GLG News by Everett Thompson

Implications: While the US macro economic fundamentals are weakening, the market for data centers remains strong.  Further, the restriction in access to finance is limiting supply growth significantly.  The benefits of this environment will be reaped by the established players.

Analysis: I see few signs of a slowdown in the demand for data center space and ancillary equipment. As the article cites, in the most recent quarter, three of the leading publicly listed data center specialists raised revenue guidance: providers Digital Realty Trust (DLR), Equinix(NMS:EQIX) (EQIX) and Switch and Data (SDXC(NAS:SDXC)).  And these providers are continuing to invest as aggressively as their balance sheets allow.  Related equipment providers like Broadcom(NMS:BRCM) also continue to report sustained revenue growth.  The sector definitely has it's laggards, with Savvis(NAS:SVVS) the easiest target. But the challenges of these companies are complex and not necessarily due to slowing growth.  While it is obvious that the data center market will be influenced by a downturn in the economy, to what extent will that influence be?  There is no reason to expect a cliff or precipitous decline given current evidence or the strong fundamentals of this industry.

Implications: Equinix momentum continues, backed by a strong balance sheet and aggressive growth plans.  While pricing growth in the Americas has fallen to inflation levels, the well diversified company will continue to see strong revenue growth.  G&A growth is their one weakness.

Analysis: There were no real surprises from the Equinix release - Equinix is by far the strongest player in the data center space today, with few true "form" competitors.  Similar to many multinational firms, Equinix is benefiting from it's now diversivied business, racking up $1 MM in quarterly revenue from currency fluctions alone.  Per cabinet revenue grew materially on the back of the currency fluctions, though US cabinet revenue grew modestly y-o-y to $1,693 from $1,681.  Another interesting statistic was their new bookings, now at 85% from existing customers.  This high figure is typically a positive; however, lack of new client growth may have contributed to an uptick.  The company is investing aggressively - you can find a schedule via their website - in the face of continued strong demand.  Slowing economy or not, I see little on the horizon that will blunt their continued growth.  Now if they could just get a hangle on the G&A growth....

Implications: 1) Despite economic uncertainty, data center demand is strong. 2) The data center economy is smarter and stronger than ever.

Analysis: In the data center economy, with basic real estate on the left of the continuum and fully-outsourced, managed hosting on the right, I see continued demand all along the spectrum, including shell space, wholesale data centers and retail colo.  And with 45% of companies with more than $100MM in revenue planning to build/renovate in 2008, versus only 22% in 2007, it would appear I am not alone.  Combine that with the continued emergence of Internet businesses and it would appear that a near-term slowdown in activity is unlikely. 

And in the event of a slowdown?  Well, Equinix has what, four million sq-ft?  They have been a very cautious investor, and are anything but overextended.  A flight to quality would only benefit their business in the long run.

Savvis may be vulnerable.  With four new facilities, and their ongoing re-positioning as an integrated, facilities-based, managed service provider, their ability to execute is going to be tested.

Implications: 1) Telecom and data center assets are highly specific infrastruture. 2) Both require significant investment and lead times to deploy. 3) For data center forecasts, the starting point is real estate.

Analysis: This article in some ways reminds me of Bernie Ebbers and his public comments on Internet growth back in the pre-bubble days.  It would be great if there were a way to accurately predict demand.

Specific to data centers, the reality is that while tracking public build announcements is relatively simple, monitoring the demand side of the equation is much more difficult.  How long will it take the market to absorb the 36MW DuPont facility in Chicago?

What I see in the market is continued demand for new product driven by the technology refresh cycle.  DuPont's facility has the latest, high-density space, making it some of the only space on the market.  They believe it will sell out before the end of 2008.  From where I sit, I'm inclined to agree.

Implications: - Virtualization, Grid / Utility Computing are maturing rapidly. - New Offerings and technologies are multiplying. - Revenue is growing rapidly. - Offerings from HP (Opsware), Savvis, Navisite are the evidence.

Analysis: I used to believe that processing power that could be accessed like power from a utility was some far off dream of fanatical technologists.  I now realize that virutualization, SaaS and grid computing are all elements of this vision and that these technologies are contibuting real revenues.

3Tera's and GridLayer's vision - see article - of rapid deployment and instant distribution is mature and in practice today at HP (Opsware), Savvis and Navisite.  These firms are mature users of vituatlization and blade technology which allows them to consolidate traditional hosting services 10:1. 

It is a good time to be in the hosting business and this virtuous cycle of consolidating infrastructure, costs and increase ease of use.

Implications: - Power usage in data centers is a huge issue. - Data center users are highlighting power costs in TCO. - IBM will gain traction owing to power and space advantages.

Analysis: Users of outsourced data center infrastructure have historically cited networking as the number one cost followed by power and lastly space.  Owing to blade technology and deflation in bandwidth pricing, the cost of power and cooling is rapidly moving up the ranks.

IBM has been aggressively advertising what appear to be order of magnitude advantages in power and space consolidation via the use of their linux-based mainframes, reducing energy consumption by as much as 98% (see article). 

With the DOE considering legislating energy efficiency in the data center, I'd say it's beginning to look a lot like Christmas at Big Blue.

Implications: - The supply of for data center real estate is growing - Investors should key on the competitive response of EQIX

Analysis: The supply of data center services has been very tight for the last several years. Demand, at last, is starting to catch up with supply - material pricing increases are behind us.

