Here we are again. Another year gone, another year yet to unfold. And unfold it will…especially in the streaming video industry. Just like last year, I’ve taken a look at many of the technologies being considered, tested, and perhaps even shunned and created not only a tech radar diagram but also updated (or added) my take on each of these technologies. Of course, at the SVTA, we will continue to keep an eye on how technologies in the streaming tech stack (or in consideration for use) are evolving throughout the year. But, just like my post last year, my predictions will focus on the technologies that streaming operators evaluate, test, and eventually implement (or abandon) as part of delivering a high-quality video experience at scale.
My Thoughts On Making Predictions
Well, I guess it’s nice to admit that I haven’t much changed my mind on this part of the exercise from last year. I still believe that Technological change does not happen overnight. Big companies are hard to move. Because new technologies can potentially interrupt revenue streams, testing and implementation can take months or even years. And that, inherently, impacts the ability to make predictions. Anyone who says, “next year will be when all streaming companies implement XYZ technology” doesn’t understand the technological inertia of an already established industry like streaming. What’s more, in some cases, it’s not the technology itself but the way in which it is implemented (i.e., cloud versus on premises).
Before I Get To Making Specific Predictions…
And…this hasn’t changed much either. The penchant for leaning into new technologies is still tempered by the pressure of profitability. I still see resources once allocated to emerging technologies (AR/XR/VR, Web3, Ultra-low Latency, Commerce) being shifted to technologies that directly support subscriber retention: service resiliency, scalability, QoE, etc.
2025: What’s Going Up and What’s Coming Down
To keep things consistent from year to year (I plan on revising this post until, well, I stop revising it), I will keep taking a simple approach to my predictions by addressing key technologies as either “coming up,” meaning there will be more resources spent towards implementation, or “coming down,” meaning there will be less resources allocated. I’m going to start with technology buckets and then dig into specific technologies.
| Bucket | Up or Down? | Thoughts |
| Security | Up | I’ve addressed this in a section below, but a lot of attention will be put on this in 2024. |
| Commerce | Down | The idea of purchasing directly from within a streaming video (or even through surrounding elements in a player) is definitely plausible. But, this just isn’t panning out right now to make it a financial priority when keeping subscribers is far more important. |
| Advertising | Up | With the growth of FAST, all streaming operators are realizing that they are leaving money on the table by not showing ads. But the implementation is complicated and hard to scale. Lots of effort going into this in 2024. |
| Interactivity | Down | Disagree with more or not, if the viewing experience sucks or the content library is lacking, no amount of interactivity, whether Web3 or otherwise, will help keep subscribers. |
| Delivery | Up | There is always emphasis on improving the delivery architecture. Interest in Open Caching continues to grow amongst operators but new delivery technologies, like MOQ, are also popping up to address specific delivery use cases. |
| Alternate Experiences [NEW for 2025] | Up | As live streaming, like sports, continues to grow in popularity, streaming operators (and rights holders) are looking for ways to differentiate their content or platforms from typical lean-back experiences. |
| Automation [NEW for 2025] | Up | Automating elements of streaming operations, such as identification of potential delivery issues and switching CDNs without human intervention, can help ensure a predictable, reliable, and resilient service. |
Note that the bracketed numbers before the technology below represent their placement in the tech radar diagram at the end.
[19] Cloud: Same as 2024
The cloud is still important in video workflows. While alternative technologies, like ASIC, have given streaming providers a reason to pull some of their cloud-based encoding back into the data center (because of cost), cloud-based resources are still a large part of the streaming picture and even more so as edge resources become critical to providing advanced (like SGAI, see below) and even ultra-responsive functionality. But there is much more thoughtfulness occurring with regards to what needs to be in the cloud and what doesn’t. Obviously elements of the workflow which need elasticity, like login and authentication, should be cloud-based. What 2025 will continue to see is much intentionality with regards to specific use cases.
[20] Cloud Side Note: Edge Compute: Same as 2024
This is where the cloud is trending up. The edge is an important part of the workflow, especially for those use cases, like ad insertion (SGAI), content steering, and data overlays (AR), which require ultra-fast responsiveness. The SVTA has even spun up an Edge working group to help address some of the best practices for utilizing edge-based resources (like ultra-low latency servers) within containerized environments.
[22] Cloud Side Note: Serverless Functions: Same as 2024
Although I am a big proponent of this technology, it’s just not really catching on. Perhaps that will change with the increased emphasis on the Edge, but, regardless, it’s still really in an experimentation phase.
[11] Cloud Side Note: Containers and Virtualization: Same as 2024
Much like cloud technologies in general, we will continue to see best practices, optimizations, and new approaches to deploying streaming video tech stack components in virtualized/containerized environments. This goes hand-in-hand with migrations to the cloud. And while we are seeing some reversal, such as with encoding moving back down to on-premises with ASIC and other technologies (as cloud encoding can get very expensive with multi-bitrate ladders across a fragmented device landscape), the continued push into DevOps is accelerating other aspects of virtualizing streaming video tech components.
