Issue link: https://cp.revolio.com/i/428453
12 | THE DATA CENTER JOURNAL www.datacenterjournal.com year would reach $3.7 trillion. In 3Q12, it had pushed that number out to 2013, and even in 3Q14, it still estimated that world- wide IT spending has yet to break $3.7 tril- lion. In other words, overall IT spending is only now reaching levels expected just before the Great Recession, at least judging from data provided by one research firm. Granted, these numbers don't speak to data centers specifically, but they should track closely. In 2013, Gartner began breaking out data center system spending in its forecast releases: in 3Q13, it esti- mated 2012 spending at $140 billion and predicted $143 billion and $149 billion for 2013 and 2014, respectively. ese num- bers yielded forecast growth rates of 2.1% and 4.1%. In 3Q14, however, the research firm revised its estimates, keeping data center system spending about flat through 2014 at $140 billion and anticipating a 2.9% increase in 2015 to $144 billion. Don't overlook DeBt Spending numbers are funny: they may tend to indicate economic activity, but in a highly debt-ridden economy, they lose meaning. For instance, data on gross domestic product includes "government consumption expenditures and gross investment," according to the U.S. Bureau of Economic Analysis. But the U.S. GDP only grew by about $1.66 trillion between its bottom during the recession (2Q09) and 2Q14. Meanwhile, the U.S. federal debt grew over $6 trillion during the same time period; in other words, every penny of GDP growth owes to at least one penny of new government debt at the federal level alone (noting, of course, that the total GDP gain is much less than five years multi- plied by $1.66 trillion). With actual (i.e., closer to generally accepted accounting principles, or GAAP) federal debt being much higher—as high as $200 trillion by some estimates—default is inevitable, whether through nonpayment or inflation via money printing. e only question is when—and how it will affect the economy. On the corporate side, major technol- ogy companies such as Apple, Cisco and IBM have engaged in huge stock-repur- chase programs to the tune of billions of dollars. In a sense, stock buybacks reduce debt, but they also reduce a company's ability to invest in production (such as data centers). And in some cases, fueled by the Federal Reserve's zero-interest-rate policy, these repurchases are simply debt transfers. IBM, for instance, has repurchased $37.7 billion in stock since 2012, according to Zero Hedge and company filings. But it has also issued $33.6 billion in debt over the same period, roughly balancing its stock buybacks. e effect of this tactic is to prop up stock prices, but it does little to increase productive capacity. For some companies, this money poured into financial gim- mickry is money that could otherwise be spent on data centers. the future Judging simply from demand, an expectation that data center spending will increase in the next few years seems reasonable. "We think that data center spending will continue to accelerate in the coming years as more companies look to improve their operational efficiency, move to a more flexible opex model and tap into more services that are accessible through multi-tenant data centers," said Choksi. "Increasingly companies will decide that managing data centers requires a lot of work and a great deal of expertise—why not look to companies that do this for a living? Let the experts run your data center, and free up your resources and your people to focus on core elements of their business." Cloud data centers will no doubt benefit from this trend. A half-decade of forecasts predicting a growth upsurge that is ever near but nev- er here raises some questions, however. Of course, this year could always be the break- out year that leaves the effects of the reces- sion far behind. Unfortunately, though, the recession instigated little structural change, and talk of a new bubble in the technology sector has become rife. Although the de- tails differ from the months leading up to the dot-com crash, the investment frenzy for companies like Twitter—which has a market capitalization of about $25 billion despite huge losses quarter aer quarter— suggests that something is amiss. One fairly certain forecast is that the cloud has the upper hand in data center spending for the foreseeable future. Choksi notes, "I don't think in-house spending will increase or return. e trend of outsourc- ing to all models—retail, wholesale and cloud—will continue to grow. What we are seeing is that companies are looking for even more ways to outsource the different elements of their data center resources. At first it was just to distribute their platform for performance or to create disaster- recovery redundancy, then it was more about moving their core to a data center provider. Now we are seeing things like R&D environments and large-scale server farms being move to outsourced models with providers of all shapes and sizes." e question is then whether companies that provide data center outsourcing services will be able to drive greater spending to meet demand. ConClusions Assessing and predicting data center spending is a difficult task. Even distin- guished research firms like Gartner must regularly revise their predictions, perhaps to offset overoptimistic input from corpo- rate executives. But Gartner isn't alone, in that the much vaunted economic recovery remains elusive for broader-economic forecasters, despite isolated indicators like record stock prices. One near certainty at this point, however, is that the data center market is making a strong transition away from in-house facilities to outsourc- ing (cloud and colocation, for instance). Resources and spending will naturally flow, therefore, from a broader base of compa- nies to a smaller one, potentially making forecasts a bit more reliable. Factors that could hinder growth in- clude rising debt (public and private) and a possible technology bubble inflated in large part by near-zero interest rates and the absence of profitable investment opportu- nities beyond the stock market. Eventually, for instance, investors will cease to fund hundred-million-dollar quarterly losses by the likes of Twitter, potentially hitting the data center market by proxy. Demand for data center services is rising, but as to how service providers (in house and in the cloud) will respond to this demand in the future, we may not know until we get there. n