Will Silicon Valley Run Out of Data Center Space?

1sockchuck writes "With capital scarce, data center developers are prioritizing projects in northern Virginia, where the Obama stimulus plan and federal shift to cloud computing are likely to boost data center demand from government agencies. This is forcing them to delay or scale back large projects in Santa Clara, setting the stage for a supply/demand imbalance in Silicon Valley, particularly for large space requirements. One potential mitigating factor: some currently occupied data center space could become available through the failure of venture-backed startups."

Comments

dream on

[Anonymous Coward]

There is a TON of unused data center space in the bay area, as dotcoms have folded up or moved.

Re:dream on

[cantcomplain]

You may be partially right--some of the facilities have space. But some that have "tons" of space are maxed out or near it for power and cooling. Some are not accepting new clients or anticipate turning away business in the next few months. I think it was 365 Main that turned me away and said the move to the cloud would be consuming their capacity. Some of the tenants at 200 Paul have space but some of it is pretty ghetto and limited by power. At least that's what I found in a cursory search dictated by a ceo that doesn't want to build out a server room (40kva w/10 tons hvac) at a new site. If you know cheap and usable co-lo space, all the sushi you can eat!

Re:dream on

[simontek2]

BDAW.com has some space, they are in the verizon building on 8th and Hope in Downtown LA. Peer1 usually has space all over the place.

Re:dream on

[Anonymous Coward]

um... not so sure bout this. I've spent months for a previous company looking for 10 racks of space. Closest we could get was in Oregon. Data space is largely limited by 3 things. Electricity, Electricity and Electricity. Can't run the systems, cool the systems, monitor the systems etc without Electricity. Most Failed dot comes at best clear up a half rack, most a virtual machine. Neither of which satisfies data center needs. Right now for a small 150-200Sq ft cage you are on a long assed waiting list. Have been for a long time.

Re:dream on

[mcrbids]

As a former 365 Main customer, I can say that our move to Herakles Data in Sacramento went smoothly, and what a difference! Great, reliable service, redundant EVERYTHING (unlike 365 Main which gave me A network feed and A power strip) with a bigger rack and more power, and at a BETTER PRICE.

Seriously, it's dramatically different, night and day in just about every way.

When 365 Main had their power woes a couple years back, it was all lawyer-speak about validating any kind of claim, because we *were* promised 99.99% uptime in the contract, but we'd have to sue to get it. They didn't even bother to respond to our legal letters and issued only cryptic, terse public notices.

When Herakles data had a core Cisco router go "half dead" a few months back, (blown enough to cause problems, but not cause backup routers to step in) causing a few hours of downtime, they sent detailed analysis of what happened, and then they SENT US THE FORMS if we wished to file a claim. It was even partially filled out! (we didn't file, just because of the trust and goodwill this fostered)

I was originally impressed with 365Main because it *is* technically a nice facility, but in terms of service quality and price, Herakles (1.5 hours drive away!) has them beat hands down.

Get a REAL data center!

No more startups

[Kohath]

No, it won't be a problem.

There will be very, very few new startups in the US. And many of the existing startups will shut down. There's just not much point in starting a business in the US any more.

- IPOs used to be plentiful, but that was before Sarbanes-Oxley made going public astronomically expensive.
- The government is sucking up most of the country's available capital [to buy votes] for stimulus and other government spending, leaving less available for business growth.
- The new stock option rules more-or-less preclude giving lower-level employees company stock so they share in the success of the company.
- Even for those that do see success, the tax rate will be 60-70% in a few years, so they won't be able to keep much of what they make. They won't be able invest the money in new startups because the taxes will take too much and there will be none left over.
- And don't forget that everyone knows businesses are villains and rich businessmen are hated. Why subject yourself to all that for such low after-tax gains?

See this article by Victor Davis Hanson.
See this article by T. J. Rodgers of Cypress Semiconductor.
See this article by Michael S. Malone.

