Netflix and Amazon have filed appeals with the Lazio Regional Administrative Court in Italy in respect to their CDNs. Last August, Italy’s Communications Regulatory Authority (Agcom) decided to extend the general authorisation requirement, under the Electronic Communications Code, on Netflix and Amazon, stating their CDNs should be regulated as telcos and pay network taxes.

As both Amazon and Netflix rightly point out, private CDN services should not be regulated like telcos, and Netflix states that “there is no legal basis to support it, and it is contrary to European and Italian law.” Agcom lacks a clear understanding of how private CDNs work; otherwise, they would know that CDNs are not electronic communications networks, and Netflix’s CDN does not carry any third-party data.

Agcom argues that CDNs are networks capable of transmitting signals and, therefore, are subject to authorization under the Electronic Communications Code. Netflix states that requiring the company to obtain a telecommunications operator’s license could result in servers being relocated elsewhere, requiring data to be delivered from further away and negatively impacting the viewing experience.

Rumors suggest that Agcom is considering adopting a regulation that takes into account the type of internet service in its decision and whether the EU, with the new Digital Networks Act, intends to move in the same direction.

For the past two days, I’ve been using Amazon’s new Vega OS-based Fire TV Stick, and I’m not impressed. Amazon claims the value of the latest model is “remarkably fast app launches,” but when tested against the Fire TV Stick Max, there’s no difference in speed.

I tested Netflix and Prime Video load times on both devices ten times each, and the average load time for Netflix on the Vega-based device was 14.03 seconds. On the Max device, the average time was 14.06 seconds. When testing Amazon’s Prime Video app, the average load time was 10.2 seconds on the Vega-based device and 11.3 seconds on the Max device. No customer is going to notice a fraction of a second in app loading times.

The Max device does boast better specs, featuring 2GB of memory and a Quad-core 2.0GHz processor, compared to the Select’s 1 GB and Quad-core 1.7 GHz. But according to Amazon, the Vega-based device doesn’t require as much memory or a faster processor because container apps in the cloud are streaming the actual app from Amazon’s cloud servers.

Setting up the device took nine minutes because the remote wouldn’t pair out of the box. It kept telling me to use new batteries, despite them being new from Amazon, as verified by a battery tester. The device required a software update as expected, followed by a four-minute restart. An hour after using the device, the remote decided to do a two-minute software update. The new Select device features a more rounded remote and has replaced the dedicated Hulu button with one for Peacock.

I’ve come across a few online mentions suggesting it is possible to sideload apps onto a Vega OS Fire TV device using a command-line interface similar to Android’s ADB utility; however, I don’t see how that is possible, and I haven’t seen anyone demonstrate it working in the wild. Unlike Android-based Fire TV devices, where you can turn on Developer Options in the About section, I’ve not found any way to do this on the Vega-based device. I “think” that only registered developers will be able to enable developer mode. Still, it’s unclear at this time, and I don’t know what the reference means in Amazon’s documentation when it lists a command for running an app that has been sideloaded.

For developers whose app is delivered via Amazon cloud app streaming, note that the Vega-based stick won’t support local device access, picture-in-picture, live TV EPG integration and continue watching row references. Amazon also notes that running the Fire OS app in a cloud container will cause the IP address seen by the developers’ content delivery network to not align with the IP address seen by the developers’ back-end servers. As a result, Amazon says local ad targeting may not function correctly, VPN inhibitors might be falsely triggered, other geo-fencing solutions might not function as expected, and High-traffic events might trigger denial of service attack warnings.

To address these problems, Amazon says apps can be configured to support IP tunneling, where all IP traffic is routed through the device. However, Amazon notes that enabling IP tunneling may increase lag in customer interactions and reduce the speed of content catalog browsing and other app interactions, which is why it is not enabled by default. Amazon advises developers to contact Developer Support if they encounter this problem, to request that their app be enabled with IP tunneling. This all sounds like a lot of work for the developers to ensure customers get the experience they expect, and I wonder how many developers will dedicate time to supporting a single Fire TV model.

