Photograph by Mint Hindustan Times via Getty Images

Machine learning may help make sense of it all.

By David Meyer
August 11, 2016

Apple’s App Store is likely to contain twice as many iOS apps by the end of 2020 than it currently holds, according to research by the app marketing company Sensor Tower.

That will mean a catalog of just over five million apps for iPhones and iPads. Sensor Tower is basing this prediction on the trends in the app store’s growth since Apple aapl opened it in 2008.

The firm said games are by far the biggest category of new apps that get added each month, to the tune of almost 21,000 new apps per month, or around 43% of the total.

Get Data Sheet, Fortune’s technology newsletter.

After that, it’s education, entertainment, business and lifestyle, all of which see between 2,200 and 2,500 new apps added monthly. At the bottom of the chart, only 134 new weather apps appeared in May.

Not coincidentally, Sensor Tower provides tools that it claims will help developers get their apps in front of people’s eyeballs. When you’re dealing with this enormous volume of apps, only a few will be easily discoverable.

That problem will largely have to be solved by Apple itself. Sensor Tower reckons that Apple’s current push into machine learning could play a part—the company just this month bought a machine-learning startup called Turi for a reported $200 million.

For more on Apple, watch our video.

Machine learning essentially involves teaching machines how to learn from data, in order to make clever predictions. It currently manifests in the form of things like smart image recognition systems, spam filtering and quasi-artificial-intelligence virtual assistants.

“We’re already seeing so-called assistants making their way into core areas of iOS itself—most prominently in the new iOS 10 iMessage and Photos refresh—so it’s not unrealistic to imagine a near-future App Store assistant that learns from user behavior to deliver customized recommendations, intelligently surfacing new apps based on behavior patterns or even conversations with others,” Sensor Tower said.

SPONSORED FINANCIAL CONTENT

You May Like