The 2026 NFL season has already drawn tens of millions of viewers across platforms. The season opener on NBC and Peacock saw 25.1 million viewers, and 18.6 million tuned into Amazon’s “Thursday Night Football” debut. Fans have more ways than ever to watch live sports, and they also expect those streams to work, no matter how many people are watching.
However, the industry is running into an infrastructure challenge. Streaming may have revolutionized where we watch football, but it hasn’t yet solved how we efficiently deliver it. The internet wasn’t built for millions of people to watch the same live moment at the same time. As leagues and platforms continue pushing viewers online, the systems powering live streams are becoming strained by scale, latency, and cost.
Fixing that problem may require revisiting one of television’s oldest strengths: broadcasting. By combining broadcast infrastructure with modern cellular networks, the industry has an opportunity to rethink how live video is delivered at scale.
Every network evolution starts with a new application
The network has gone through this kind of transition before. In the late 1990s, consumers began spending hours online through dial-up connections. But that created real issues for telephone networks that were built around voice calls. Data sessions lasted much longer than traditional phone calls, putting new pressure on infrastructure designed for an entirely different application. In response, the industry evolved from a voice network into a data network.
Today, we’re seeing a similar shift, with the data network having quietly become a video network. Video now accounts for roughly 75% of mobile traffic. Yet, much of the infrastructure underneath still treats live video as another form of data that needs to be delivered from one point to another. Live sports are making that mismatch harder to ignore.
Live sports turn a shared moment into millions of streams
Live video is fundamentally different from traditional video on-demand. A hit series can be watched over an entire weekend, and the demand spreads out over time. A live NFL game, on the other hand, results in millions of viewers hitting play within a narrow window, expecting to see the exact same content simultaneously. What looks like one shared experience to fans becomes millions of separate delivery requests for the network.
This is because much of internet video delivery still relies on a unicast, or a one-to-one, model. Take the season opener mentioned earlier, for example. If all 25 million people had streamed that game, the network would have duplicated the same video feed 25 million times to reach every screen.
When networks inevitably become congested due to this, adaptive streaming can lower the bitrate or resolution to get viewers watching quickly. But during a live game, that can mean the picture suddenly gets worse at the exact moment fans care about most. These aren’t necessarily failures by streaming platforms. They are signs that the delivery model is being pushed by a new type of workload.
More capacity isn’t the whole answer
The traditional response to growing video demand has been straightforward. We’ve added more capacity, fiber, and bandwidth. Those investments remain important, but a bigger pipe doesn’t solve the underlying problem when millions of people are requesting identical content at the same time. The issue is how efficiently the network can distribute that data. That distinction becomes especially important at the edge of the network, where content ultimately has to reach the viewer.
The bottleneck has moved to the edge
Every duplicate stream consumes capacity. When the same live event is being watched by millions of people in the same market, continuing to deliver that content as millions of individual streams becomes difficult to justify. This is where broadcast earns its place again.
Enter ATSC 3.0, a modern broadcast standard that bridges traditional TV signals with internet-protocol data. Instead of creating a separate delivery path for every viewer, ATSC 3.0 uses an advanced “one-to-many” architecture to send a single, high-capacity stream to everyone at once. That is what makes broadcast uniquely suited for live events.
The idea isn’t to replace cellular, as the two networks are built for different jobs. While cellular is designed for individualized, two-way connections, broadcast is designed for mass delivery. For live sports, that creates an opportunity to use each where it makes the most sense. Broadcast for the shared video everyone is watching, and cellular for the individualized data and interactions around it.
Broadcasters have already rolled out ATSC 3.0 signals to more than 76% of U.S. households, so the bones are already in place. Building on it is the next step: getting ATSC 3.0 reception into more of the devices fans watch on, and getting broadcasters, streamers and wireless carriers working from the same playbook.
The next evolution of sports streaming infrastructure
The demand for streaming live sports is undeniable, from the NFL to the Olympics and World Cup. The question is whether the delivery architecture will adapt to how audiences today actually watch.
It has to. Leagues and platforms that keep moving live audiences online without rethinking delivery will keep paying more for a worse picture at the biggest moments. The next era of sports streaming won’t be won with bigger pipes. It will be won by using the right network for the job.
Broadcast didn’t disappear because of streaming. Streaming is about to need it.
Conrad Clemson is CEO of EdgeBeam Wireless and a longtime technology executive focused on broadcast-IP transformation, network infrastructure, and large-scale content delivery.