Steam Earned an Estimated $4.3B in 2017, but Benefits Flow to Handful of Titles

Steam Earned an Estimated $4.3B in 2017, but Benefits Flow to Handful of Titles

Valve, as a private company, doesn’t have to disclose its earnings, governance, or a great deal of other information about its own operation. Luckily, there’s third-party analytic firms like SteamSpy to fill in the gap. According to Sergey Galyonkin, the firm — which rose to fame with games like Half-Life and Team Fortress 2, and which now controls the overwhelming majority of the PC digital download space — earned a massive $4.3B in 2017, $800M more than the estimated $3.5B it earned through the same period in 2016. While SteamSpy’s figures can’t account for every bit of revenue Valve might have earned, the data it’s missing from DLC and microtransactions only understate the true size of Valve’s giant.

2017 was the best year for Valve so far: Steam's share of the market grew to $4.3B, not counting in-apps and DLCs.

If you're at #GDC18, come to my talk at 4PM at 2005 West for even more numbers and graphs.

— Steam Spy (@Steam_Spy) March 22, 2018

At first glance, this data seems to be nothing but good news for PC gaming, which has often chafed under reports that it’s a fraction the size of the console market. But there’s another issue: The profits overwhelmingly flow to a bare handful of titles. PUBG was always going to be a major heavyweight in 2017, but it hauled in $600M in revenue — dwarfing the earnings of the second-highest selling game, CS: Go, at $120 million. GTA V was in third place, at $83 million. Call of Duty: World War II came in fourth place, with $41 million in earnings from 840,000 copies sold. Of the Top 20 games, seven were released before 2017, including Civilization VI, The Witcher III: Wild Hunt, Rainbow Six Siege, Dark Souls III, and Rocket League. Rounding out the Top 20? Cuphead, with 1.3 million sales and revenue of $22 million.

Of the 21,000+ games available on Steam, fully half of the company’s revenue was delivered by the Top 100 titles. In other words, 0.5 percent of the titles available on Steam delivered 50 percent of the earnings. To some extent, that’s to be expected — some percentage of the games available on Steam are going to be old titles, and the original Half-Life isn’t exactly tearing up the sales charts these days. Then again, 7,672 games were released on Steam in 2017, which puts that 21,000 figure in perspective. That’s up from 565 in 2013, 1,772 in 2014, 2,964 in 2015, and 4,207 in 2016.

As Ryan noted earlier this year, PUBG was hauling in more players than multiple Steam games combined.
As Ryan noted earlier this year, PUBG was hauling in more players than multiple Steam games combined.

These figures suggest two things: First, that computer gaming is following the movie and music industries in squeezing out what used to be thought of as the ‘middle’ of the market. With skyrocketing dev costs and Steam becoming an endless morass of games, there’s less and less room between being big enough to catch the notice of a core group of gamers (or streamers, or influencers, or what-have-you) and vanishing into the abyss, never to be heard from again. In movies and music, these trends have played out differently — the movie industry may currently be suffering from an over-reliance on tentpole franchise films, for example — but the push and pull between the biggest box office draws or musicians and everybody else continues to favor the biggest, most-entrenched operators in all three genres.

I wouldn’t try to draw too many cross-genre parallels out of this, but there are some interesting similarities. Of course, one major difference between the gaming industry and movies or music is Steam itself. With 80 percent+ of the game distribution market, Steam has a lock on gaming that no other media outlet has ever enjoyed. Gamers, at least thus far, continue to view this as a good thing.

Continue reading

Google’s AutoML Creates Machine Learning Models Without Programming Experience

The gist of Cloud AutoML is that almost anyone can bring a catalog of images, import tags for the images, and create a functional machine learning model based on that.

Google’s Cloud TPU Matches Volta in Machine Learning at Much Lower Prices

Google and Nvidia both offer competitive machine learning products — but Google is beating Nvidia on costs, at least in certain tests.

Chinese Firm Claims It Scans Workers’ Brainwaves to Boost Earnings

A Chinese firm is claiming to have made over $300 million by monitoring the brainwaves of its 40,000+ employees at all times.

Intel Positions Xeon as Machine Learning Competitor in Inference Workloads

Intel is tossing its own hat into the machine learning ring, positioning Xeon as a potent inference workhorse.