Before a web3 DePIN offer can win against the $100B+ annual run-rate AWS, they should have a point-of-view regarding how AWS won as a starting point. Typically the rhetoric is “centralized Big Tech bad, decentralized good” but AWS was once the scrappy insurgent against incumbents.
So how did they win and are their lessons for web3?
Context: Who Were the Incumbents?
Currently, we see AWS, Google, and Microsoft as the main players in cloud infrastructure.
They are seen as unstoppable Big Tech that powers the Internet, with AWS leading the pack in cloud infrastructure services.
But that wasn't always the case.
At one point, AWS was the underdog that few believed in, even those who had an advantage over AWS.
So who were the incumbents that AWS had to overthrow?
The Storage and Compute Gorillas
At the time, giants like IBM, HP, EMC and Dell dominated the compute and storage landscape.
These vendors owned the CIO relationship for enterprise-scale infrastructure expenditures. When AWS launched, it had only two primitives — EC2 for compute and S3 for storage. While the offerings have expanded beyond that, those two — exactly in the wheelhouse of these established companies — gave it the traction to win.
So in theory, the storage and compute vendors — the ones who already had the technology and relationships — could have controlled the new era — but they didn’t.
Cisco - The Networking Choke Point
Cisco entered the compute and storage fray with a unified data center offering combining their core strength, network, with integrated compute and storage. They remain a leader in networking infrastructure, and from that position, could have also commoditized and integrated the storage and compute into a cloud offering.
Like the storage and compute competitors, they had massive reach through a formidable channel strategy, and relationships with decision-makers. But they, too, did not make a dent in cloud.
VMWare’s Missed Opportunity
The second adjacent giant was VMWare. Virtualization technology became the enabler for cloud computing. AWS used an open source hypervisor, Zen, for their offering; and while VMWare later tried to enter cloud, it failed.
VMWare could have leveraged its core enabling technology into a leadership position in cloud computing. But it didn’t. Why?
Equinix - The Data Center / Colo Giant
The third group of potential winners included data center providers like Equinix — companies that controlled the physical layer where the networking, compute and storage would come together to deliver the service.
Equinix already had relationships with all parts of the infrastructure stack; just as important, was the high density of network and network service providers. In many ways, they had the ultimate choke-point for accessing the infrastructure over the Internet.
They even had early visibility (likely) into AWS’s business since AWS needed colocation and data-centers on a lease-basis before building their own.
But the data centers didn’t end up beating AWS to the punch.
Google and Microsoft
Google already operated massive, horizontally-scaled infrastructure to power its search engine and services. Microsoft dominated enterprise software and owned the developer ecosystem that was the natural customer base for cloud solutions.
Although Google and Microsoft have since become two of the major cloud players, they were both initially beaten — and beaten decisively — by AWS.
Microsoft, especially, had the developer market which AWS appealed to initially; and a head-start on the channel as deals grew larger and which AWS built for themselves.
While both are strong contenders, AWS still had a lead; they missed the nascent opportunity despite having an apparent advantage.
Salesforce: The Original Cloud
Salesforce was founded earlier, and even entered the market via its acquisition of a Platform as a Service provider, Heroku, which did have much developer love in its early days.
In the same way Amazon's retail side provided insights and inspiration for AWS’s underlying services, Salesforce, as a dominant CRM and marketing application, offered valuable perspective on what cloud infrastructure can deliver to application developers.
Like Amazon, they already had to deal with scale and could feel their own pains of building customer-facing applications.
Yet despite its purchase of Heroku in 2010, Salesforce was unable to leverage the existing scale into a large cloud infrastructure business.
Framing the Question
While I began this post asking “How did AWS win?” the starting point was “Who were the incumbents?” These incumbents weren’t other start-ups; instead, they were significant, formidable technology companies with meaningful advantages in scale, go-to-market, and technology.
So how did AWS win?
I continue this later in part 2, but I welcome all answer (ideally your own ideas, hypothesis and creativity; but I can’t stop you from using an LLM, either).