-night matchup between the Dallas Cowboys and defending champion Philadelphia Eagles averaged 28.3 million viewers. Weather delayed the game by more than an hour in the second half, after the audience had peaked at 34.3 million viewers, which likely affected its final average.

This year’s Seahawks-Patriots opener drew 25.1 million viewers across NBC and Peacock, a decline of roughly 11% from Cowboys-Eagles. The Wednesday game still cleared 25 million viewers for the fourth straight season and delivered NBC’s biggest audience among adults ages 18–49 for an NFL opener since 2021.

The lower overall figure, then, was not a sign that the kickoff game lacked reach, but it did give the league an early indication that moving the traditional Thursday opener came with a different viewing pattern.

Sports Business Journal’s Austin Karp looked at all the viewership data and handed the league with a reality check: “The NFL averaged 19.5 million viewers for its games in Week 1, down 13% from a record 22.3 million last year, when there were no competing Sunday afternoon national windows on CBS and Fox.

“Two years ago, NFL Week 1 drew 21.5 million viewers. Those three figures mark the three best NFL Week 1 audiences over the last decade, as the league has been able to take advantage of Nielsen measurement changes such as Big Data, improved out-of-home measurement, co-viewing changes and streaming improvements.”

Last season, Fox had the late-afternoon national window to itself. This year, CBS and Fox went head-to-head, with Packers-Vikings drawing 18.6 million viewers on CBS and Commanders-Eagles reaching 20.3 million on Fox. The two games were watched by a combined 38.9 million viewers, but because they were played at the same time, that audience was split across two broadcasts rather than concentrated on one.

The same thing happened earlier in the day. Fox and CBS both had 1 p.m. ET windows, with their games averaging 10.8 million and 14.1 million viewers, respectively. A year earlier, CBS had one national early window that averaged 17.1 million. Those additional games matter because the NFL’s 19.5 million figure is an average audience per game, not a count of unique people who watched football during the week.

The Australia game also makes the streaming part of the story harder to reduce to a simple loss for traditional television. The 49ers-Rams matchup averaged 18.5 million U.S. viewers on Netflix, peaked at 21.3 million and drew 21.7 million globally. It also averaged 4 million viewers among adults 18–34, the highest non-Christmas streaming audience for an NFL regular-season game. In Australia, it became the country’s most-watched NFL game.

Netflix’s expanded role is nevertheless part of a broader shift in where NFL audiences are being counted. The league now has games moving through traditional networks, Amazon Prime Video, Netflix and NFL+, while NBC and CBS also distribute games through their streaming services. That makes the old idea of one television network capturing an entire national audience less common, particularly when multiple games are available at the same time.

“It’s not the first inning anymore,” Fox Sports president of insights and analytics Mike Mulvihill wrote on X recently, taking a swipe at Netflix’s game window. “Putting sports exclusively on streaming has failed. [It] alienates longtime fans without reaching new ones.”

Mulvihill’s criticism points to one side of the streaming debate, but the Week 1 numbers do not establish that streaming caused the overall decline. The Netflix game itself drew 18.5 million U.S. viewers, while Amazon’s NFL audience also grew among full-season rights holders in 2025. The more measurable issue for Week 1 was that the NFL spread its inventory across more games, windows and platforms, making a single per-game average less straightforward to compare with a year earlier.

It’s not just about legacy media houses, though. All of these factors have even started to take a swing at alt-casts like The ManningCast.

“The ManningCast on ESPN2 pulled in 705,000 viewers, down 16% from the show last year in Week 1 and the lowest figure for the alt-cast in Week 1 since it began in 2021,” writes Karp.

The ManningCast’s decline also came on a night when the main broadcast was doing the opposite. Broncos-Chiefs averaged 23.2 million viewers across ABC, ESPN, ESPN2, ESPN Deportes and NFL+, making it the most-watched Week 1 Monday Night Football audience of ESPN’s era and its third-largest Monday Night Football audience since 2006. The ManningCast accounted for 705,000 of those viewers on ESPN2.

That contrast runs through much of the Week 1 data. The overall average was down, but the league still produced individual audiences of 25.1 million for the opener, 25.7 million for Sunday Night Football and 23.2 million for Monday Night Football. Week 1 remained the third-most-watched opening week of the past decade.

This is also where the 13% drop needs a little context. The 2026 Week 1 calculation covered eight games, compared with six in 2025. More games meant more viewing opportunities, but also more audiences being divided across individual broadcasts. When the two additional early-afternoon windows are removed, the NFL’s average rises to 21.9 million, 2% higher than the comparable figure from last year.

That does not erase the decline. Some individual windows were down, too. The Wednesday Patriots-Seahawks opener averaged 25.1 million viewers, an 11% drop from the previous year’s kickoff game. But it still produced NBC’s strongest 18–49 audience for an NFL opener since 2021. And the Sunday night Cowboys-Giants game reached 25.7 million viewers, up 4% from last year’s Week 1 Sunday Night Football opener and the show’s most-watched opener since 2015.

There is another variable in the comparison that is easy to miss: Nielsen changed elements of its Big Data + Panel methodology for the 2026–27 television season. The NFL’s chief data and analytics officer, Paul Ballew, had raised concerns before the season that some of the changes could have negative effects on sports measurement. That does not establish that the methodology caused this year’s decline, but it is another reason the 2026 figure should not be treated as a perfectly like-for-like measurement with earlier seasons.

For the NFL, the challenge is making that increasingly scattered audience work without making the path to the games harder to follow. That becomes more relevant as the season adds Amazon’s Thursday night schedule, more Netflix exclusives and the league’s other streaming options.

There have already been network outages on NFL+ and audio glitches on ESPN. Those incidents do not explain the 13% Week 1 decline, but they add another layer to the access question. If the league continues distributing games across more platforms and windows, reliability becomes part of the viewing equation as well.