One category sticks out that has yet to be addressed with a cloud service: Algorithms.

I think there’s a market opportunity to crease an AAS (algorithms as a service) company which provides simple APIs to implementations of common algorithms:

• Want to extract the color scheme from an image? Use our ‘Image Library Algorithms’ API
• Need to calculate an efficient path between two points? See our “Distance Algorithms API” documentation
• Trying to format text on lines of equal length? We have a killer Knuth and Plass line breaking implementation.

Algorithms as a service would give you development efficiency, problem scalability (access to CPU farms), and confidence in the results. The obvious negative of AAS (and most cloud services) is network latency. If an algorithm is trivial to implement and computationally simple, it should certainly be done locally. If an algorithm is CPU intensive and takes longer than 300ms to compute, the network latency becomes increasingly marginal and is a good candidate for a cloud solution (provided you can get the data to the server efficiently).