For an understanding of how video playback in Amazon cloud app streaming works, Amazon’s developer page provides a detailed step-by-step: (copied word for word from Amazon’s page)

1. Customer UI interaction on device are sent to the app instance running on Fire OS in a cloud container.
2. A customer interaction with app instance results in a content play request.
3. App instance calls to developer CDN gets routed to on-device streaming player.
4. App instance uses MediaDRM and MediaCodec interfaces which are bridged to device streaming player.
5. Device media player fetches content from developer CDN.
6. Device streaming player sends non-AV payload to app instance for state synchronization.
7. Device streaming player decodes the AV payload using on-device secure video pipe.
8. Playback information is synchronized between device and media player in app instance context.
9. App instance streams UI to device client.
10. App instance UI is composited along with the video content and rendered on the device screen.

The Vega-based Fire TV Stick currently costs $14 more than the newest Fire TV Stick 4K model, although pricing varies based on whatever discount Amazon is running at the time. No consumer will choose the more expensive model simply because it runs on an OS owned by Amazon. Of course, Amazon’s pricing strategy could quickly change at any time, and I expect it will.

I believe it will enable Amazon to offer the cheapest Fire TV model ever, as Vega OS is a simpler platform that requires less from the device’s hardware compared to Fire OS. That should be just fine for consumers who don’t care and overlook the limited capabilities and limited app ecosystem. A significant limitation is that the Select device does not support live view picture-in-picture. If customers purchase it under the assumption that it’s faster than the Max device, they will be disappointed.

John Ourland reports that Apple is in the final stages of negotiations with F1, and wants F1 to shut down its streaming service F1 TV in the US as part of the deal. Rumors are that Apple would like to announce the agreement at the upcoming U.S. Grand Prix, taking place October 17-19 in Austin. Since last year, rumors have circulated that Apple would pay at least $150 million per year for the US rights, although this figure has not been confirmed. It was also previously reported that ESPN was still interested in Formula 1, but only if it could license select races, rather than the entire season.

Although many claim in their reporting that F1 TV is profitable, and Liberty Media may not want to shut it down, I have not seen the company release any data to confirm this. Liberty Media’s financial reports and earnings calls indicate that F1 TV is experiencing “continued growth in F1 TV subscriptions,” but the company has not provided specific profitability details for F1 TV. (Please correct me with a source in the comments if I am wrong.) F1 could also keep the F1 TV service running in the US, but change the offering by limiting the content just to race archives, or a similar format. A complete shutdown of the F1 TV service in the US may not be necessary to finalize a deal with Apple.

ESPN is giving up its US broadcast rights for Formula 1 at the end of the 2025 season, rights that Apple would acquire if a deal is reached. Halfway through the 2025 season, on July 8th, ESPN announced they were averaging 1.3 million viewers per race, up 7% over the season-to-date average from 2024. One of the most significant problems for any company licensing content is that Formula 1 races occur worldwide, and many take place early in the morning or overnight in the U.S. Outside the US, multiple companies hold the rights.

Sky Sports currently holds F1’s media rights for the UK, Germany, and Italy until at least 2029. And last year, F1 secured a new broadcast deal with BeIN Sports across the Middle East, North Africa (MENA) and Turkey through 2033, extended its contract with Viaplay in the Netherlands and Nordic countries through 2029, and signed a streaming platform, FanCode, in India through 2025. DAZN also acquired F1 rights in Portugal for the next three seasons.

What is one of Netflix’s key advantages? It offers no discounts. Netflix has never offered the ability to sign up for a year and pay a lower price. It has also never provided special pricing around the holidays or at any other time during the year, nor offered discounts to students or members of the military. Along with price increases, Netflix’s pricing strategy is a key factor in its ARPU’s significant rise over the years, while other SVOD platforms have struggled to increase their ARPU, thereby impacting their profitability.

