Data Center The Modular Approach

This building we have in Singapore, it's going to support between 20 and 30 megawatts of IT load. It would be crazy to build all that out at one time, even if I was one of the biggest users in the world. And the reality is a data center is a long-lived asset. typically 20, 25 years of useful life in a data center. So building something like that all at once carries a huge amount of financial risk to make those investment commitments. Then you have some technology and product risks. We've seen densities increase. We've seen a big revolution from single cord to dual cord, back to single cord, back to dual cord. Distribution architectures are changing. So the idea that you're going to build one thing and it's going to serve your needs over this entire period is a really difficult proposition to make. So by building in smaller, more modular chunks, you're reducing your risk on your finance side, you're reducing your product and execution risk, and you're generally doing a better job at managing your business. And so that, that's really the key driver for modularization is being able to tune what you do to your business requirements. The dangerous part is if you overbuild, you're wasting money. So we have a lot of experience where clients have come and they've been in the, in the throes of a particular vendor. We need 500 watts per foot, we need certain capacities, we need a maximum amount. If you overbuild that data center, the main thing you've done is you've parked capital and that capital is stranded. And that's not a good situation for companies in modern times where everybody's trying to squeeze every last dollar and be as efficient as they can. So when we talk to clients at the CFO and the board level about building, we're usually talking about risk and finance. Technology is frequently off the table. And the idea about a modular approach is it allows the boards and the business people to deliver in an incremental basis. I'll give you an example. In 2008, global financial crisis hits. Digital, at the time, we had something like 40 projects in process. In one particular project we had, it was about a 13,000 square meter data center building with 6 data centers programmed in it. We had built 2, leased 2, and we had 1 under construction. We were able to finish that project and spend maybe $2, $3 million more to finish it out and then halt construction on the rest. We had competitors and end users in that marketplace who had $100 or $200 million projects. that they couldn't stop. So the ability to pull back on a modular build is probably one of the most important things that CFOs and finance teams look for. On the IT side, it allows you to tune demand because, as you know, especially if you're in the enterprise space, acquisitions, divestitures, new products and technology changes, you don't want to make a decision in 2010 that locks you into an architecture that in 2012 your new initiative can't be supported with, right? And so that's one of the drivers for modular.