Summary: An interesting angle to the shift from asset-heavy pureplay managed hosting to a more asset-light position with reselling of clouds like Amazon, Azure and Google is the question of what to do with significant investments in data centre capacity that have already been made.
Details: In digging through filings, we can see that Rackspace has some significant wholesale data centre lease obligations with up to and above a decade remaining. With DuPont Fabros, Rackspace has an average term remaining of 8.6 years with estimated annual base rent of over $33m. With DuPont, Rackspace is in three buildings across two markets. With Digital Realty, Rackspace has average term remaining of 12.5 years with estimated annual rent of around $19m. It is in four Digital locations and Dallas and Slough are two of the most prominent.
Why is this important?: If Rackspace continues with its push of customers to third party clouds, and de-emphasizes the traditional managed hosting and private cloud business, it would become increasingly difficult to fill up this capacity based on prior projections. It is fair to presume that Rackspace made its data centre commitments with a very different business model in mind. It was not that long ago that Rackspace was competing in public cloud and wasn’t really thinking about managing third party clouds. Its need for infrastructure was steady and growing. But all of a sudden, that is changing. It is clear that ‘traditional’ services are still being sold and could potentially fill up this capacity but sales and support resources are being diverted to the third party cloud model and ‘traditional’ is going to get left short. The traditional business has also slowed for many years and less and less investment will only perpetuate the situation. And there is precedent. The likes of Logicworks and Datapipe might only have somewhere around 5-10% of revenue coming from managed third party cloud. But they are trending heavily in that direction and word on the street is that the bulk of new engagements are completely asset-light. And the deal sizes and profitability with managed services are attractive. Given the growth potential in this service type and Rackspace’s stated desire to get into this game in a more meaningful way, the same situation can quite easily happen to Rackspace. The majority of its pipeline could go asset-light very quickly. Returning to the original question, why is this important? Because Rackspace has already made such huge commitments it will have to figure out how to utilize this capacity that is not going to fill as easily and avoid wasting valuable CapEx.
What can Rackspace do?: The best course of action would, of course, be to keep selling bare metal and private cloud and fill up that space and power. And that will certainly keep happening. But to reiterate our comments above, there is every reason to believe this is going to get harder. Traditional is slowing and Rackspace is likely to continue trending in this direction. One option could be to consolidate and fold aging data centres into these wholesale locations. We noted how Rackspace recently had begun evaluating what to do with a legacy Grapevine data centre it had built back in the day. Given the situation, it makes complete sense to shut it down, move the customers to Digital or DuPont and sell the facility. So step one is to fold and consolidate legacy data centres (within constraints of data location rules and preferences), save the CapEx required to upgrade and then monetize. There might also be an opportunity to absorb legacy data centres from companies it acquires. Rackspace just acquired TriCore – its biggest acquisition to date and one that its relatively asset-heavy – and it is likely to be looking at other deals. It could also sub-lease to its channel and technology partners it works closely with. Sub-leasing on a larger scale, like Facebook had done in recent years, seems unlikely.
Angle: Naturally, Rackspace will have room to re-negotiate in light of its changing circumstances and likely has a decent alignment with power commitments contracted for and power that is available if growth demanded it. To be clear, it is not a dire situation for Rackspace. This just means that Rackspace will have to be creative as it tries to optimize in light of a new business model. And that is the main takeaway here. Providers are going to be moving resources from CapEx-heavy areas like hardware and data centre procurement and investing in more human-intensive managed services capabilities and software development. The ‘new world’ is more about services and capabilities and less about infrastructure. This is where providers will be competing. From the perspective of data centre operators, there is simply going to be less upside on raw compute and infrastructure from the likes of Rackspace as they move to managed third party cloud models. It is unsurprising that data centre operators, as a result, have focused more on wholesale and bringing in the big clouds as tenants. It is the best way to capture where compute infrastructure is moving.
The post The shift to asset-light; data centre implications for Rackspace and others appeared first on Structure Research | Cloud, Hosting & Data Centres.