The Data Center Grant Controversy

There’s been recent news about the data center that had been proposed for Page County.  It seems that the Page County Board of Supervisors have voted to discontinue their sponsorship of this proposal, setting the stage for a new round of sturm and angst about a $300,000 grant that was provided by the Commonwealth of Virginia to support its development.  At the heart of the current discussion lies the question, “Who will pay it back?”

Now, I’ve posted fairly extensively on this topic – 10 posts before this one, with the last in December 2010 (check the Page County Data center label at the end of this post if you’d like to read them), and in one of them there is even a link where the local radio show interviewed me on the topic.  I’d like to be able to say that my posts focused mainly on the feasibility of bringing that kind of industry and development to Page County.  I thought it was a folly then and still do, but it might be easy for the champion of the proposal and past county supervisors to think that my posts were personal and directed at them.

You see, in my mind, I can envision this brainstorming session way back when, where some supervisors, some economic development board members, and the new businessman in town got together and decided that tough times called for a big bet – and in their hubris, or naivete, or both, they decided that bringing cutting edge industry, like a data center, to Page County would bring a lot of good paying jobs that would save the local economy, and they talked themselves into it. 

It didn’t work out like they thought it would; there were simply too many factors at work against the concept.

I’ll leave other interested stakeholders – who have made valid observations about the unraveling of this concept – to dissect and distill the lessons that Page County needs to learn from this pursuit.  There are many, and we’ve only just begun that exploration.

Meanwhile, the controversy over the grant is percolating along, having made the front page of the local paper for two or three weeks running.  It seems there was a performance clause, where if the results – new jobs – were not delivered, the grant would have to be repaid.  The supervisors’ activities have accelerated the state’s action on this, and there has been a call for repayment.  The supervisors have passed that bill right along to the project’s champion, just like the mayor on that old show “Carter Country” might have handed something unpleasant along to the sheriff:  “Handle it, Roy!”

If there is a specific performance clause in this grant, someone’s got to pay, I suppose.  I know how important capital is to small firms, and it is likely that this particular infusion into the company at hand created a job or two, just not the ones at the data center.  More likely, they were administrative positions that help run the business and supported marketing activities related to the development of the data center concept.  Also, we know that there were a number of feasibility studies, design drawings and meetings, and engineering studies that were done – all of which seem to be logical expenditures that would have taken place during this adventure.

Now, I’m not apologizing on behalf of anyone here, but the wheels got in motion from group think – the county board, the economic development authority at the time, and the business.  Maybe the burden of repaying the grant is something that all of them are responsible for, not just the entrepreneur. 

That’s just my two cents this morning.  I know that many folks won’t agree with me on this one. 

Investment Questions about Attracting Data Centers to Page County

Some alert readers on the new Page County blog site found a recent article out of Roanoke about the negotiations that surrounded the announcement that Microsoft would be investing $500 million in a data center in Mecklenburg, Virginia. It turns out that the town of Christianburg, Virginia was also considered, but events conspired against that town’s bid. All of this is chronicled in an article titled, “Christianburg’s Miss Brings Pricey Lessons,” located at http://www.roanoke.com/news/nrv/wb/269714.


This article says that the figures aren’t official, but were gathered a review of emails and other communications regarding Christianburg’s pursuit of that deal suggests that “While the county grappled with finding money for basic services such as education, the localities offered to forgo $62 million to $117 million in taxes to attract the facility.” Typically, these deals involve upfront cash, discounts on land and utilities infrastructure, and rebates on tax payments.

What we do know about the eventual deal is that Microsoft will receive upfront payments of $2.1 million from the state, another $4.8 million from the tobacco related community revitalization money, state sales tax exemptions, $50,000 in state hiring and training benefits, $2 million in real estate, $3.95 million for local water and sewer connections, and a 20-year rebate on personal property taxes – the first three years of this benefit were valued at $12 million. All of this for a company whose June 2010 balance sheet shows current assets of $56 billion, including more than $5.5 billion in cash on hand.

Leaving aside for a minute questions about how the Page County data center will bring nearly 100 good paying job to the county, while the Microsoft operation, a 4th generation facility (Page County’s is a 3rd generation facility) will bring 50 jobs estimated at $50,000 annually, imagine how these benefits would impact established businesses in the county that have to pay local taxes. According to this article, the state expects that it would break even on the project by 2021 – but due to the personal property tax waiver, the county would still be paying for these jobs on into the 2030 decade!

Then there are other benefits at the Mecklenburg location: the high-wattage electricity is in place, there is premium broadband already built, other utilities (water and sewer) are ready, and the site is located next to a US highway. If these are the features of a location that Microsoft values, it leaves the suggestion that Page County’s Project Clover site is acceptable open to question. We currently don’t have these assets to offer; they will have to be built.

