Investment Notes #8: Sirius XM a media business with a better economics than Netflix

For some time now, I have been receiving consistent notifications on Berkshire increasing its stake in Sirius XM Holdings (SIRI). In fact, the stock has been on my radar since 2017, when I first visited Omaha. I must have picked up this stock idea during one of the dinners. I did not buy it then. SIRI’s stock first moved up but has then declined and is roughly half compared to 2017.

Berkshire, however, has been gradually increasing its stake. It first bought Liberty Media tracker stocks in 2016 and started actively purchasing SIRI in early 2024. After buying about 5 million shares in December, the company disclosed more transactions on 30-31 January and 3 February (2.3 million shares). Berkshire is now the largest shareholder, with a 35.4% stake.

At first glance, this is a classic value trap. It is cheap (6x LTM PE) but has high leverage (Net Debt at $10bn is higher than the Mkt Cap, Net Debt/EBITDA at 3.7x) and, most importantly, stagnant revenue growth and bleak prospects in light of the growing popularity of streaming.

However, after spending some time researching the company, I think there is more to the company’s investment case. SIRI may lack significant growth potential, but it has a surprisingly deep moat (durable competitive advantages), while its business economics is in some ways better than Netflix's. It can “engineer” per-share growth through increased buybacks, with an additional upside from equity investments (worth almost $1bn).

Here is what I have discovered.

SiriusXM is a subscription-based satellite radio and audio entertainment company that provides ad-free music, live sports, talk shows, and news programming to millions of users across the US and Canada. The company operates two primary businesses: a satellite radio service, integrated into vehicles and standalone devices, and a streaming service, which offers on-demand content via mobile apps and smart devices.

1. Satellite coverage – the only service that provides uninterrupted radio coverage across the US and Canada. Traditional AM/FM stations or smartphones do not provide such access. Besides, consumers often must pay for the extra mobile data to download music or podcasts. No other company offers coverage via satellites. SIRI operates five satellites, each requiring an FCC licence.

2. Unique content – through Sirius XM, consumers can listen to live sports coverage and exclusive content such as the Howard Stern Show. SIRI’s choice of content is vast. Consumers can find the content they like regardless of where they are on the political spectrum or the music genre they prefer.

3. Pre-installed with a free trial – consumers do not need to take extra effort to buy a separate device or find an app online. SiriusXM is usually pre-installed in most cars. A 3-month trial is generally provided for new car owners (even if the vehicle itself is pre-owned).

If you search for customer reviews on SiriusXM, you will find a pretty low rating of about 2, with some negative comments around the company’s push to sell subscriptions or the high price of the service. However, this is misleading since a quick customer review search on Spotify, for example, also brings quite low ratings. I think neither of these companies pays much attention to their ratings on those platforms, as it is mostly the unhappy users who leave such unsolicited reviews.

OEMs and auto dealers are the most important partners for SIRI because they are financially incentivised to install SiriusXM into cars. If a customer decides to stay with the service after the trial period, OEMs and/or dealers receive a share of future revenue. SiriusXM is a reason for OEMs and dealers to try to charge more for a car while also earning a share of future revenue streams.

Being the only national satellite broadcaster and available in most cars, SIRI has a large customer base of 32.9 million (mainly in the US) and another 5.9 million Pandora active monthly users.

While the general perception could be that the radio business is a melting ‘ice cube’ in the age of streaming, the data does not support such a view, at least not for car radios.

SIRI has a very low monthly churn rate (derived by dividing the monthly average of self-pay deactivations by the average number of self-pay subscribers for the period). For 2024, it was at 1.6%.

Netflix does not disclose this number, but third-party sources estimate it to be over 2%, higher than SIRI. Spotify stopped disclosing it, either. Back in 2015-17, its monthly churn was at 5.5-7.7%. The latest estimates suggest it is over 4%.

A churn of 1.6% means an average customer stays with the service for 62.5 months (5 years and eleven weeks), much longer than 43 months at Netflix and just 25 months at Spotify.

This stickiness can be explained by SIRI’s unique live sports content and no alternative with the same breadth and depth of content in the automotive space.

Even more staggering is SIRI’s subscription price relative to Netflix. The basic plan starts at $10/month and can reach $30/month, with an average monthly revenue per user (ARPU) of $15. Netflix’s ARPU is just $11.7, although this is skewed by international subscribers. An average subscriber in North America spends c. $17 a month.