Astute investors will place greater focus on the competition and focus in on build plans, "proactive" churn figures, and solicit detail on how Equinix will lock in peak pricing agreements.

Implications: 1. Savvis is shoring up its finances to free up cash for growth. 2. Expect more of same from former Equinix CFO, Phil Koen. 3. Capital intensive nature of data center business may constrain overall growth

Analysis: I would never bet against Phil Koen - he is an astute and committed businessman.  And he has success to draw from, using the experience he gained while CFO, and subsequently President Asia-Pacific, at Equinix.  Even today, Equinix Asia Pacific is arguably the more profitable division of the company.

So, Phil has brought on board some of his previous colleagues, focused on cutting costs and shoring up the balance sheet, reducing debt and interest expenses, and you can bet he will continue to explore higher-level, value added services. 

However, Savvis is a facilities-based service provider, which means without facilities, they will be revenue constrained. Will they be able to fully leverage their relationship with DRT?  Savvis will likely need significant sources of financing, both internal and third-party, if they are to fully develop their revenue potential.

Implications: 1. Europe has never been as dynamic as the US market. 2. Equinix paid a pretty penny to get into the market - 7% cap before upping the ante. 3. There is little additional upside from cross-border sales collaboration.

Analysis: This is the second acquisition by Equinix of a foreign-based service provider, the first being Pihana Pacific at the end of 2002.  With little overlap in footprint, and nearly identical service strategy, on the surface, this seems like a good move.

However, my suspicion is that the acquisition will drag down the 35% y-o-y growth previously attained by Equinix.  And there is little benefit to a global footprint to the majority of Equinix clients.  Analysts would be advised to judge the European operations on a standalone basis with minor benefit from US MNCs.

Implications: 1. Many data center users are not used to planning 18 months in advance. 2. DLR pre-built space is a compelling product in a supply contrained market. 3. DLR is aggressively building with some 250k sq-ft raised floor and another 250k of shell space planned.

Analysis:

DLR is taking away one of the biggest issues for data center users, and that is one of schedule.  In this supply constrained market, the only thing preventing companies like Equinix, Savvis, and DLR from selling more is supply.  And supply that is available in six months or less.

With their pre-built space, DLR is solving the timing and planning issue for many companies.  This will result in continued strong growth and upward increases in pricing for this technical real estate provider.

Implications: Equinix continues to lead the market for premium datacenter space.  Investments reflect the strong demand for its services and strategic shift to property owner instead of property manager (lessee).

Analysis:

• Regulatory changes (SOX / HIPPA) creating demand for upgrades / replacements.
• Industry maturity and demand soaking up preexisting supply.
• Telecom penetration reaching the tipping point and increasing demand on existing service providers.
• Technological advances rendering existing stock of technical real estate obsolete.

Equinix is capitalizing on this market opportunity via strong development, though it's losing the multi-carrier features that made its brand.  Equinix, always a conservative investor, is leveraging its strong balance sheet for campus style development.  These moves place Equinix in an unequalled position in the retail datacenter space. 

Implications: TIVO should be acquired by EarthLink.  Combining TIVO with a CDN, an eyeball network like Earthlink, or other media related ISP would create a category killer for content delivery.

Analysis: I am often asked, "What does the future of content delivery and in-home entertainment look like?"

The answer is a big, in-home hard drive with software that manages the automated downloading and display of content with a very easy-to-use interface.  Guess what?  I just described TIVO.

Does a consumer really care which pipe the latest programming came down?  With TIVO plugged into the Internet, and some better network functionality to serve content to PCs, there is nothing more to do.  The video that TIVO records today may come from cable or a satellite, but it could just as easily come from the Internet.  Shall we call it "virtual" cable?

Oh, but what about bandwidth limitations?  With technology like that from BitTorrent installed on every TIVO shipped, my suspicion is those last-mile limitations would disappear very quickly.

With TIVO, or ReplayTV for that matter, any content network would be able to claim the first end-to-end delivery.

Potential suitors are numerous - they are the who's who of the Internet: Apple, Akamai, Internap, Limelight, Yahoo!, Google, Microsoft, etc..

Implications: The acquisition of Netli will accelerate Akamai's penetration of realtime application markets, a strategy that is well underway.

Analysis: This acquisition is meaningful for the following reasons:

- It will bolster Akamai's IP in network optimization, allowing it to defend its market.

- It will support high gross margins via higher value added network services (versus commodity download services of Limelight).

- It will bring on several marquee B2B customers who use Netli for pieces of their supplier facing websites.

Implications: Google is attempting to grow beyond its core advertising business.

Google will make some interesting moves, but like many businesses that stray too far from their core, their legacy business will remain the most profitable.

Good article and comments, but there's a lot of hype.

Analysis: - Google buys/builds datacenters for many reasons in locations with cheap power. Their overwhelming need today is power not bandwidth.  The bandwidth purchases are for their next stage of development.

- Google's strategy is similar to Yahoo!BB's of Japan - nothing new. They are likely building a network to facilitate direct access to the end users, both businesses and eyeballs, thus owning the user from end-to-end.

- Google has a clear interest in software as a service (Salesforce acquisition?). 

- BitTorrent does not account for 50% of traffic - they DO use bandwidth from otherwise unused resources at the edge of the network and via peering connections which is not accounted for in overall Internet usage statistics.