Christianburg’s story doesn’t end well – we know that Mecklenburg was eventually selected, for one thing. Apparently, during the finalization of the discussions, a small sink hole opened up at the selected site, and the site engineering that would be required to offset these risks were estimated to cost up to $30 million. Christianburg is in karst terrain, just as Page County is. The sink hole will make this site a challenge to market to other businesses.

Using these figures, it’s easy to say that turning down Fibrowatt was an excellent move on the part of our board of supervisors, since it would have required us to compete on these financial terms, with a broader impact to our quality of life, for fewer jobs that probably wouldn’t have paid as well.

But I would also like to find some kind of encouragement for Page County in its pursuit of this kind of industry and those much needed jobs, something I’ve heard referred to as “Page 2.0”. I have limited insight into whether we could make some of the required investments to attract outside business, and the marketing emphasis on the Project Clover land as a potential site seems to miss the mark.

The whole thing has to start with a plan, though – and for now, that is not something we have from current EDA leadership. Our board of supervisors would do well to consider that as the terms for some current EDA members are coming to an end – lengthy service is much appreciated, but at last, it’s simply time for a change in leadership.

The new Page County blog site is http://pagecountyblog.com/,  and of course, my past posts on this topic can be found by clicking the Page County Data Center, Fibrowatt, Project Clover, or Page County EDA labels at the end of this post.

Doubling Down on the Data Center?

A friend recently sent me some news about a prospective new data center in Mecklenburg County, Virginia. A few months ago, when I was researching data centers more extensively, I'd read that there were several counties - all similar economically to Page County, and not doing so well by most measures – who had jumped on this thought that data centers might be a good route for future development. My reading showed that a lot of initiative was focused on the former tobacco regions in Virginia.

The news my friend sent me is summarized below:

Governor McDonnell Announces Microsoft to Locate Major Data Center in Mecklenburg County

• Will be Largest Investment Project in History of Southern Virginia
• Company to Invest up to $499 million and Create 50 New Jobs
• Virginia Beats out North Carolina and Texas for Company’s Most Advanced Data Center

RICHMOND- Governor Bob McDonnell announced today that Microsoft Corp. will invest up to $499 million to locate their latest generation data center (Gen4) in Mecklenburg County, in what will be the largest economic investment in Southern Virginia history. Using modular technology and advanced cooling mechanics, the center will be Microsoft’s most advanced data center. The project will create 50 new jobs and Virginia successfully competed against North Carolina and Texas to secure this sought after facility.

From: http://www2.timesdispatch.com/news/2010/aug/27/datacentergater-ar-476140/

When I think about this news, the first thing I want to analyze is location – which is strategic in Mecklenburg County:  power coming off the Kerr Lake Dam (Wikipedia: John H. Kerr Dam currently produces over 426 million KWh), easy access to Interstate 85 and rail corridors, located between the two state capitals, and positioned between the financial sectors in Richmond and Charlotte. With rail cutting through there, a good communications infrastructure already exists.

The reference to modular data centers is also interesting. It’s an approach that makes the building as flexible as possible, allowing for easy changes to technology. In fact, as I read the press releases, the only potential negative I see is the use of modular data centers in a hurricane prone area. Then again, it's Microsoft, and they'll have that figured out.

This news once again brings to mind some rhetorical questions about Page County as the potential location for a large scale data center:

• What's our current communications bandwith at Project Clover?
• What is the power capacity, and where does the power come from?
• And mechanical capacity (water for cooling)?
• What major roads get us there?
• What data intensive industries are nearby?
• This $500-million investment in a Gen4 data center will create 50 jobs...how will the Page County Data Center, a Gen3 facility, create 96 jobs?

The headline in the Page News and Courier said something our data center project doubling down. Call me a skeptic, but until we see more information, basic math applies: two times zero is still zero.

Hawksbill Cabin on the Radio -

The guys on the Speak Out radio show invited me on to talk about the letter to the PNC editor I wrote last week.  They've already got the material up as an mp3 and podcast.  Here's a link: 


http://wraa.libsyn.com/index.php?post_id=625290

I'm on from 13:10 to 23:40 -
 
I will figure out how to get a link up on the page to the show in the next few days.

Page County Data Center Wrap Up

I edited the material in these posts down for a PNC letter to the editor, which appears as submitted below.  A lot has happened in Page County on these matters just this week, so this material is almost moot at this point. 
====

Dear Editor,


Recently I finished a data center feasibility study, so I thought I would share some of what I learned. I’ll focus on just two elements of my concerns in this letter: the competition, and how quickly data center technology changes.

Digital Realty Trust and Terremark are nearby in Ashburn and Culpeper. Digital has an occupancy cycle of just six months, and has land and infrastructure for expansion. Terremark has a secure data center “outside the blast zone” up and running, with 50,000 square feet of space ready to go. Without signed agreements to occupy the data center before it is built, can Page County really expect compete with them in this marketplace?

