we don't run SQL on vmware but if you were going to do it you would have to watch the I/O. even SQLCAT's testing showed that all the I/O goes through the hypervisor software and it can become a problem on busy databases.
there is also cost. by the time you buy the vmware licenses for everything it's a lot of money. compare it to clustering and you'll have to decide if it's right for you.
vmware originally had a killer ROI on solutions where someone would ask for a 1U server to run an app and it would take 10%CPU at most and a little RAM. Buying a few machines to run vmware was a lot cheaper than server creep since you have to buy a lot of accessories for servers. $3000 racks that do nothing but sit there and power. with SQL and new servers you can consolidate a lot of databases to very few servers. vmware/clustering is just for availability since there is no reason to run a server with only one or two small databases unless you have to for security/customer or some other reason. if your main concern is having current data at a DR location check out some of the disk backup software where it ships your backups to another unit
clustering will give you a better return on investment since you can run on both nodes at the same time. with vmware you have to have a server bought and paid for sit around and do nothing.
it's like backing up to disk vs tape and dedupe. every solution has costs, issues, problems and advantages and you have to decide if it's right for your environment