Amazon just recently announced that users with a Fire TV can sign up for Luna, their new cloud gaming service. Although cloud gaming has gained more traction as the world's GPU shortages continue, no clear winner has emerged from the pack. Google, Microsoft, Amazon, and NVIDIA are all trying to position themselves well while Valve continues to be the de facto gaming platform with a business model similar to iTunes' old approach.

Today's cloud gaming landscape has a lot in common with the music streaming wars of the early 2010s. Apple was the dominant player with a resounding success in iTunes as audio shifted from physical to digital, but Spotify was able to come in and take the lion's share of premium subscribers. In 2020, Apple reported having over 70 million active users on Apple Music while Spotify announced a whopping 133 million premium users around the same time.

Apple, Google, and Tidal all tried to keep up with Spotify as most people shifted away from having a personal music library and embraced streaming services. Each had a somewhat different approach that ended up falling short of the seamless experience Spotify has today, and I think it's worth looking at how history might repeat itself again.

Spotify's success

The first thing that stands out when looking at the competition is how long Spotify was able to establish itself as the only streaming service with a unique business model. From 2011 to the end of 2013, Spotify had to compete with other newcomers to the space (e.g. Rdio, a rebooted Napster) but didn't have to worry about beating iTunes at their own game. The biggest shift in their approach was to offer a full catalog of music that is now fairly common with SaaS models for consumer products. The app itself was a smart blend of a familiar modern take on the classic media player while completely changing how things worked behind the scenes. Users still had their playlists, a view of all their songs, seamless queueing and shuffling of tracks, and easy access on their computers and mobile devices, so the core user stories hadn't changed. The big shift was asking users to not think about ownership of songs anymore. Although there was some initial hesitation, Spotify offered clear upside with no longer having to buy individual songs, not having to worry about disk space, and not having to manage your own library with endless tags for genre, artist, name, album, etc. The conveniences the platform simply outweighed abstract concerns around "ownership" of digital assets.

Where others fell short

Apple had a commanding lead in the music industry after leading the shift away from physical media to a (legal) digital library. Their one-two punch of a comprehensive digital catalog combined with their best-in-class hardware created a business that steamrolled the competition. Even when they launched Apple Music, they reached 10 million streaming users in only 6 months (it took Spotify 6 years to reach the same growth). Their biggest gap was giving Spotify almost 4 years to get ahead in the streaming space. Like many market leaders, Apple was clearly ahead of the pack, which meant they were more concerned about iterating and improving their existing offerings instead of trying to stay ahead of new, disruptive approaches to their industry.

Google Play Music tried to appease users by appealing to their existing music libraries and making them available for online and mobile streaming while slowly building up a catalog of songs and albums available for purchase. Much like Apple, they let Spotify have years to establish themselves as the premier true streaming service while their more novel approach of encouraging users to bring their libraries to the cloud didn't quite take off. The fact that the primary player was in-browser and was mixed in with YouTube Premium (then Red) led to what is now a common concern with Google products of whether or not the company was taking this seriously. Today, those concerns have been validated with the ongoing shutdown of Google Play Music and the transition to YouTube Music. I get the feeling that Google could have been a major contender in this space given their absurdly strong engineering bench and the widespread adoption of Android internationally, but their lack of focus and commitment seemed to be the main thing missing from their strategy.

Tidal was perhaps the most unique contender in the music streaming wars as a platform specifically for audiophiles with high quality bitrates and exclusive partnership, but they were a latecomer and launched in 2014. Tidal was the only platform that presented itself as a luxury streaming service, which led to them seeing the same level of adoption that luxury products have. Higher quality streaming means higher storage and bandwidth costs. The increased cost per play was pushed back to the user, leading to a more expensive product. Tidal tried to justify the higher prices by having exclusive, high-quality artists available only on their platform (e.g. Beyonce, Jay-Z, Kanye), but, in the end, Spotify's larger collection with a high enough quality bar to appease the masses ended up winning. Today, all of the Tidal-exclusive artists have shifted over and are available on Spotify. The main lesson from them is that having the highest quality doesn't matter if the competition's quality is still pretty good.

Cloud Gaming

Fast-forward almost 10 years, and it feels like we're watching a sequel with some of the same characters and a rehashed plotline. Steam today looks a lot like Apple did in 2011 with a dominant position as the default platform for one-time purchases and library management. Although Steam has seen an increase in competition from the likes of Epic, GOG, and other competitors, it easily reaches over 25 million concurrent users on any given day. For comparison, Fortnite reached 12.3 million concurrent users for their Travis Scott concert (their biggest promotional event ever). The main draw with Steam is that it's the best place to go to have all of your games in one place and to purchase new games as they come out.

Although Steam isn't going anywhere anytime soon, they haven't made clear strides in streaming games as a service or changing their fundamental business model. Much like Apple, Steam led the charge in shifting everyone away from physical goods to a digital library and have been happy to ride that train of cash all the way to the bank.

One point where the comparison between gaming and music falls short is the scale of the engineering problem. Spotify's engineering team has done a phenomenal job scaling their business globally, but the limitations around music streaming aren't as complex as cloud gaming. Audio streams require fast access to sequential files and can allow for some delays when skipping around tracks. The relatively small size of each track makes the data footprint per user somewhat manageable for a given session.

Cloud gaming requires enough installations of a given game to meet demand. Games today are anywhere 1-2 GB to 40+ GB for major blockbusters, compared to the measly 3-7 MB per audio track. Each game session also requires actually running the game, which is a fairly CPU, GPU, and memory intensive task. Most modern servers can easily stream multiple files at once, but the idea of a single machine running multiple copies of Cyberpunk 2077 is a bit of a stretch. Despite all of these complications, there are still a few companies that could try and build a service that meets the needs of a large portion of gamers.