[18] Cloud Side Note: Cloud Encoding: Slightly down
While cloud encoding has been around for a long time, the economics of it at scale are starting to point to a shift back towards on-premises encoding. It will be important to watch the cloud-encoding service providers and see how they address costs as streaming operators consider migration (maybe not all encoding, but a lot of it) back down to the data center.
[2] AR, XR, VR: Leveling Out (maybe a little up)
The biggest issue with adoption here is content. 2024 saw the launch of not only Apple’s spatial computing device (and their Spatial Video format) but also a new version of the Oculus, the 3S, with a price point that is well within reach for many consumers. One of the obvious use cases for VR is sports and the NBA has really learned into this. It will be interesting to see if the combination of a lower cost HMD (not Apple; I wanted to buy one but I just couldn’t justify it and the Oculus 3, not the 3S which has slightly lesser hardware, was reasonable and a solid piece of tech) and in-demand sports content can supercharge VR adoption. It’s interesting that we are also starting to see a resurgence of AR glasses hit the market and that the newer HMDs, like Oculus and Apple, provide an amazing AR experience.
[3] Ultra-low Latency Streaming: Up
Although the few years prior to 2024 ULL streaming was the topic-du-jour, 2024 saw a bit of a decline in interest. The problem: infrastructure. Running a ULL service, such as for sports betting, required different approaches to scalability. This resulted in sports providers running dual infrastructure: HLS for people watching, WebRTC (or similar) for those wanting to bet. This ensured that the operator could run a minimal ULL service and avoid having to provision too much extra capacity. But with MOQ exploding on the scene (more below), there is renewed interest because MOQ promises what WebRTC couldn’t provide: the same scalability as HTTP. It still remains to be seen if MOQ can do everything it proposes (field testing under pressure is a big test) but if it can, it will be a powerful addition to a streaming operator’s arsenal of technologies.
[24] MOQ: Way up
I have argued with people a lot regarding the viability of MOQ. I don’t believe it’s a silver bullet for anything but it does clearly fix the shortcomings of WebRTC so as a ULL approach, it seems like it will be a really powerful technology. But it still needs a lot of testing. In lab tests, the SVTA Networking and Transport working group, has shown that QUIC is not necessarily superior to TCP. It depends on the use case. For ULL? Absolutely. And while there are a few known implementations (such as Meta), there is very little adoption so far. 2025 might be a real year for testing this out.
[1] Web3: Slightly up
The big issue here is that Web3 encompasses a LOT of things: blockchain, NFTs, crypto, DePIN, etc. While I feel that the crypto side of things isn’t doing well (lots of operators/publishers have backed away from NFTs and crypto can attract a fair bit of nefarious actors), the DePIN side has really shown some potential interest. SVTA grant member Blockcast, for example, is attempting to marry Web3 technologies and approaches, like blockchain and DePIN, to innovate in the delivery space. Of course there are other plays in the market, like Livepeer and Pipe, which haven’t really panned out as of yet which is why this is only slightly up. There is definitely some interest here by the streaming operators to see how they can use Web3 technologies under the hood.
[8] Web3 Sidenote: Blockchain: Slightly Up
Blockchain is a legitimate technology. Distributed ledgers provide an immutable way of verification. That is powerful when it comes to things like streaming security,
[12] NEW for 2025: C2PA: Up
As the new “security kid” on the block (no, it’s not really about security but about proving content provenance which aligns more with Blockchain than anything else), C2PA provides a powerful way to assure the owner of the content is the agency posting it. Of course this is a great way to combat deep fakes and other video doppelgänger issues. SVTA member, EZDRM, has been demonstrating an incredible service similar to how DRM is handled for enabling content producers to take advantage of C2PA signing and verification. The cool thing about their solution is that it can be used to take down pirate streams, in real-time, that are leeching from the CDN.
[13] NEW for 2025: Common Access Token (CAT): Way Up
The streaming industry has been waiting for this standard to be published by CTA-WAVE and SVTA for a few years now. On December XXX, it was published and is now widely available. This will greatly simplify how CDN access tokens are handled significantly reducing the complexity for streaming operators in a multi-CDN delivery environment. Of course, it will require CDNs to implement (which many have already been working towards based on the earlier specs in the CAT standard) but downward pressure from CDN customers will definitely force this to happen. While this solution alone won’t solve piracy from the CDN, it will be a very important part of the overall puzzle to mitigating CDN leeching.
[14] NEW f0r 2025: Multicast: Slightly Up
If there is one thing that 2024 showed us, it’s that internet capacity (and the way ISPs peer and connect to one another to enable content to flow through the internet) is not what we need. With roughly about 1,700 Tbps of capacity, 300MM people watching a 3.5 Mbps video would take up all the available pipe. That’s bad. Multicast, which has been around for decades but hasn’t really taken off (because of an ecosystem issue which was covered in the SVTA white paper on multicast), is an obvious solution to the capacity issue. By removing all those duplicative unicast bits, capacity is increased naturally. And with the advent of new approaches like AMT and TreeDN, Multicast can be implemented without specific hardware or requiring end-client software. 2025 will probably be a year of testing this at scale.
