Hyperscalers: Cloud TAM

This is a snippet of the transcript, sign up to read more.

With that in mind, my initial question for you is, to what extent, in your role, over the last five years, how this has changed over time and how you are thinking about it right now, when you are moving workloads to the cloud, to a third party hyperscaler, do you distinguish between base and burst as an influence for whether you should maintain it on-prem or push it to the cloud?

But as you move the permanent workload into your full production environment in the cloud, then you're going to need that capacity. There will be some spikes, like large order volumes, month-end activity, or when you're running an overnight supply chain process where all your orders have to flow to supply chain systems. But you have to contract for that full capacity because the pay-as-you-go premium that cloud providers charge is just too high.

This is a snippet of the transcript, sign up to read more.

Instead of simply moving the unpredictable, financially challenging workloads, the bursty ones, to the cloud where they can handle them due to their extensive fixed cost capacity, why would you want to move everything, including the predictable base workloads, to the cloud? What is the business case for this approach?

I believe that innovation is key, particularly in the data space where there's so much happening. To truly leverage this innovation, you need to move to the cloud. Traditional descriptive analytics are rarely done on-premise these days, unless you're a global bank or a government with a lot of legacy systems. For such entities, it's not easy to transition to the cloud, so they still run a lot of on-premise operations. However, in the last two or three companies I've been part of, we've moved to the cloud to leverage the capabilities of platforms like Snowflake, Databricks, or Azure Synapse. These applications allow for data ingestion, data pipelining, and analytics. For instance, generative AI and OpenAI are difficult to build on your own. It's easier to leverage the Microsoft ecosystem, especially if you already have a large footprint.

This is a snippet of the transcript, sign up to read more.

Could you share, not necessarily specific to your company, what percentage of companies like yours have their workloads on the cloud? And where do you see that percentage in five years? This would be helpful to know.

Indeed, the usage varies based on the industry. At a CPG, a manufacturing company, we have traditionally used more of what we call COTS or SaaS products due to our heavy SaaS footprint. Therefore, our cloud mix between SaaS, PaaS, and IaaS is around 60% to 70%. The remaining 20% is on the edge, meaning it's on-premises but not in the data center. Then, there's about 10% to 15% in some sort of on-premises colocation. However, if you're a global bank, a significant part of your infrastructure would still be on-premises, and some would never move. Different banks are at different levels of maturity, like Wells Fargo, which has made a big migration to the cloud, and JP Morgan. Different workloads also have different requirements. For example, if they run their HRMS system, they're going to run it on the cloud.

This is a snippet of the transcript, sign up to read more.