Cloud is Not Cheap

Lyft is going public. Its S-1 filing reveals that "[Lyft] committed to spend an aggregate of at least $300 million between January 2019 and December 2021, with a minimum amount of $80 million in each of the three years, on AWS services." If its usage of Amazon's cloud doesn't hit or exceed that $300 million threshold, Lyft will have to pay the difference. 

It echoes Snap's (Snapchat parent company) S-1 filing in 2017: "[Snap] have committed to spend $2 billion with Google Cloud over the next five years and have built our software and computer systems to use computing, storage capabilities, bandwidth, and other services provided by Google, some of which do not have an alternative in the market."

Cloud is great for a startup to grow its business when it lacks of capital and skills to build the IT infrastructure. But once they grow up, cloud may not make sense any more. Some large-scale web companies have actually moved from cloud to their own data centers. Dropbox shared their story in the S-1 filing. Over two years, Dropbox shaved $74.6 million off its operational expenses primarily because of the move.

It is not only about cost but also about innovations. If compute is part of your core competency, building own infrastructure is worth of heavy investment as every business has its unique compute challenges. With AWS, Dropbox struggled to handle the explosion of data it was dealing with. While building in-house data centers, Dropbox had designed and built its own storage systems. Codenamed Diskotech, the system hosts 500 petabytes of data with 3-5x better performance at tail latency.

It is time for a reality check of your cloud strategy. Checkout my previous post on private cloud strategy.