Enter the new players: Google, NVIDIA, Microsoft, and Amazon

Google Stadia

For almost two decades now, Google's name has been synonymous with engineering excellence, creating some of the world's most efficient storage solutions in BigTable and Spanner while acquiring and scaling the world's most successful streaming service in YouTube. Supporting cloud gaming requires infrastructure that is an order of magnitude higher than music streaming, but Google has continued to make massive strides in cloud technology across their products. I believe that Google is more than capable of making this work, but it remains to be seen if they are truly committing to cloud gaming.

Just last week, Google announced the shutdown of its internal game studio, which raised some questions around how seriously Google wanted to take gaming. I don't think the studio itself is vital to the success of Stadia since they can still see real success here if they can get publishers and studios to sign licensing deals through Stadia. The SaaS model where they can create hosting options for any game developer is a viable strategy, if they want to pursue it. Unfortunately, their business model relies on purchasing each game through Stadia, which creates an extremely high barrier to entry for people who might want to buy a game on a platform they already trust (like Steam) but are curious about testing out cloud gaming.

GeForce Now

NVIDIA's recent dominance as the premier GPU provider in recent years has led to them being synonymous with gaming and, more recently, cryptocurrency mining. The increased appetite for cloud gaming comes in part from supply chain issues they and AMD have had. It will be interesting to see if NVIDIA can try to address frustrations they are largely blamed for by spinning up another, potentially lucrative, business line.

NVIDIA has taken a novel approach by whitelisting a set of games through existing platforms. Users who own those games and are members can play on GeForce Now without having to acquire a new digital license, a huge benefit over Stadia. This greatly decreases the barrier to entry for NVIDIA's target audience, and their focus on existing NVIDIA GPU owners makes it easy for them to target their core demographic. The main drawback with NVIDIA's approach is that they clearly aren't ready to scale. For the launch of Cyberpunk 2077, one of the most hotly anticipated games in recent years, NVIDIA's cloud had queue times of over an hour to start playing, even for people paying for the premier tier of GeForce Now. They also have hard limits on the length of each play session.

It feels like NVIDIA has learned a key lesson from the music streaming wars while making the same fatal mistake as so many others. They have a low barrier of entry where people don't need to feel like they have to re-invest in a brand new platform since they can bring their games (or a subset of them) with them, but the experience itself feels fundamentally off. People play games for leisure and don't always schedule time for them. This is especially felt by gamers who might not have as much free time and can't wait more than 15-20 minutes to start playing. Add in forced session limits and having to re-enter the queue and you have a platform that clearly feels worse than the current gaming experience for most people.

Xbox Game Pass

Microsoft gets an honorable mention here because of how much they are investing in gaming, but they aren't a streaming service. Yet. Game Pass provides the same buffet that Spotify does, but the main question is around quality. Most modern AAA games aren't available on Game Pass immediately, and the biggest names are either from indie companies or are hits from a few years ago (e.g. Final Fantasy XV launched in 2016 but was one of the premier Game Pass games). It remains to be seen whether Microsoft can continue adding heavyweight to their lineup, similar to how Spotify eventually was able to come to an agreement with the Taylor Swifts, Beatles, and Led Zepplin's of the world. They have the capabilities to build out their cloud solution on Azure and are clearly investing in improving their catalog with their relatively recent acquisition of ZeniMax Media. What remains to be seen if they'll be willing to pivot to cloud gaming or if their catalog of games with a monthly subscription will be enough as the supply chain for GPUs recover.

Amazon Luna

Although Amazon is entering the gaming space a bit later than the competition, nobody has succeeded in creating a healthy cloud gaming ecosystem yet, so their delay might not matter. Google and Microsoft are juggernauts in their own right when it comes to self-hosted infrastructure, but Amazon's AWS has long been the reigning champion of cloud-hosted services. They are famous for relying on the same infrastructure they offer to enterprise customers, including Netflix, Airbnb, and many others. Their track record should speak for itself, but I think it's worth being explicit that, if Amazon can't scale to meet cloud gaming's needs, nobody can.

Luna currently offers a catalog of games for a standard monthly subscription, but they also have a large question mark of whether they'll be able to put together a compelling enough repository of games once they are fully live. That being said, if they decide to go all-in on licensing and find a price point that is reasonable enough for all parties, they will have all the advantages of Xbox Game Pass with the convenience and ease of access that comes with cloud gaming.

Summary

At this point, there is no clear winner, and any of these players could completely shift their strategy. Even Steam could decide to heavily invest in having their own cloud, potentially through a partnership with Amazon funded by their massive war chest from years of high-profit sales on digital games. If the music streaming wars were any indication, creating a seamless experience for people that is at least as convenient as gaming is today is a must-have for anyone to succeed. I strongly believe that Stadia's current cost of entry and GeForce Now's queues and playtime limitations are too high of a cost to convince people to switch over from sticking with Steam, Epic, and GOG. If Amazon is able to beef up their catalog or Microsoft decides to move into cloud gaming, they each have a good chance at getting a significant lead over the competition. It might be strange, but I'm honestly betting on one of them leaping ahead of the pack and gaining an advantage similar to the one Spotify enjoyed for a few years before everyone else caught up.

At the same time, the entire industry could shift. Riot Games sidestepped this entire conversation by building a game as a service with its own launcher and a low enough set of technical requirements to make League of Legends accessible. Many others outside the MMO genre have followed suit with Genshin Impact as the most recent massive success. While some of us are busy thinking about the future of cloud gaming, we might miss a fundamental shift in the gaming industry itself if more and more games end up being standalone services and capture the majority of industry.