However, if you consider the amount of content Netflix has to create relative to SIRI and both companies charging roughly the same monthly fee, it is not hard to see which company generates more cash for its shareholders. Netflix competes with streaming giants like Amazon Prime and Disney, whereas SIRI has no direct rivals in the satellite car radio segment.

Speaking of content, Netflix spent $21bn on its videos in 2024, which is 54% of its revenue, compared to SIRI’s c. 40%. Clearly, the trend favours Netflix, which can spread its content costs over a larger audience, benefiting from an economy of scale. It is easy to see how Netflix will overtake SIRI in terms of relative content costs in the future. However, SIRI’s current advantage remains significant.

Thanks to its strong relationships with OEMS and auto dealers, SIRI does not have to spend much to acquire new customers as they get introduced to its services almost automatically when they first settle into their cars. In fact, it gets paid by OEMs for the equipment it provides which is part of the overall Revenue reported by SIRI. Equipment revenue was $182mn in 2024 (2% of total sales).

SIRI spends c. $15 to acquire a new subscriber who, on average, pays the same amount in the first month (1-month payback).

The lifetime value (LTV) of a typical SIRI subscriber is around $561, based on the following formula:

The value of SIRI’s subscribers has increased over time as ARPU appreciated, while the gross margin and the churn rate have remained stable. With declining acquisition costs, the critical value of LTV to SAC has rapidly increased (from 9x in 2013 to 39 in 2024).

SIRI has expanded its EBITDA margin over time, another testament to its business quality. The only cost item that has gone up relative to sales is the programming & content (from 16% in 2009 to 33% in 2024). All other costs have remained flat or declined relative to sales. Subscriber acquisition costs saw the most significant reduction.

How the margins look in the future is more critical for the stock, and here, the outlook is much less clear. It is worth flagging that SIRI’s EBITDA margin actually peaked at 36% in 2017, then dropped but remained relatively stable in the past six years at c. 29% (including share-based compensation expenses [SBC], or 31-32% if the SBC expenses are excluded. The EBITDA above includes share-based compensation and is about 200 basis points lower than what the company reports as it adds back SBC expenses.

The company has been on a solid growth trajectory since 2009, with a 15-year sales CAGR of 8.7%. Thanks to improving margins, EBITDA has outpaced the top line, growing at an annualised rate of 10.5%. Net income benefited from the increased financial leverage and increased by 13.7%

Buyback has boosted EPS growth further to 14.9%.

The last chart on the sources and uses of capital is quite telling. During 2009-2024, SIRI generated $23bn of cumulative cash flows from operations (CFO) and raised additional capital of $3bn (primarily via debt and preferred shares).

It has returned $20bn (!) back to shareholders via buyback and dividends, reinvesting just $5bn back in the business and spending $0.8bn on acquisitions. In other words, the past shareholder distributions are 2.5x more than the current Mkt Cap ($8.8bn).

The Board approved a new $1.166bn buyback programme in September 2024, following the completion of the “split-off” deal with Liberty Media. So far, SIRI has been slow at repurchasing its shares spending just $18mn on buyback up to 31 January 2025. The company plans to reduce its leverage from 3.7x to 3.6x by the end of 2026. SIRI should be more aggressive with buyback over time.

Such generous distributions point to a highly cash-generative nature of SIRI’s business model. As a subscription company, SIRI gets cash upfront, benefiting from the negative working capital. It carries almost zero inventory and pays its suppliers after c. 54 days.

The company also requires limited fixed assets to run the business (the book value of PP&E is just $2.1bn as of 31 December 2024).

On my estimates, the company’s ROCE is c. 20%, although this includes the book value of the FCC licences, which is a subjective estimate. Excluding them, the company generates closer to 100% return on its capital.

The main challenge for SIRI is the rising popularity of streaming services. This can be one reason behind the slowdown in ARPU. The latter grew by 40% from 2010 until 2020, but since 2022 has remained broadly flat since then. There is a particular challenge with younger consumers who prefer Spotify or other alternatives to SIRI.

The second issue is that SIRI service is mainly associated with and consumed while driving. It serves North American drivers and primarily middle-aged men. This is a large audience, but it is not growing.

