As cloud-computing prices keep falling, the whole IT business will change.
IF THERE were a prize for corporate secrecy, Amazon would have an excellent chance of winning. Interviewing its executives can be like pulling teeth. Even trivial details are not revealed, such as the approximate location of the office of Jeff Bezos, the founder and chief executive, in the company’s headquarters in Seattle. Unsurprisingly, then, its quarterly earnings calls are mostly a dull affair. But financial analysts and many in the information-technology (IT) industry will pay close attention when the e-commerce giant releases results for this year’s first three months, on April 23rd. Nearly a decade after it launched Amazon Web Services (AWS), the company will enlighten its shareholders about the size, growth and profitability of its cloud-computing business.
The disclosure is meant to reassure investors: they are getting worried about Amazon’s chronic lack of profits and the amount of money it is spending—last year the company as a whole invested nearly $9 billion. But the announcement will also signal that cloud-computing (in which data are stored and crunched on remote servers) has come of age: AWS’s revenues are thought to have reached $5 billion in 2014 and to be growing at more than 50% annually. Analysts have already assigned AWS a valuation of $44 billion—putting it in the same league as incumbent computer-makers such as Hewlett-Packard, which has a market capitalisation of $60 billion.
Nothing to stand on
- Banking on de-banking
- Risky business
- Engaged tone
- The cheap, convenient cloud
- Ever more from Moore
- A very British business
Although AWS is much larger than the cloud business of rivals such as Microsoft and Google, they are also growing quickly (see chart). Gartner, a market-research firm, expects the global market for cloud-computing services to reach $176 billion this year. That is still only 4% of all IT spending, but it is growing fast, as most other parts of the industry are stagnant or even declining. By 2017 cloud spending will total $240 billion, Gartner predicts.
As cloud providers rush to build new data centres, and battle for market share, businesses are finding that the cost of putting their computing and data storage into the online cloud is getting ever cheaper. In the past three years prices are down by around a quarter, according to Citigroup, a bank; and further significant falls look all but inevitable. Some providers, such as Microsoft, have started providing their services free to startups, in the hope of turning them into paying customers as they grow.
The IT industry is going through a wrenching change, in a sense reversing the trend of the 1990s in which corporate computing shifted from giant, centralised mainframes to smaller, dispersed desktops and servers. Like most advances in IT, the advent of cheap online processing and storage has been brought about by Moore’s Law—the rule of thumb that the density of transistors on a microchip doubles about every two years, which celebrates its 50th anniversary on April 19th. And fortunately, Moore’s Law seems to have some life left in it yet (see article).
But for providers of cloud services, and their shareholders, the question is: will we ever make any money from this? So far, all that many of them have done is run up losses. The providers may find it hard to differentiate their services, forcing them to compete on price and thus to expand rapidly to achieve economies of scale. The risk is that they end up somewhat like airlines and mass-market carmakers: chronically afflicted by overcapacity, constantly struggling to achieve a decent margin and perennially hoping that their competitors will keel over first.
Just as airlines’ and carmakers’ losses are tourists’ and motorists’ gains, businesses and other organisations that make heavy use of IT services will enjoy big savings. Rather than buying lots of servers and hiring armies of engineers to maintain them, they will increasingly rent processing power and storage space cheaply; and whereas a system administrator might now manage a few dozen in-house servers, he will be able to oversee hundreds of “virtual machines” in the cloud.
This is not to say that all corporate computing will be pushed into the cloud immediately. Thanks to early cloud providers like Salesforce, many firms already manage and store in the cloud such things as their customer accounts, mailing lists and employee-monitoring systems. Companies have also put into the cloud many of what are called their “systems of engagement”, that is, the services that handle their interfaces with the public, such as smartphone apps. Any software that companies design for themselves—for instance, in “sharing economy” firms like Uber—is increasingly being developed and tested in the cloud, as well as run there.
However, many businesses are stuck running important parts of their operations on ancient bits of software, and are aware that replacing such legacy systems can be fraught with difficulty. Many such applications would need to be rewritten to run in the cloud, says John Rymer of Forrester, another market-research firm. And they are often too embedded in a firm’s day-to-day operations to be replaced easily with newer, cloud-ready software.
Heavily regulated businesses, such as banks, face a different problem. Although they may be convinced by the cost advantages and convenience of moving their systems into the ever-cheaper cloud, they will have to persuade both their regulators and the insurers which cover them against any data leakages or system breakdowns.
Cloud geeks, legacy geeks
So businesses are beginning to split their IT departments into two groups, explains David Mitchell Smith of Gartner. One lot grapples with keeping a diminishing number of creaking, legacy systems going, while the other develops and manages the snazzy new systems that run in the cloud.
Similarly, there are now two kinds of IT firm: those native to the cloud, led by Amazon, on the one hand; and the incumbent sellers of hardware and software, on the other, which are struggling to adapt to the new age. Software firms are not just having to rewrite their applications so they can run in the cloud, but also to switch from a business model in which they get much of their revenue from large, upfront licence fees to one in which they receive smaller, recurring subscription payments. A few have already made headway in this, such as Adobe, a maker of publishing software. Others are only at the beginning of the journey, including Oracle and SAP, two big providers of business applications.
Adapting is going to be harder for makers of corporate computer hardware like HP and Dell. Businesses that used to buy servers from them, perhaps a dozen or two at a time, will increasingly rent computing services in the cloud instead. And although the cloud providers will need huge quantities of servers themselves, they are increasingly designing their own, and having them built by low-margin contract manufacturers. As for the software that the cloud-computer providers use to manage their giant data centres, they are tending to go for free, “open source” software instead of the paid-for variety.
Computer-makers face a hard choice: try to establish a strong position in the cloud, or focus on their legacy business. IBM has opted for the former: in 2013 it bought SoftLayer, a cloud-service provider, and has invested in it heavily since. HP seems confused: earlier this month it seemed to announce that it would stop competing with AWS and other cloud-service providers—only to recant a few days later. Catching up will be hard for both firms: unlike Google they do not have a lucrative business in internet-search advertising to rely on as they make the change. Neither do they have cash-cows like Microsoft’s Windows and Office software, nor shareholders as patient as Amazon’s.
Meteorologically speaking, then, big IT users can look forward to blue skies containing cheap, fluffy cumulus clouds. Providers of online computing may be under a grey, drizzly stratus of low profitability. And legacy makers of hardware and software should expect stormy cumulonimbus formations overhead.
Powered by Facebook Comments