Also, large data centers are obsolete in five years. Microsoft, Google, and Intel usually just sell off to other companies ready to “grow into” their hand-me-down facilities, and build new ones.

Sun, HP and other technology vendors have started using sea freight containers that don’t even need a facility to house them anymore. The containers cost as little as $500,000 each, and ten boxes may do the job the entire $16.5 million Page County data center can do as designed. By the way, people cannot work in these containers, and they add few technology jobs - they are simply replaced when they wear out.

With the County stretched so thin in so many ways, it doesn’t seem like Page County is ready to swing for the fences in the data center industry. It is hard to imagine that such a high risk, high cost idea is the one that is going to save our bacon. There are many low risk, low cost goals and objectives in the EDA Strategic Plan that are being ignored because of the data center hoopla – why not pick a few of these and go for some solid base hits?

Best regards,
“Cabin Jim”

About this Page County Data Center...post 5, Final

Post 5 – Summing Up


You know, as a Page County taxpayer, I have only brought the topic up because I genuinely want to see us prosper and grow – and I honestly want for our Board of Supervisors to be successful. I hope these posts provide them with some additional data they can use as they consider next steps for the County and EDA.

I mentioned some preliminary concerns in my first post, when I said that it is hard to believe that anyone is entertaining the concept of building one of these high risk, high cost facilities in Page County. The industry just changes so fast, and technology has changed this much since the idea first came up…this just seems to be a case where the due diligence has to be thorough and accurate.

Why would a lender give EDA, Premier Technical Services, or anyone else a $20+ million construction loan for a data center facility, especially if the borrower has never built, operated, or perhaps even worked in one?

- and especially to build one on speculation, in a location where the infrastructure has not yet even been extended to support this kind of technical facility.

I am not sure where the financing for such a facility would come from today’s financial market – unless there are signed contracts with tenants for the space. Especially when well financed – self financed – professionals like Digital and Terremark are already swarming the market with competing, up-to-date modular solutions, already built or under construction, that won’t go obsolete as new generation technology comes on line.

In conclusion, it seems like the County is just stretched so thin in so many ways, I don’t think Page County is ready to swing for the fences in the data center industry. It is hard to imagine that such a high risk, high cost idea is the one that is going to save our bacon. There are many low risk, low cost goals and objectives in the EDA Strategic Plan that are being ignored because of the data center hoopla.

Why not pick a few of these and go for some solid base hits?

About this Page County Data Center...post 4

Post 4 – The Changing Technology


Moore’s law still applies (http://en.wikipedia.org/wiki/Moore's_law ), so capability, capacity, and infrastructure requirements continue to grow at a very fast pace. I just want to give two examples of change here – the equipment and the alternatives to data center facilities.

Large companies in this business that own their own facilities plan on obsolescence in three to five years, before selling it off to other companies ready to “grow into” their hand-me-down facilities. Is this a market that you, with your backgrounds as long-time Page County residents, can see us competing in – and succeeding in?

The racks and servers themselves populating data centers can also be obsolete in two to five years and their owners will have to replace them. However, this process can also mean an update to the facility at the same time the new equipment is brought in. If our data center were fully populated with 500 or so racks, when the recap cycle begins we’d need to recap those at a rate of one every other day. That is a lot of equipment to ship in and out of Page County – we can handle the truck traffic, but our roads would need to be improved to handle shipping sensitive, expensive equipment in and out of here.

Major technology vendors have even come up with a technology that lets users avoid the cost of a data center facility altogether: companies like Sun, HP, Microsoft, Google, etc., are advertising new modular systems that use sea freight containers and can be installed anywhere – they don’t even need a facility to house them anymore. These boxes, costing from $500,000 to $1.5 million each (an example is at http://nebula.nasa.gov/about/  - the data center is in the white container next to the antenna) are designed to allow users to avoid hard facility costs, like the $16.5 million Page County construction budget.

By the way, people cannot work in these containers. They are simply built with enough redundancy that no service is required…and they are replaced when they wear out.

Next Post – Summing Up

About this Page County Data Center...post 3

Post 3 – The Costs


I recall from past PNC articles that the PTS strategy for a data center included cost estimates of $16.5 million. Costs will surely have increased by now, maybe even to $25 million or more, but there is also significant infrastructure to be built.

For power, it’s reasonable to think that the facility will need upwards of five megawatts of electrical power – during the Fibrowatt discussions, we learned that the County’s entire power usage was around 25 megawatts, so this plant will require almost 20 percent of the County’s entire power load!

