S3's launch and the forthcoming Storage 2.0 bloodbath

One of the companies I led a discussion on at the TechDirt Greenhouse last week was Openomy, which is a small storage startup developed by two young entrepreneurs. Openomy’s pitch was that their Open API was a key differentiator as compared to their competition (great overview on TechCrunch here).

My comment to them was that their differentiation (having open APIs) would only last a week (it was more an image than an inside knowledge of Amazon’s announcement), and that an ad-supported storage offering was challenging from two standpoint: one, the idea of a company crawling my personal files and data to ad-target me freaks me out – sort of implying that I have darker secrets on my hard drive than my gmail archive – and two, the notion of using ad support in a storage context was far fetched. Unlike so many other Web 2.0 companies storage and bandwidth are still expensive at scale, and storage does not benefit from economies of scale delivered by web caches. I likened this offering with the free ISP’s of the late 90’s that all disappeared from the face of the earth. One of my comments is that I was expecting a bloodbath in the online storage market b/c of the large number of companies attempting to penetrate it, as big as that market is.

Fast forward 3 days (gee), Amazon.com announced S3 (for “Simple Storage System”) yesterday, a grid storage infrastructure that is available through a set of web APIs at a cheap cost of $15c per GB of storage (per month) and $20c per GB of bandwidth used. It is actually interesting that Amazon decided to meter both elements.

Because of the economies of scale available to Amazon, it is clear that their cost of provisioning is much lower than any startup can hope to reach (unless it has built a unique storage back-end). Will all Storage 2.0 startups jump on S3 and use it as a wholesale filesystem, building their differentiation in the application and services they will develop on top ? I don’t think that this will be sustainable in the long term, especially as 1Gb of bandwidth can be provisioned for less than $20c at scale. However, as Mike Arrington wrote, it is going to create an additional pressure on all the companies looking at competing in the space, for which the cheap and dirty online backup is no longer the killer app.

Note that I am not saying that it is no longer possible to create interesting Storage 2.0 companies, it still was last week with the expected release of G-Drive, and it still is today. It just became harder.

 

Comments

[Ian Sefferman]

Hey Jeff,
Thanks again for leading the discussion this weekend. Maurice and I left the conference and on the plane starting talking about how we feel very confident that we have a new direction and new focus and can equally challenge the large players.

I wrote a little argument about why Amazon just missed the goal with S3 on my blog today:
http://www.iseff.com/2006/03/aws-s3-what-amazon-got-wrong.html

The one sentence summary of the argument is that essentially this just moves the problem of your data existing in 'silos' throughout the internet down one level (to Amazon, rather than each individual web service) -- you still don't have a single access point or control over your data.

Check out the blog for more info, and thanks again. Hopefully we can speak further about this. :)

Ian

[Aaron Levie]

Jeff,

Looks like I'm making the rounds trying to give an online file storage company's point of view. Anyone in the business, who has already prepared to scale, does not have costs higher than what Amazon is now charging. When you also add up the double bandwidth surcharge (file from user to you, then file to amazon, and vice versa), this becomes impractical for a large scale online file storage company. As far as I can tell, this is offering is focused on the crowd that traditionally uses 'web services.' It's great for mashups, non-storage startups, and others who need quick and easy access to a backend. Why doesn't this push our price points lower? Because you still need to pay for developers, webservers, marketing, bandwidth, and any other fixed or variable costs related to your operations. In Box.net's case, we have set our price points in a way that allows us to offer a free package without any advertising/spam, and a premium service with more storage and features.

This has turned into way too long of a comment... shoot me an email if you want more :)

Overall thought: http://blog.box.net/?p=28