Does AWS know itself yet?

I’m looking forward to new AWS CEO Adam Selipsky’s re:Invent keynote on November 30th, 2021. It’s a chance to see how well Adam has settled into his new role and the first big public appearance where announcements are involved. This means it is also a great chance to test a theory I’ve been forming.

AWS is the world’s leading utility providing operational excellence as a service to enterprise customers.

If the test of a theory is its ability to predict, how can we test this? What should we expect to see from Adam’s keynote?

Let’s break this theory apart into its core components and address each in turn. First, AWS is a utility. Second, AWS provides operational excellence as a service. Finally, AWS serves enterprise customers.

AWS is a utility

AWS carries many characteristics of a public utility. With 25 regions as of November 29th, 2021, and new regions costing more than three billion USD each, AWS is “a natural monopoly … due to large demands for capital and other barriers to entry.” AWS is also “affected by the public interest and of great importance to health, welfare, or the economy.” AWS also has a “need for ‘unused capacity.’”

Looking beyond the legal characteristics, AWS is more akin to a mobile network operator (MNO) than it is to an over-the-top (OTT) service provider. AWS offers native services but admits that third-party offerings like Vercel and Snowflake offer a more robust feature set and better user experience.

This means that AWS should invest the most capital and effort — and ostensibly be the most successful — in its core service offerings. Historically, this has been the case, with innovations like Firecracker MicroVM and the Arm-based AWS Graviton processor driving huge leaps forward in capabilities while minimizing resource usage.

For Adam’s keynote, the tests here are which services receive the most keynote airtime and which innovations receive the warmest welcome from customers. The first is measured in time; the second can be measured by the number and quality of reactions on social media during and immediately after the keynote.

AWS provides operational excellence as a service

Operational excellence stands out as the number one reason to run your workloads on AWS. Unless you are running a competitor service, you cannot finance your way to the level of expertise and experience that AWS has across all operations domains from energy sourcing to patching and maintenance.

AWS has entire service offerings devoted to operating open-source solutions, such as AWS Managed Services for Kafka, Amazon Elastic Map Reduce (EMR), and Amazon OpenSearch Service. Customers like these services because they can run their existing applications with minimal modifications while pushing the toil, or undifferentiated heavy lifting, onto AWS.

For Adam’s keynote, the tests here are how much time is spent talking about operational experience and any announcements of open-source solutions-as-a-service (OSSaaS?).

AWS serves enterprise customers

This may be a somewhat controversial claim, at least among AWS employees, but AWS is an offering for enterprises and specialized platforms. Everything from volume-based discount pricing to support makes more sense at the enterprise scale. Small and medium businesses are better served by offerings like Netlify and Vercel that offer usable abstractions on top of AWS.

Yes, AWS has its AWS Activate program for startups and in some cases that makes sense, but those cases are the “specialized platforms” mentioned previously. They’re also few and far between.

This doesn’t mean that you can’t build successfully on AWS if you are not an enterprise. What it does mean is that you won’t be the focus for their customer development, or at least even if you have a voice in the room, you won’t be swaying feature development. That goes to the enterprises, and that’s an entirely reasonable business decision given how much money they spend on AWS.

For Adam’s keynote, the tests here are what percentage of reference customers are enterprises and how many times are the words governance and ‘VPC’ used? That last one might seem pejorative, but the VPC and the jungle of complexity it brings is a harbinger of enterprise requirements.

Does AWS know itself yet?

If the theory holds, AWS has a fantastic future ahead of itself, so long as it recognizes what it is and leans into its strengths. Despite its infamous OP1 planning cycles, AWS has not been successful at working in concert as a unified organization to achieve a single, clear mission. In fact, that sort of focused effort is antithetical to the storied “two-pizza rule.” While that rule was fundamental to Amazon’s rapid growth in the 1990s, it’s not an approach that carries a utility to long-term success.

Adam’s task is to stake out a future for AWS and demonstrate that he can lead AWS to realize that vision. Hopefully, tomorrow’s keynote shows us that he recognizes AWS’ strengths, position, and — most importantly — customers.