This is something that can go in several directions:
- The increased use of collaborative methods to develop solutions (as opposed to applications)
- A move away from complex design and development projects towards simpler models (a more top-down approach)
- The use of small simpler applications combined to build more complex applications. (a more bottom-up or mixed/service-oriented approach)
While Jason Fried of 37Signals believes the idea of enterprise software is "dead", it's more likely a space that small projects just don't play in. Also there may be a great many more small projects going compared to large projects but it's difficult to argue either way that a number of small projects equals one large project or vice-versa. (Obviously, I'm not defining what "small" and "large" means; think of those as what they mean to you)
On the other hand, small projects may be a difficult idea for some organizations to handle, especially when they are focused on going for large ones. It's a matter of overhead; With each small project, the relative size of the overhead for running the project may seem a lot higher than the overhead of a large project.
However, the techniques needed to operate a large project many not be the same as the ones needed for small ones.
The real winner is the one who figures out what kind of project management techniques are most appropriate for any given project, anticipating its complexity and size. Of course, if you could see the future and know how much work was involved right away, it'd be a simple trick.