That two Hispanic surnames made the top 10 list in 2000 seems to reflect the country's growing Hispanic population. The proportion of Asian surnames in the population also increased. I'm happy to report that my surname, Liu, rose from a rank of 2300 in 1990 to 650 in 2000. That's higher than the surnames of World Almanac editors Joyce (#948), Janssen (#3101), and Steinitz (not among top 5,000) but well below Thomas (#14).
More than 6 million surnames were identified in Census 2000 records. About 26% of the population sample, or 60 million people, accounted for 275 of the most common surnames. Around 65% (4 million) of all surnames identified were unique to one individual.
The entire list of 1,000 most common surnames from the 2000 Census can be found at the link below. In addition, the Census document "Demographic Aspects of Surnames from Census 2000" (pdf) ranks surnames by their frequency among different races.
Frequently Occurring Surnames from Census 2000 (Census Bureau)
Genealogy page, including frequently occurring surnames from the 1990 Census (Census Bureau)
"In U.S. Name Count, Garcias Are Catching Up with Joneses" (includes name search function) (NYTimes.com)