Around the beginning of the 21st century, companies such as Amazon and Google grew massively. In order to consolidate their growth, they used clustering techniques: not only did they have better servers, but they also relied on many more of them working together.
In a scenario such as this, deciding how to store your data is key. If you take an Entity and spread its information throughout multiple servers, in multiple nodes of a cluster, the effort needed to control transactions is high. The same thing applies if you want to fetch an Entity. So if you can design your Entity in a way that is persisted in the node of a cluster, it makes things much easier. That's one of the reasons why Aggregate Design is so important.
If you want to know more about the history of Aggregate Design outside of Domain-Driven Design, take a look at NoSQL Distilled: A Brief Guide to the Emerging World of Polyglot Persistence.