The $7 Billion Dollar Cloud Mystery
I feel like I’m playing a game of Clue all by myself (it wouldn’t be the first time).
No one seems to have questioned IBM’s $7 Billion dollar 2015 target on Cloud computing. It was just swallowed up in the news as if it were written on stone tablets and delivered by Moses himself.
Well, I have a confession to make. The figure is a mystery – at least to me.
Ever since IBM announced its $7 Billion dollar 2015 Cloud forecast I’ve been thinking about it. The figure strikes me as being very conservative even by IBM standards. Granted, my “non-calculation” is just a gut feel, and knowing IBM they probably had input from 100+ finance, business experts, and McKinsey consultants to determine that number.
However, unlike the velocity of Internet adoption where the network and infrastructure had to be put in place, my feeling is that the velocity of Cloud adoption will be much faster than its historical paradigm shift analogues – Client/Server or Internet adoption.
It’s true that Cloud data centers have yet to be built out in many countries in order to comply with privacy and regulatory constraints, and those take time to build. However, between now and then many enterprise customers will be buying infrastructure components to build out their private clouds. Seems to me that a lot of that hardware, software, and services will be included in that $7 Billion dollar figure – 2015 is 4 years and 9 months away (end of year), but even 4 years (beginning of 2015) would be $1.4B per year.
Sounds like a lot, but IBM operates in 20+ industries, 170+ countries, and has an extensive portfolio of hardware, software, and services. A leading market position and sitting on top of a powder keg we call Cloud and they only see $7B by 2015? If we divided the $1.4B per year by say, just the G20 countries, that’s $70M per country. Seems low.
Lets not forget:
- Global Fortune 1000 building private (enterprise) clouds and hybrid clouds
- Governments building private clouds
- Financial markets and their (very secure and expensive) private clouds
- IBM’s own Smart Cloud offerings
Heck, even the U.S. government plans to spend $20B this year alone on Cloud, and we know IBM will be bidding on those opportunities. I’m speculating here, though I doubt they are calculating an average $1.4B on a straight line basis – more than likely Cloud growth will be closer to an exponential curve (how steep we’ll find out).
The global cloud computing market is expected to grow from $37.8 billion in 2010 to $121.1 billion in 2015 at a CAGR of 26.2% from 2010 to 2015. SaaS is the largest segment of the cloud computing services market, accounting for 73% of the market’s revenues 2010. - Salisonline Report
Maybe IBM won’t be playing a part in the SaaS portion of the market that would leave a $33B market – a 21% market share would get them close to $7B. Are they forgoing the SaaS market?
A couple of additional data points:
- Gartner estimated the Cloud market to be $148 billion by 2014.
- IDC estimates the Cloud market to be $55 billion in 2014
I’ll be the first to admit that the numbers regarding the overall size and growth rate of the Cloud computing market are all over the place, so lets pick the lowest number – $55B in 2014 and use IBM’s $7B 2015 number. IBM expects to be the market leader and win just 12.5% of the overall market? How highly fragmented do they envision the enterprise Cloud market becoming?
IBM’s data must be telling them something else.
Why just a few years back, in 2009, they were going to buy Sun for $7B. A collapsing Sun was worth more than IBM can manage between now and 2015 in the global Cloud market?
Between now and 2015 we can expect both a growth in infrastructure sales for new data centers and high-end virtualization optimized gear, and a decline in traditional hardware gear as organizations prepare for a shift to the Cloud – though a rapid decline in its traditional (hardware) business would not be included in a 2015 Cloud figure.
Perhaps enterprise (private) Cloud adoption will be much slower than the rest of the market, and that is IBM’s primary target market – large enterprise. Could that account for the overly conservative figure?
Could they be accounting for the threat of increasing infrastructure commoditization? As evidenced by Facebook’s OpenCompute Project – just as opensource outpaced traditional software development, opencompute seeks to do the same with hardware. Could that account for the conservatism? Doubtful. Even if it were the case, surely that’s a much smaller impact that the overall global Cloud computing market.
