I work with a team that has a sharepoint site currently runnning and its lists are linked to an access DB. My question is really on the investment level, what would be the reasons to upgrade DB to sharepoint if only a few 100 users access this site. Is there a real benefit to replacing the DB with a version of SQL Server, escpecially if about to replace Access 2002 with 2007 some day. I know SQL Server can handle more memory and traffic for more users, but I'm looking for more reasons than that if there are any.
If you have multiple users accessing the database it's always worth it to replace Access. Maybe the newer versions aren't as bad but in my experience earlier versions (2002 and earlier) had some problems with data becoming corrupted in a multi-user environment. I'd rather use SQL Server Express than Access. SQL Server Express will give you all the benefits of SQL Server (there are some limitations like the maximum size of the database and the number of processors it can use) but allows an easy upgrade path to a full copy of SQL Server in the future.
You might want to upgrade to the express version of SQL Server, which is free, and can handle databases up to 4GB. If you later decide that you want to move to another version of SQL Server, it will be much easier.
The biggest reason to migrate away from Access is stability and maintainability. Any real DB (SQL Server, Sybase, Oracle, DB2, UDB, ...) will have better backup features, better management features, and provide much better control of access and table management.
Let me stress first - I use SQL Server, I like SQL Server and it's unquestionably a stronger DBMS than Access with better performance and reliability. There's also SQL Server Express as a low-end option which is quite good enough for many purposes but gives the architectural advantages of SQL Server over Access.
That said - way back in the day, when I were a lad and budgets were tight.... I have used Access as a back-end for interactive data storage on sites with significantly more than a few hundred users (total, never measured concurrent). It worked, I don't think we ever saw data issues and wasn't particularly slow. Remember a web app is a rather different usage case from a windows app because users aren't continually connected to the database so concurrency issues aren't quite the same.
Access will remain a weak point in the architecture and it's well worth understanding its limitations (transactional integrity in particular) so you can make informed decisions on its use, but I wouldn't regard it as a given that it must be replaced. If it does the job now and there's no reason to believe it won't continue doing its job for the foreseeable future, consider letting sleeping dogs lie. The ROI from replacing it may not be there.