To address the challenge of rising streaming alternatives, SIRI has been investing more in unique content and its app. As a result, its content spending relative to sales has been rapidly increasing (40% in 2024) and will likely keep climbing in the future.

The longer-term challenge is the rise of EVs. There were reports that EV manufacturers prefer not to install Sirius XM equipment to extend the reach and optimise their batteries as it adds weight to the vehicle. SIRI, however, has recently signed new agreements with Tesla and Rivian, suggesting EV companies are not too concerned with the weight issue.

The more serious issue is the rise of autonomous driving. While this is likely to be many years ahead, it is possible that consumers will have time and can watch full videos as opposed to just audio since they do not need to focus on driving anymore.

This is likely to take many years to become a reality. But even then, audio content will have its place, like today, when we listen to podcasts or audiobooks, even though we can always watch Netflix. SIRI will continue to exist, although its competitive advantages will erode in such a scenario as it will have to invest even more in content and possibly reduce prices to remain competitive.

The rise of shared mobility, or simply Uber/Lyft, is also a challenge, but personal driving has not disappeared so far. People will use such services more often in big cities and on special occasions keeping at least one car in a family for weekends and/or work.

SIRI trades on EV/EBITDA and P/E of 6.8x and 6.0x, based on 2024 results, and 7.1x and 7.6x, based on the latest guidance and consensus estimates for 2025. While definitely cheap, these multiples are not very useful since there are few direct peers. Besides, low P/E is partially a function of high leverage (3.7x).

An alternative approach would be to compare EV per subscriber, although this is also not perfect since some streaming companies have free ad-supported tiers (e.g. Spotify). SIRI also has ad-supported tiers, with ad revenue accounting for 20% of total sales. It has almost 40 million subscribers across its SiriusXM and Pandora services (33 million at SiriusXM). With an EV of $18.4bn, the market values one SIRI subscriber at $471, a modest 16% discount on our estimate of the subscriber’s LTV ($561).

My preferred approach is to estimate a sustainable (normalised) level of FCF that can be distributed to shareholders and compared to the current Mkt Cap (FCF yield to equity).

In 2024, SIRI generated $2.7bn of EBITDA in 2024 and $1bn of FCF to equity ($2.8/share). For 2025, management guides $2.6bn in EBITDA and FCF at $1.15bn. On these numbers, SIRI is currently valued at FCF yields of 12% and 14% (2024/2025E), respectively.

This is quite attractive. However, the total shareholder returns over the long term (assuming no change in P/E or other relative multiples) equals to the yield plus growth. So, if SIRI does not grow in the future, I expect to earn a c. 14% return if I invest in SIRI today.

In my base case, SIRI’s top line and margins should remain flat. The company will struggle to grow its market share. If inflation stays at 2-3%, SIRI would actually lose market share without raising prices. This can happen since streaming will become more attractive in the longer term (the young audience seems to prefer it already).

While SIRI will face increasing pressure to spend more on content, it can benefit from optimised G&A and marketing spend to sustain current margins.

In this base case, the company can “engineer” per-share growth through more aggressive buyback. For 2025, the company plans to cut debt by $700mn, which leaves just $450mn for shareholder distributions. With a quarterly dividend of $0.27, SIRI spends $360mn on dividends, meaning it can only afford to repurchase $90mn. However, starting in 2026, the company can accelerate buyback as its leverage drops closer to its target 3.5x level.

I expect about $600-700mn annual buyback from 2026, which at current valuation means SIRI will be able to reduce its share count by c. 7.8%, which is the equivalent of 8.5% growth.

It is also worth adding that SIRI sits on c. $1bn equity investments, which are comprised of a minority stake in SoundCloud ($75mn cost), a 70% economic interest in SiriusXM Canada (33% voting), and investments in clean energy technologies ($0.9bn). The latter generates tax losses and reduces the effective tax rate for the company. Most of this value should be reflected in FCF, so I do not see the need to add investments in clean energy technologies to SIRI’s valuation.

I think SIRI is attractively valued, with additional upside coming from growing share repurchases and equity investments.

However, without significant top-line growth and with the declining popularity of their service in the long run, I do not see sufficient upside in the stock. It may become more attractive if the share price drops or it finds new ways to grow subscribers, raise subscription fees or better monetise its subscriber base.