I was speaking with Jeff Bezos here at ETech earlier today about the Amazon Web Services business, and we got around to the topic of competitors-to-be. He asked me who I thought might go into that business. (Not that Jeff needs advice from me.) I said that it would be limited to companies with a strong reputation for datacenter infrastructure: Google, Microsoft and IBM came to mind quickly, although I discounted Microsoft because when it comes to compute services, I felt it would be politically difficult for them to offer LAMP-based servers. Jeff added HP to my list, which makes good sense.
But we agreed that it’s one thing to venture into the storage/servers-on-demand business and it’s another to do so successfully. As Jeff pointed out, it’s a high-volume low-margin business, in which Amazon has been notably successful. What about these other players? How might they succeed under those circumstances? HP’s heritage is quite the opposite — high margin — although with the Compaq acquisition and subsequent upheavals, they have had some success in the lower-margin PC business. Microsoft doesn’t have much of a track record in this commodity-type market. Google probably does. IBM certainly doesn’t — Global Services is a high-margin business — but if they started an entirely separate business unit with an all-new economic model, they might be able to do it. IBM certainly has the infrastructure experience. So my guess is that Google is the #1 potential competitor to AWS, with IBM and HP as less-likely players.
Whom am I leaving out here?
Blogarithms 
eBay or Yahoo come to mind.
LikeLike
Don’t discount Microsoft just because of LAMP lack; remember they get SW licenses for “free”. IBM has the experience, but can they cut the margins to compete? Any system vendor can build the boxes, so add Dell. Storage is an important cost contributor, so add EMC, especially with VMware to leverage, but probably as a team play (with Dell?). Don’t overlook the large presence of a computer-savvy retailer like Wal-Mart or Target. Maybe one of the not-so-Baby Bells will step up, or Best Buy/Speakeasy.
LikeLike
You can expect every web hosting company to eventually go after Amazon. For S3 especially, the gross margins on data transfer are huge and the storage side ain’t bad either assuming a 3-5 year useful life.
LikeLike
A pile of Sun shipping containers. 🙂 I’d suggest to Jeff that Amazon should buy Vertica, complete the stack, distribute globally and forget about the competition until Google joins the party.
LikeLike
Most notably, you left out possibly the largest player in this space, the Sun Grid (http://www.sun.com/service/sungrid/index.jsp).
LikeLike
AWS will likely encounter Gperil. Google’s distributed server farms are a potential EC2 threat. Gmail is a watered-down version of S3, to jump-start: loosen the file size restriction from 10MB to 5GB, strengthen Google Checkout security with secret keys, and Gbrand. Its dark fiber binge makes Google (now the world’s largest owner or controller of fiber bandwidth) the net’s de facto speed daemon so why wouldn’t Google open APIs to spare Hz and GB? Gturk could impress more than a few requesters with ajaxy flowcharting (i.e., Yahoo Pipes + Coghead), etc.
If Amazon enjoys strong patent protection, threats to AWS might even come from foreign startups. It’s conceivable that bitorrent could compete with S3 if nodes somehow became motivated to reliably host strangers’ encrypted data. A non-US based Mturk imitation could weaken Amazon’s (not to mention oDesk’s) market share if it: reverse engineered the Mturk API, provided tax cover for workers and undercut the 10% requester fee.
LikeLike