As shared hosting goes, the servers can handle quite a bit of traffic if the site is setup to handle a lot of traffic.
That is to say, if you think you are going to get really pounded, then using a freeware dynamic blogging software is not the way to go.
The easiest would be to have purely static content. That requires almost no CPU, no mysql access, no lags of any sort before being served.
Static content is not always what you need. In that case, you could don a hybrid idea. That is to say, you have dynamic content, but you don't serve it dynamically. In other words, let people post, add comments, etc, etc, and have a static page generated from that dynamic content every ten minutes. In that way, you serve static pages that change often enough for most purposes.
And if you truly need dynamic content all the time, the servers handle it fairly well (to hit the quota, you have to use more CPU than an entire 2 GHz opteron for over three minutes, or two full cpus for 1.5 minutes, etc). I have seen very little well written and optimised code that churns a cpu that much. If you do happen to have that, and you expect that much traffic, then a dedicated offering may be what you need. Of course, if you do full dynamic at those kind of loads (10k simultaneous users) you would need a cluster of servers.
On the other hand, I have worked for several companies that assumed a roll out would give them exponential increases in loads, traffic, etc, and very rarely does it pan out (the only exception was years ago at Holiday Inn Corporate where marketing didn't tell the I.T. groups about a superbowl commercial that strongly pushed the website. It crushed the network connection. But I digress.)
Anyway, to end my rant, I would just say, if you expect a lot of traffic, the easiest (and often best) thing is to use static pages, at least until the heavy load has passed.
Principal Systems Administrator