Amazon.com announced that it is giving some cloud computing away for free, presumably to get developers hooked on its service. 

What I've seen with countless startup/entrepreneur customers in our company's cloud services is that this kind of offer can be more of a curse than a boon for the startup:

• The successful startups inevitably need far more resources that a fraction of a physical core which the Amazon offer of a micro-instance provides, so this offer really wouldn't have any effect on the finances of a startup.  If the startup used this offer as a criteria to choose a provider, they would be ignoring the factors that correlate with successful deployments.
• The startup customers of ours that scaled successfully understood how to architect their systems for growth and institute processes that allowed them to provide a reliable service, such as change control, deployment management, incident response, resource management, and performance management. These skills came from staff members who were not developers, so creating an offer to appeal to developers works against the long-term success of a startup by placing operations in the hands of people without the skill-set and motivations that result in successful production IT management. Until self-service managed cloud offerings appear that eliminate the needs for these skill-sets and experience, they are still required to scale successfully, or they have to be outsourced.
• By limiting the offer with restrictions in so many dimensions (due to Amazon's rather complex offering), this "free" offer places the burden on the user to compute how to stay within these bounds (actually, an impossible task) which ends up "rewarding" the successful startup with large and unexpected bills.   I've seen people at countless Cloud Camps complaining about this "surprise bill" effect.  To some extent, this is inevitable with cloud computing and is even a benefit in the sense that customers can take advantage of scalability (if they design for it and enable it) but the "freeness" of the offer is fundamentally misleading. The alternative for the user would be to resolve to stay within the bounds of the "free" instance, which throws away the benefits of cloud computing over say, colocation (not to mention that you can get a lot more horsepower from a colocated server per dollar despite its other limitations.)
• The aim of this offer - to lock customers in with a free tidbit - works out in Amazon's favor but some of our most successful startup customers (some of whom came to us from Amazon) ended up deploying architectures that were not compatible with Amazon's architectural restrictions and performance limitations.  In other words, vendor lock-in wouldn't or didn't serve them. One very successful customer of ours is currently spending $1M/mo with Amazon and desperately struggling with how to move out of its dependency on Amazon.

None of my points are meant to say Amazon is a bad choice since that's clearly not the case, but the right answer to what cloud provider to choose for a company is "it depends" and what it depends on shouldn't be reduced to the price of one free micro-instance.

What ENKI would like to offer instead of a "free micro-instance" is our SSP (Startup Success Program) which offers up to 50% discounts on the first 32 cores (equivalent to 96 of Amazon's ECUs) as well as other benefits, in exchange for allowing us to assist you in managing your services so that they are reliable, cost-effective, and appropriately performing for your target market, as well as participating in joint marketing which can benefit your brand.   If you are interested, please contact us.