The hCard microformat is a way of taking normal HTML (or XHTML) and putting additional contextual information into it so that software looking at the codes will be able to recognize the information for what it is.
An example of this would be the following:
(props to the hCard creator for creating this data)
If you examine the source to the above, you'll notice that it contains a bunch of XML/HTML that tags individual data, such as the first name, last name, company information, etc. These tags can be used by programs such as web browsers to automatically identify data that could be easily be put to use in other contexts. Imagine being able to click on an address in a Google Maps listing and have that added directly to your address book. This type of work paves the way.
One benefit of the microformat "movement" is that it doesn't require a wholesale adoption of new formats, just an agreement to add some more tagging to pages that already exist. In this form, the visual presentation remains the same, but computers can use the data to do things for users that were very difficult in the past.
It's an interesting way to handle the problem of formatting information for both human and computer interpretation—an area where Google is particularly interested. So, it should be no surprise that these microformats are being adopted for use within Google products, possibly as an inducement for people to use them outside of Google products. Imagine how much easier it is for google to classify data on a page once these formats become more popular.