With Disney, Paramount, Peacock, and WBD routinely offering pricing discounts of up to 50% or more, they lose money from every subscriber who signs up with special pricing. Their long-term strategy is a gamble that they can retain the customer once the discount expires and generate revenue from the subscriber over multiple years. This strategy doesn’t always work, and Netflix never has to take that gamble.

In Q4 of 2020, Netflix’s ARPU in NA and Canada was $13.84. At the end of Q4 of 2024, Netflix reported an ARPU of $17.26, with price increases in 2022 and 2023. For comparison, Disney’s ARPU for Disney+ domestic in Q4 2022 was $6.10, and by Q4 2024, Disney+ ARPU had risen to only $7.70, with three price increases in 2022, 2023 and 2024. Disney’s price increases were higher than Netflix’s, and more often, and yet they still haven’t seen the ARPU growth that Netflix achieved.

It is also worth noting that in 2022, Netflix introduced its new AVOD plan for $6.99. Within two years, more than half of new Netflix subscribers signed up for the plan in countries where it is available, according to Netflix. And yet, Netflix’s APRU still grew, even with an unknown number of consumers, like myself, downgrading from a $19.99 plan to the $6.99 plan. While Netflix faced significant criticism from analysts and the media in its early years of streaming, with many suggesting that the company was missing out on signing up more subscribers due to a lack of discounting, Netflix’s pricing strategy was correct all along.

One could argue that Netflix’s “discount” strategy was allowing users to share their account, thereby building a base of users that could be converted to paid users over time. As we’ve seen from Netflix’s numbers, this strategy has been successful. And while Netflix doesn’t offer a monthly discount, users can pause their account for a maximum of three months, with the option to extend the pause one month at a time before it ends. For those who take advantage of pausing their account, that’s essentially a discount; however, we have no data from Netflix to indicate what percentage of users have ever utilized this feature.

As we’ve heard from Disney during their earnings calls, Disney’s ARPU growth has also been negatively impacted due to “wholesale” deals, where they receive lower payments per subscriber from carriers that offer free Disney+ accounts with home internet and wireless services. As Disney recently highlighted in an SEC filing, “Wholesale arrangements have a lower average monthly revenue per paid subscriber than subscribers that we acquire directly or through third-party platforms.” While Netflix also has some wholesale deals with Verizon, T-Mobile, Xfinity (Comcast), and Dish, insiders at these companies tell me that Netflix doesn’t offer the same level of discount on its wholesale pricing as Disney, or in some cases, doesn’t offer any discount at all. Netflix knows its service is the premium brand and the most essential OTT service in any carrier bundle.

Netflix is very selective about who it partners with on wholesale deals. In contrast, Disney was offering wholesale deals to anyone who wanted them, including TV and streaming device manufacturers, Walmart+, wireless carriers, ISPs, and even Salesforce, which gave away a free 12-month subscription to Disney+ just for streaming one session from their Dreamforce event in 2021. Disney has also given away, or offered highly discounted Disney+ subscriptions to Disney fan club members, including a 3-year lock-in price of $3.92 per month, when the service launched in 2019.

We’ve also seen similar discounts with Peacock’s streaming service. In the lead-up to the 2024 Olympics, Peacock offered a special one-year deal priced at $19.99, representing a 66% discount from the list price. Comcast’s goal was to utilize the Olympics to expand the subscriber base for Peacock, with the hope that new subscribers would remain on the platform after one year and then pay the full price. The company gained 3 million net new subs in Q3 of 2024, but then saw no net new subs in Q4. With the special one-year deal expiring last quarter, it will be interesting to see the number of Peacock subscribers Comcast reports in its earnings on October 30th and whether Comcast will discuss how many discounted subscribers who signed up just before the Olympics are still active customers.

As consumers, we all like discounts. However, they would not have helped Netflix’s balance sheet, and the company knew it. Netflix has always played the long game with its business, and its pricing strategy has been executed perfectly to date. Here’s a list of APRU numbers from all the streaming services that disclose them.

