I was having a deep conversation about puppies and socks with Box CEO Aaron Levie in the BoxWorks press room Tuesday afternoon when, as Levie is wont to do, he said something true.

“It’s about the platform. It’s always about the platform.”

See also: How Facebook Beat MySpace: From College Dorm To Platform

This came as a punchline to a string of vaguely off-color jokes about memes (Levie’s brain is a cloud-storage service for Millennial Internet arcana). Nonetheless, it’s serious business.

Enter The Platform

This year’s BoxWorks user conference was something of a coming-out party for the Box Platform, the distilled set of storage-and-collaboration services meant for developers to embed in their apps. Box first teased us with what would become the Platform in April at its BoxDev conference, and in the subsequent months, it was available only as a closed beta.

That changes in October, when a free developer version and a paid enterprise version of the Platform become generally available. When that happens, Box will go from a business that charges companies a fee per employee who uses its service to a business that charges developers a fee per user.

It takes Box out of the business of competing with Dropbox and Google Drive, in other words, and puts them in a category more like Amazon Web Services—or Salesforce, Stripe, and Twilio, three platform businesses Levie compared Box to in his keynote.

Box took its first steps towards this model in 2014, when it started charging developers for its content-storage service based on “API actions”—views, shares, etc. But this is a more decisive move towards profiting from apps developers build on top of Box.

Box Platform Enterprise Edition has a pricing structure loosely similar to the older Content API, with a free developer tier for building and testing, and a paid tier starting at $500 a month for 100 app users.

Jeetu Patel, who recently joined Box as its chief strategy officer and senior vice president of platform, told me that the per-user price will fall rapidly with volume and that “pricing will not be the reason people don’t do business with us.”

The risk for Box will be that developers will go straight to, say, Amazon or Microsoft, buy storage in bulk, and build apps on their own. Levie and Patel are betting that customers like Raymond James Financial—which demonstrated a feature for viewing mutual-fund prospectuses and other brokerage documents at BoxWorks—will find it cheaper to use Box.

Why Box’s Platform Could Make Dropbox Irrelevant

It’s no secret that Dropbox is floundering when it comes to its product direction—particularly with respect to developers. At the TechCrunch Disrupt conference earlier this month, Dropbox CEO Drew Houston barely talked about the company’s platform.

See also: Dropbox Must Avoid The Fate Of Your Fax Machine

Dropbox hasn’t held a big developer conference since its first (and so far only) DBX event in 2013. And the Dropbox platform advertises itself as primarily a way to access files in users’ Dropbox accounts.

Contrast that to Box, which sees Platform as a way to get past the idea of making users sign up for accounts in the first place. Instead, app users get the features of Box’s storage and collaboration services within whatever Box-enabled app they’re using.

Remember MySpace? It, too, touted the number of users it had. It took years for Facebook to surpass it in raw numbers. When Facebook rolled out its platform, however, it was game over. Facebook had an army of developers at its back, and MySpace’s belated effort to catch up in the platform race never caught up.

Social networks are very different beasts than enterprise services. If they share anything, it’s the power of platforms. Dropbox’s API lets developers access Dropbox. Box lets them build apps we can’t imagine. If Dropbox doesn’t get its developer story together, and soon, it will look like many a service that had a heyday of popularity—and then faded.

Photo by Owen Thomas for ReadWrite