It's not really the land of opportunity any more -- not unless you know just the right people in government or the environmental industrial complex to steer you an earmark. And even those will run out in a few years after all the money is spent and all the output from the country's slowly-declining future production is borrowed and spent.

There will be plenty of vacant data center space.

Re:No more startups

[YesIAmAScript]

Gloom and doom never gets old. We were definitely doomed in 2000 also. And 1994.

I don't think Sarbanes-Oxley has anything to do with the number of IPOs. IPOs are such a ridiculous money making scam that increasing the overhead a few percent isn't doing to dampen anything. I'd say the reduction in the number of IPOs is more to do with the stock market in the toilet than anything. Companies don't go public to succeed anymore, they go public to get a huge stock pop, and you won't get that right now.
You're right about the changes in stock options to low-level employees, but my understanding is there are no new rules, companies are voluntarily expensing stock options now. Which, to be honest, they should have done all along.
The tax rate will be 60%? Marginal tax rates aren't even at 30% for most people right now. Remember, it's Silicon Valley, your house costs $600,000 and the tax on the interest is deductible. I know people with 14 exemptions on their W-4, virtually all of which come from them paying $28,000 in interest on their house each year. That means if you're making $100,000, 28% of your income isn't subject or state or local taxes at all!

- And don't forget that everyone knows businesses are villains and rich businessmen are hated. Why subject yourself to all that for such low after-tax gains?

Good point, now what are you going to do with your business degree? I'd be lying if I said I was disappointed that your ilk might no longer be rushing to Silicon Valley to make their money for nothing but calling themselves executives and quoting tax regulations.

Re:No more startups

[Kohath]

Which other countries were a better place than the US to start a company in the 1950s? It was just after WWII and most countries were rebuilding.

Which countries are a better place than the US to start a business in 2009? And in 2011 when the taxes start kicking in? And what about the years after that? There are lots of countries that won't be capping their carbon emissions, for example.

Also, government as a percentage of GDP was about half as large in the 1950s. And the regulatory burden was much lower.

Re:I say DIG

[magarity]

Meh, there are plenty of sites available to build your own hardened data center.

Re:Rust Belt

[magarity]

meaning, you can find people with the necessary skills in other areas
 
What skills? A few people at the data center have mad skillz but most of them just need to know how to swap out tapes and failed hard drives and to not press the reset button until expressly told to do so. Most of the administration of the software, which is what needs the most training and skill, happens by people who are off site. Usually in India.

Why...

[QuietLagoon]

... do the data centers need to be located in Silicon Valley? I thought this thing called The Internet negated the need for geographical co-location.

Re:Why...

[Anonymous Coward]

* When you're a small company building something out, you need to have colo near your engineers, since they WILL have to visit it

* If you later become big the staffing isn't as big a problem (you could just pay a couple engineers to live near your data center in Bumfuck, NW) However, by then you have a lot of infrastructure in your current datacenter, so moving is hard/expensive too.

* Even if you're large you need to consider the risk of having a datacenter located far from most of your employees - if something REALLY bad happens will your skeleton staff near the co-lo be able to deal with it as quickly? How much is avoiding that delay worth to you?

* The people on this thread who are saying "just run a fat pipe to the middle of nowhere" have no idea how hard/expensive that is. Getting a truly redundant (i.e. different geographic paths) OC-192 to your shiny new data center in Montana is going to cost millions.

Finally, imagine if you're in the datacenter-building business and want to invest millions into building out a new facility. Would you want to put a low-cost facility in the middle of nowhere and hope you can find clients who are willing to host there (and risk having the project turn into a total bust)? Or would you rather enter a market like Silicon Valley where there are plenty of companies looking for space, with new ones showing up all the time?

The arguments in favor of building data centers off the beaten path aren't without merit, but they only work for the largest of networks were the question is "We already have 5 datacetners, where should we put #6?" (Think: Google, Microsoft, Federal Govt, etc...) For most organizations that just want to rent a few cages of space there are real reasons that they continue to cluster in a few tech-heavy parts of the country.