Maybe they’re accounting for systems rationalization resulting from the impact of x-86 multi-core - 32nm Westmere-EX E7 Xeons, which support up to 10 cores deliver up to 40% better performance than the 8-core Nehalem-EX chips. Supposedly, 1 rack of E7 Xeons can replace 18 racks of 2006 dual-core servers, and AMD will follow shortly with 16-core Interlagos processors mitigating the market need for some of IBM’s lower end RISC based systems and potentially rationalizing IBM’s System x line up.
After all, Cloud computing is largely a horizontal scaling game, and the x86 architecture is where that game is being played out. Besides, IBM stated long ago that they intended to reduce their hardware exposure as it was increasingly being commoditized, and the higher margins were in software and services.
Could it be that given the competition, with legacy hardware players migrating towards software (HP) and services (Dell), and new players – Salesforce, Amazon, and Google targeting the enterprise, that IBM is accounting for an extensive competitive impact on their emerging Cloud business? Is IBM de-leveraging their 50,000+ member partner channel? Not for now.
A 12.5% “leadership” market position is a lot of market share to lose – that would make for a much smaller IBM.
IBM hedged their bets 10 years ago and has changed its business mix considerably from hardware which was 24% of pre-tax income in 2000. In 2010 it accounts for just 8% of PTI while packaged software is now 44% PTI vs. 25% in 2000 (IBM 2010 Annual Report).With the Cloud train on the horizon that’s a huge high margin business target and one that will eventually be shrinking as Cloud sweeps the globe.
A Cloud computing model undercuts both traditional hardware and software – though that would be a statement on IBM’s overall business, not its $7B Cloud figure.
By 2021, ten years out, the Cloud party is over and we’ll have ~65-70% of all workloads migrated to the Cloud and ~30-35% of workloads remaining in traditional IT. That’s total and complete speculation on my part. Nevertheless, I use a broad brush to point out that (IMHO) the golden years of Cloud are upon us in these next 3-8 years, yet IBM, a $100B IT leader only plans to capitalize on much of Cloud’s sweet spot (2011-2015) to the tune of $7B? In a market estimated to be $55B to as much as $148B by 2014?
Are they underestimating the speed of migration to the Cloud? The size of the Cloud market? Its growth rate?
Big Data in the Cloud will drive storage sales, though the price of storage continues its decline. However, the volumes of data are growing faster than most company infrastructures can support and running analytics against that data is very compute intensive – so it’s either more gear, more (private or hybrid) cloud, or running those jobs in public clouds. All of which should be good (Cloud) news to IBM.
Building out data centers is a capital-intensive business – maybe they’re seeing more industry consolidation mitigating the number of potential large-scale providers, and those large-scale service providers can take advantage of economies of scale and provide services at lower price points than smaller competitors – presenting a barrier to entry and driving the smaller players out.
Even if one agrees with Larry Dignan’s recent article (quote and link below), that shift is not likely to be a radical decline between now and 2015 – and even if it was, IBM should be able to capitalize as the enterprise (private cloud) data center outsourcer of choice. Wouldn’t those dollars be considered Cloud?
Why hardware giants starting to sound like Amazon Web Services? They are all facing a stark reality years from now: No one wants to build their own data centers. Data centers are meant to be outsourced to some other company or taken to the cloud. -Larry Dignan [Source: ZDNet Article]
So what could account for such a conservative estimate? Slow moving Fortune 1000? Industry consolidation? Slow Cloud adoption? Infrastructure commoditization? Slow economic growth? Growth in multi-core and virtualization compressing number of servers? Smaller Cloud market size forecasts? Slower Cloud growth rates?
Maybe I’m over thinking this, and it’s nothing more than appeasing Wall St. by being able to “surprise to the upside” every year between now and 2015.
Do you find IBM’s $7B dollar estimate low given their place as a global IT provider? Have any ideas on what they may be seeing, which would account for such a conservative figure? I’d love to hear your thoughts.
Until then, the mystery continues (at least for me). Was it Mrs. White with the candlestick?
By Ray DePena