[4] Context-Aware Encoding: Same as 2024
There is no doubt that encoding optimization will remain a big pain-point in 2025, especially as live events grow in size (such as Netflix’s boxing match which hit 60MM concurrents). Pressure to reduce bits while retaining quality will continue to drive innovation, like context-aware encoding.
[21] No-code Platforms: Same as 2024
There is still a big demand in the industry for this kind of software approach to defining and building workflows.
[23] QUIC: Same as 2024
The interest in QUIC is largely driven by MOQ, at least in the streaming industry. So with MOQ being so hot (at least for experimentation and testing), it’s natural for QUIC interest to ride along with it.
[5] Open Caching: Same as 2024
Operator networks continue to show interest in Open Caching. The primary reason is simple: they want (and need) edge-based caching. With providers like Qwilt offering a managed Open Caching implementation, it makes it even easier. Of course, operators will still host specific boxes in their networks, like Netflix OpenConnect, but operators really want a single solution for any kind of content, not just a single operator. Open Caching, as a specification, is also growing considerably, expanding out the available features accessible through the API specification.
[6] Server-Side Ad Insertion: Same as 2024
Even with FAST, as a business model, taking a hit in 2024, ads are still a critical piece of the streaming experience (especially in live sports). So this isn’t going anywhere. In addition, the Ad tech stack is complex and messy so there will be continued effort to bring services and technologies to bear that simplify things and improve interoperability.
[15] NEW f0r 2025: Server-guide Ad Insertion: Up
The new kid on the block and the best of both words: client and server-side. While SSAI is great at avoiding client-side ad blockers, it loses the dynamic nature of CSAI which can request ads based on user behavior. Enter SGAI. There will be a lot of experimentation and implementation throughout 2025.
[7] Real-time Watermarking: Same as 2024 (maybe up a little)
With the push on Edge resources, real-time watermarking will be a prime use case. Watermarking can be very effective as part of a multi-layered security strategy.
[9] Player Code-base Unification: Same as 2024
The SVTA Common Media Library (CML) is now a public repo that has been adopted by both dash.js and hls.js. 2025 will see a continued push to get the other JavaScript players, like Shaka and video,js, to adopt the CML as a way to mitigate player fragmentation.
[10] Content Steering: Same as 2024
With Apple moving their content steering out of the HLS format spec and into its own IETF spec, it’s clear that content steering will be an important part of streaming workflows. 2025 will probably see a lot of load and scale testing. What remains to be seen is if the industry can align the DASH content steering specification with HLS so that there is a single approach which doesn’t require slightly different implementations to use in a heterogenous format environment (which pretty much every streaming provider is right now).
[16] NEW for 2025: AI: Up
There is definitely a lot of hype about AI and how it can be applied to streaming video. From generating video to personalizing it to improving content recommendations and streaming video operations, AI is being tested and applied throughout the workflow. While it remains to be seen where it might have the most cost-effective impact (as connecting to OpenAI and other AI services can be quite expensive), there will continue to be a lot of assessment and testing throughout 2025.
[17] NEW for 2025: AI Sidenote: Generative AI: Way Up
While the general field of AI is showing potential for applications across the workflow, generative AI is perhaps making the biggest splash. Because it provides a way to generate new content, whether that is targeted advertising, personalized video, or even entirely new video content developed through a prompt, there is a lot of experimentation happening which we will continue to see publicized and explored throughout 2025.
Deeper Perspective: Technology Radar
Of course, the concepts of “going up” or “coming down” are pretty nebulous when it comes to actually doing something! As such, I’ve put together a TechRadar with all of the predictions and how they might be addressed in 2025. Note that the categories aren’t the best so I am working on hosting this to be able to change the category names but at least it’s a placeholder for now. Given that I have just pasted in a screenshot of the techradar below, I have added bracketed numbers to each of the predictions above to be able to find them in the image below.
A Final Thought About Security in 2025 (Same as 2024)
Right now, a lot of theft is happening. It may be actual files off a network or stream leakage (where pirates are stealing from CDNs and source URLs to rebroadcast through their own services; although CAT and C2PA could feasibly address a lot of this). Whatever the case, it’s happening and it’s costing content owners a lot of lost revenue (in the billions of dollars). And there are a lot of ways to protect content and protect the network but not everyone is doing what they need to and not every protection technology is as effective as it needs to be. 2025 will continue to see streaming operators leaning more heavily into security to protect valuable content assets. This will include not only continued work in improving DRM and watermarking, but also other aspects like CDNs and conditional access. It will be interesting to watch how C2PA is turned into a service (EZDRM is already demoing this very effectively) so that streaming operators can terminate rogue sessions against the CDN.