Most clients are sensitive to power availability and quality, so the building will probably need a second power source, especially if the clients will be government organizations (American Operations, the company noted in last week's PNC, is essentially a business development specialist for the federal market sector). Some of these clients are sensitive enough about real-time access to their data they might require a second, generator back-up in addition to the grid connection.

I learned that the power company will pay for the first connection, but the developer/owner/builder is usually responsible for the second grid connection, which can cost $1 million per megawatt; for five megawatts, Page County will need an additional $5 and $10 million just for the power infrastructure.

…and power is just the first of three critical needs for a data center – it will also need cooling, and communications, this last to be addressed by the Broadband Authority.  For our purposes, it's enough just to look at the power costs.

Next Post – The Changing Technology

Some Splainin on the Data Center Posts

I received some questions on why I left the roles and responsibilities here in a relatively unclear state.  That's because they need some clarification as far as the data center discussions go.

I understand that the EDA's role is courting and attracting business to Page County.  However, a lot of people are appearing in the PNC making statements about the data center - very few of them associated with PTS, the company who announced they wanted to build it in the first place.  The letters I am referrring to are the recent ones by Steve Dave McClure and Bill Shuler, where it's understood they have roles within the County (McClure calls himself the CTO, Shuler is vice chair of the EDA and BBA), yet here they are championing a data center...it gets very confusing to me about whether this is a public sector effort or private sector initiative - it belongs firmly in the private sector... 

Since so many of these people making public statements aren't being clear on their roles, I have left a lot of these details in a not specific state.

I just don't have the visibility into these roles to assign who's responsible and who isn't. It seems like one big "them" to me, and that's my point.  Without clear lines of responsibility, without a game plan, and without due diligence, it's ripe for SNAFUs.  That's why I have posted this stuff the way I did.

Cabin Jim

About this Page County Data Center...post 2

Post 2 – The Competition


As we think about who we would compete with if Page County were to have a data center, let’s focus on just two nearby companies with data center facilities.  I've heard that there may be existing buildings in industrial parks down in Harrisonburg along I-81 that may be suitable for data centers, but I'm not sure about them.  We can get an idea of what to expect, not only in building a data center here, but in finding customers for it, from Digital Realty Trust and Terremark, who operate data center facilities in the Ashburn and Culpeper areas.

• Digital (NYSE Stock Symbol DRT; http://digitalrealtytrust.com/ ) has pioneered a modular development process that not only reduces the construction/occupancy cycle from as long as 3 years (which is what we can likely expect for Page County) to one that can be as short as six months. Digital is poised to expand their facilities in the Dulles area, already acquiring the needed land and infrastructure connections.

• Terremark (NASDAQ Symbol TRMK; http://www.terremark.com/technology-platform/nap-of-the-capital-region.aspx ) already has a secure data center – “outside the blast zone” – up and running, ready to accommodate clients requiring space as large as 50,000 square feet or more.

These are large, specialized organizations with quick and easy access to the financial resources needed to build these facilities. Can Page County really compete with them in this marketplace? You can’t bet the farm and get into this market speculatively – without signed agreements to occupy the space before it is built, especially if you are new to the market, like we are. 

And we don't even have the infrastructure trunks extended to the planned locations.  What are the schedule and cost implications of that issue?

Next Post – The Costs

About this Page County Data Center...

Post 1 – Intro


I’ve been thinking about the discussions that have been showing up in the Page News & Courier the last few weeks about the prospect of a data center in Page County. I submitted a letter to the PNC, but in case they don’t pick it up –

I guess the use of words like “bickering” and “playing on the internet” in last week’s paper have finally set me off, so I have a series of posts ready to put up over the next few days, based on some things I’ve learned being around the data center industry the last few years.

Data center technology and processes change very fast, so examining risks and understanding the deal thoroughly are critical. My research has convinced me that this is a prospect to approach with eyes wide open, frankly skeptical of buying into the concept as a panacea that will easily cure all economic problems.

To be honest, as well intentioned as those involved in this prospect may be, it is hard to believe that anyone is entertaining the concept of building one of these high risk, high cost facilities in Page County. My letter to PNC is brief so it fits their guidelines; in these posts, I’ll focus on three elements of my concerns: the competition, the costs, and how quickly the technology for data centers is changing.

I don’t think anybody is doing anything intentionally wrong, and I would never presume to suggest that. The industry just changes so fast. Technology has changed this much since the idea first came up, and this just seems to be a case where the due diligence has to be thorough and accurate.

It’s probably best to be prudent, don’t you think?

The County is just stretched so thin, in so many ways, I don’t think Page County is ready to swing for the fences in the data center industry. It is hard to imagine that such a high risk, high cost idea is the one that is going to save our bacon. There are many low risk, low cost goals and objectives in the EDA Strategic Plan that are being ignored because of the data center discussions. Why not pick a few of these and go for some solid base hits.

Next Post – The Competiton