Why Streaming Giants are Buying Uncertainty
Streaming was built on freedom from the television schedule. Viewers could watch anything, anywhere and whenever they wished. Now the calendar is returning.
On Christmas Day 2025, Americans streamed a record 55.1bn minutes of television. Netflix and Amazon Prime Video together captured 22.5% of all TV use as three NFL games ran across the two services. Netflix then released new episodes of Stranger Things. It looked less like on-demand television than a broadcast schedule rebuilt inside an app.
This is not a reversal. It is what happens when streaming matures. Once every service has a deep library, the scarce product is no longer another programme. It is a reason to open the app now.
A selective diversification
Sport diversifies a streaming service, but not in the way a stock diversifies a portfolio. It adds a different form of consumer behaviour. Scripted series and films compete for time-shiftable leisure; a match starts at a fixed hour, is discussed in real time and loses part of its value once the result is known.
The paradox is useful. The schedule is predictable to the buyer, while the outcome is unpredictable to the viewer. Sports economics calls this the uncertainty-of-outcome hypothesis: evenly matched contests should be more attractive because the result is less certain. The empirical literature is mixed; close games do not automatically produce larger audiences. The narrower commercial claim is stronger. When play begins, nobody knows how it ends. That concentrates attention in a way a catalogue rarely can.
In L.E.K. research, 62% of fans said that without live sport they would move their attention outside streaming platforms altogether; the share rose to 79% among younger viewers. Sport therefore offers behavioural diversification: incremental viewing, simultaneous ad inventory and a recurring acquisition moment. It can hedge a platform against an indifferent release week. It does not hedge the cost of the rights.
Netflix buys moments; Amazon buys habit
Netflix's live portfolio remains selective. It has mixed weekly WWE programming with boxing, NFL games at Christmas and future global football events. The 2025 Lions-Vikings Christmas game averaged 27.5m US viewers, then the largest audience for a streamed NFL game. Netflix's July 2026 shareholder letter said advertisers were showing strong interest in a live slate that includes the Women's World Cup, an expanded NFL schedule, WWE and MLB events; it expects roughly $3bn of advertising revenue this year.
The objective is not to become a conventional sports network. A tent-pole event creates a mass arrival point that Netflix can place beside its entertainment franchises and advertising product. Christmas 2025 made the logic visible: football assembled the live audience, while Stranger Things gave it somewhere to go next.
Amazon has chosen recurrence. Its 15-game Thursday Night Football slate averaged 15.3m viewers in 2025, 16% more than the previous season. Amazon's figures, drawing on Nielsen, put the audience's median age almost seven years below that of NFL viewers on linear networks, with household income 28% higher. For Amazon, the same right can be subscription utility, advertising inventory, customer data and a weekly Prime habit.
Apple illustrates a third model. Its MLS partnership created one global, no-blackout destination. Yet after initially selling MLS Season Pass separately, Apple folded every 2026 regular-season match into the standard Apple TV subscription. The reasonable inference is that reach and the value of the core service had become at least as important as extracting a second subscription.
The bundle did not die; it fragmented
Streamers spent an estimated $12.5bn on sports rights in 2025, according to Ampere Analysis - about one-fifth of the $64bn global market. Yet the largest properties are not moving neatly from television to streaming. They are being divided across both.
The NBA's 11-year US agreements with Disney, NBCUniversal and Amazon, which began in 2025-26, put every national game on a broadly distributed streaming service. They also lifted the number of regular-season games on broadcast television to about 75 from a minimum of 15. Streaming won a central role; broadcast gained reach.
For fans, the result resembles the cable bundle after it has been broken into apps. Deloitte's 2026 survey found that 90% of US households have a paid video service and subscribe to four on average. Meanwhile, 41% of consumers had cancelled at least one service in the previous six months. The NBA made its own app a universal signpost to games across three media groups - a useful feature and an admission that discovery has become a problem.
Sport may reduce churn inside one platform while fragmentation increases churn across the system. Rights-holders gain more bidders. Fans inherit more logins, prices and questions about where the game is.
The rights bill is not a moat
A large audience is not proof of a profitable right. Sport is usually rented intellectual property: the contract expires, the auction returns and the price can rise. A scripted library may generate viewing for years; a live match creates most of its value in a few hours. The service must also carry peak technical capacity for the moment when everyone arrives together.
The return can appear in several places - subscriptions, retention, advertising, commerce or the wider ecosystem - which makes it strategically useful and financially difficult to isolate. Platforms disclose audiences more readily than standalone rights economics. For a service without advertising scale or a complementary business, an expensive sports package can be less a moat than a category error.
Sport is therefore selective diversification. A tent-pole can acquire users; a weekly package can create habit; a global league can deepen an ecosystem. Buying everything would merely replace programming risk with rights inflation.
The calendar becomes the product
The next phase is likely to look less like a streaming takeover than a re-bundling. Leagues need reach, platforms need habit and fans need fewer points of friction. Shared packages, free windows, aggregators and cross-service bundles will coexist with selected exclusive events. The NBA's hybrid architecture and Apple's wider MLS access already point in that direction.
The streaming war began as a contest over libraries. It is becoming a contest over calendars. Sport's uncertainty is the hook; its regularity is the business model. The winner will not be the platform with the most rights, but the one that knows exactly what job each right is being paid to do.
Research Note | Interpretations and Assumptions TSL interpretation: “Diversification” refers to attention, scheduling, acquisition and advertising - not a demonstrated reduction in financial volatility. Platform audience figures are drawn from company releases citing Nielsen and should not be read as standalone proof of profitability. The academic evidence that closer contests always increase demand is mixed. Facts are current to 27 July 2026.
Sources and Methodology
Nielsen | Streaming records and NFL viewing on Christmas Day 2025
L.E.K. Consulting | The streaming war for national sports rights
Buraimo & Simmons | Do sports fans really value uncertainty of outcome?
Netflix | Q2 2026 shareholder letter
Netflix | 2025 NFL Christmas viewership
Amazon | Thursday Night Football 2025 viewership and audience profile
Apple and MLS | Original MLS Season Pass structure
Major League Soccer | Every 2026 regular-season match included with Apple TV
NBA | 11-year media agreements with Disney, NBCUniversal and Amazon
Deloitte | 2026 Digital Media Trends
Ampere Analysis | Streamers' $12.5bn sports-rights spend in 2025