A story in Tuesday's Wall Street Journal (subscription) examines Digital Realty Trust, the largest landlord in technology real estate, and wonders aloud about future demand for data center space. "Digital Realty Trust Inc. is going where few real-estate investment trusts have dared to go since the technology-stock bubble burst in late 2000," the Journal's Michael Corkery writes.
While noting the case for the "bulls" buying Digital Realty shares, the article features comments from several analysts who worry about a repeat of the tech bust, and the potential impact upon DRT's operations. "There is a real risk that some of the company's tenants may not be in business within the next few years," Citigroup Inc. analyst Jonathan Litt wrote in a report late last year. There is also concern that the shift to blade servers and smaller equipment will translate into smaller lease requirements for enterprise companies.
The story may have had a chilling effect on other stocks in the sector. On Thursday Piper Jaffray downgraded Akamai (from Outperform to Market Perform), and Matrix Research lowered its recommendation on Infocrossing (from Buy to Hold).
Tenant quality is a huge issue for data center operators and landlords. But there are significant difference between the demand supporting the current boom in data center industry, and the dot-com hype that fueled the rampant overbuilding in 1999-2001.
There are several key demand drivers supporting the current market. Let's look at three of these:
The Enterprise Market: As more companies adopt blade servers, data center consolidation is becoming a more meaningful trend. For many companies, this may mean smaller space requirements. But where is that space located - in company-owned facilities, or provider-owned data centers? This balance is shifting as well, with more companies mothballing in-house operations in favor of off-site third-party data centers. Meanwhile, a wave of legislation requiring corporate data retention will translate into growing need for storage. All those e-mails and company documents will be digital eventually. Enterprise demand never materializes quite as fast as service providers hope and expect. But it's not going away either, and in most scenarios will support growth over the long term.
The Web Hosting Industry: The dot-com bubble saw hundreds of venture-funded start-ups lease data center space to accommodate demand that never arrived. The growth in the web hosting industry in the past three years has been very different. The hosting companies that thrived in the post-bust industry were the ones that focused on organic growth and profitability. They didn't buy data center space until they actually needed it, and then bought distressed properties cheap. As a result, most of the expansion space has become occupied, and their facilities costs are not onerous. Shared hosting and managed hosting have thrived, and in many cases justified providers' capital investments. The dedicated hosting sector is a different animal. Customer demand for discount dedicated servers has been enormous, but it's also the most capital-intensive form of hosting, requiring an up-front investment in hardware that is recovered over time. This poses the risk that large providers could face funding issues, but in the current market struggling firms would be acquired, with minimal impact on space requirements.
Everything Over IP (Video, Audio, Phone): Cheap storage, higher broadband adoption and improved business models have paved the way for a massive shift in how media is delivered. Movies, music, video games and phone traffic are all moving to IP networks and on-demand delivery. As we've previously noted, this will create demand for data center space. Media files are much larger than the text and image files that comprise most of existing web traffic. Even with blade servers and a gradual transition to media-over-IP, a collapse in demand for data center space is unlikely, and the potential upside is significant.
Will these trends treat all providers and landlords equally well? No. There will be winners and losers. Digital Realty's business model is focused on facilties with superior infrastructure and existing tenants (preferably enterprise tenants), as shown by its 93 percent occupancy rate. That's not an accident, and its caution and focus will serve it well going forward.
A caveat: There are a number of signs that the enthusiasm for "Web 2.0" technologies may be getting ahead of itself. The VC investments in Internet start-ups are starting to look more speculative, and it wouldn't shock me to a see a "correction" in sentiment soon. A reversal in Wall Street sentiment on technology could easily batter data center stocks, which can create challenges in running and funding these businesses. But that won't change the fact that the demand outlook and tenant quality are fundamentally improved.