A recent post at The Hollywood Reporter asserts that “most cord cutters and cord nevers are flocking” to YouTube TV, which isn’t accurate and overlooks the actual numbers, as detailed below. I often see this notion in the media that YouTube TV is picking up more cord-cutters than it actually is. The author also states that “Alphabet barely cares about YouTube TV’s financial performance,” which couldn’t be further from the truth, and every employee in the YouTube TV division I talk to would attest to the contrary.

Alphabet views the P&L of its YouTube TV business as one that needs to become profitable, and the author’s notion that Alphabet doesn’t care about YouTube’s financial performance is inaccurate. In January, YouTube raised the price of YouTube TV by $10 to $82.99 per month, due to rising content costs. In conversations I’ve had with executives at Alphabet, they have indicated that the YouTube TV service is not being used as a loss leader. If the company didn’t care about the financials as the author suggests, YouTube TV would not have dropped channels that they say cost too much to license. In addition, the pricing for the service has increased from $65 in 2020 to $83 in 2025, with the cost of the service more than doubling since its launch.

In February 2024, YouTube reported that it had “more than 8 million subscribers.” They have not given out a number since then, but many are reporting via third-party sources that YouTube TV has “about” or “close to” 10 million subscribers today. If the reported 10 million number is accurate, YouTube TV gained 2 million or fewer subscribers in 1.9 years. Over the same 1.9 years, more than 8 million cord-cutters have been reported in public earnings from Comcast, Altice, Charter, Verizon, Breezeline, DISH, and WideOpenWest, with an unknown number from DIRECTV, which no longer reports subscriber numbers. Adding in the not-yet-reported cord-cutting numbers from Q3 of this year, and the number of cord-cutting households in the past 1.9 years will be approximately 9 million.

An aspect that many may not be aware of, which is factored into YouTube TV subscriber numbers, is that YouTube counts consumers on trial plans who have not converted to paying members as subscribers. That’s why they never use the language of “paying members” but rather subscribers. Adding up the math, and assuming the 10 million YouTube TV members are accurate, this means that only 22% of cord-cutters have signed up for YouTube TV in the past 1.9 years, with that number dropping to under 20% if you add in those trialing YouTube TV for free. That would not be, as the author suggests, “most cord cutters,” but instead only one out of five cord cutters, or possibly even less, depending on how many trial subscribers YouTube TV has.

These numbers make sense, given that the cost of VMPD services has increased significantly over the past five years. For many, YouTube TV ($82.99 per month), Hulu+ Live TV ($88.99 per month), Fubo ($84.99 per month) and DIRECTV Stream ($84.99 per month) are no longer a simple replacement for cable TV, as they were when the services first entered the market. Over the past five years, more than 20 million households have cut the cord, with less than half opting for vMVPD services, according to public data provided by Disney, Fubo, DISH, and YouTube. That doesn’t take into account how many might have returned to cable for various reasons, nor does it consider those who leave a service and then come back (churn), which also needs to be acknowledged when discussing numbers.

All numbers are sourced from public filings, with the only estimate being for DIRECTV, which does not provide its own numbers but is estimated by Wall Street to have lost at least 1 million subscribers in 2024.

Disney announced that it will no longer provide ARPU numbers to Wall Street, following Netflix and Roku, which have also stopped reporting them. As this data continues to disappear from the market, here’s a breakdown of the latest ARPU numbers from services that continue to release them. For OTT services that had SVOD as their only business model, ARPU numbers provided valuable insights into their business. However, with nearly every OTT service incorporating advertising into its plans, as well as bundles offering wholesale pricing and some providing free streaming, a single ARPU number no longer provides as much insight into the business as it did previously.

Comparing ARPU data from previous years is challenging, as streaming services have increased pricing more frequently than before, and many have also expanded their services internationally with different pricing models and, in some cases, bundles offering promotional pricing. Some are unaware that streaming services receive lower monthly payments for wholesale subscribers. Disney, for instance, includes this language in their financials, saying, “Wholesale arrangements have a lower average monthly revenue per paid subscriber than subscribers that we acquire directly or through third-party platforms.”

Additionally, no OTT service details the percentage of users who have an ad-supported or ad-free plan, and only Paramount breaks out its share of DTC revenue that comes from advertising versus subscription fees. We also don’t know how many sign up for a service during special, highly discounted pricing offers or the percentage of consumers who buy the service via a vMVPD or aggregator, such as Amazon Channels. All of these factors significantly impact ARPU numbers, but we don’t know to what degree or for what percentage of users within a service. APRU is still an important metric, but it doesn’t provide a complete picture of how an OTT business is performing, which is why some companies have decided to eliminate releasing ARPU and subscriber numbers to Wall Street. All dates below are calendar quarters.

• Disney+: US and Canada ARPU, $8.09, Subscription plus ads (Q2 2025)
• Disney+: International ARPU, $7.67, Subscription plus ads (Q2 2025)
• Hulu: SVOD Only ARPU, $12.40, Subscription plus ads (Q2 2025)
• Hulu: Live TV+ SVOD ARPU, $100.07, Subscription plus ads (Q2 2025)
• ESPN+: US ARPU, $6.36, Subscription plus ads (Q1 2025)
• fuboTV: North American ARPU, $85.37, Subscription plus ads (Q1 2025)
• fuboTV: Rest Of World ARPU, $7.76, Subscription plus ads (Q1 2025)
• Paramount+: Global ARPU, $7.80, Subscription plus ads (Q2 2025)
• WBD: Global D2C ARPU, $7.14, Subscription plus ads (Q2 2025)
• WBD: Domestic D2C ARPU, $11.16, Subscription plus ads (Q2 2025)
• WBD: International D2C ARPU, $3.85, Subscription plus advertising (Q2 2025)
• Peacock: US ARPU, $10.02, Subscription plus ads (Q2 2025)
• Starz: US ARPU, $5.91, Subscription only, no ads (Q2 2025)
• Roku: Global ARPU, $3.45, [$41.49 trailing 12 months], Ads only (Q4 2024)
• Netflix: US and Canada ARPU, $17.26, Subscription plus ads (Q4 2024)
• Netflix: EMEA ARPU, $11.11, Subscription plus ads (Q4 2024)
• Netflix: LATAM ARPU, $8.00, Subscription plus ads (Q4 2024)
• Netflix: APAC ARPU, $7.34, Subscription plus ads (Q4 2024)
• Vizio: Global ARPU, $3.09, [$37.17, trailing 12 months], Ads only (Q3 2024)

Older, last reported APRU figures:

• iQiyi: China ARPU, $2.00, (Q4 2022)
• LionsgatePlay: India ARPU, $0.50, Advertising (Q4 2022)
• Pluto TV: Global ARPU, $1.64, Domestic ARPU, $2.54, Advertising (Q4 2021)
• Eros Now: India ARPU, Premium subscriber “in the range of $1.20 to $1.30”, Subscription (Q4 2022)

For all the price increases in OTT services over the past 12 months, many have seen YoY ARPU remain flat, or go down, due to wholesale pricing deals and, as Disney said during Q2 earnings, “lower advertising revenue.” No ARPU data has been released for AMC+, Acorn TV, Amazon Prime Video, Apple TV+, CuriosityStream, DAZN, NFL+, Sling TV, Tubi, YouTube TV, and numerous other services

While a private company, my thanks to Gilles Domartini, CEO and Founder of Cleeng, for allowing me to disclose details about the company’s revenue. Cleeng has been in the market for fourteen years, having raised a total of €10 million and ended 2024 with €20 million in revenue, with a 7% EBITDA margin, up from €5 million in revenue in 2021. The company counts 40 customers across four continents, targeting those with a monthly budget of at least €4,000. The company ended 2024 with 140 employees and reported that 65% of its revenue comes from the US market. These are great numbers for Cleeng, as they demonstrate growth and profitability without requiring a significant investment of capital.

During the call, I also had the opportunity to see Cleeng’s recently announced Pro platform, which enables content owners to launch a subscription management platform free of charge for up to 10,000 subscribers. You can check out more details on that platform here.