According to a new Penn State study on Airbnb, 17% of hosts in 12 major US cities rent out two—or more—residential properties year-round, drawing criticism from the hotel industry.

The study was sponsored by the American Hotel & Lodging Association and claims the full-time "multi-unit operators," as the study calls them, take in over $500M/year—nearly 40% of the Airbnb revenue in the cities studied.

The study is the latest in a line of attacks against Airbnb. A NYC City Council meeting last year weighed serious fines against the firm, and Airbnb CEO Brian Chesky was forced to pull his own couch in S.F.—the first Airbnb listing—off the site when it came out he had been renting it out illegally.

Among the markets surveyed, NYC, Miami, LA and San Fran had the most full-time operators, data from Airdna, a firm that analyzes Airbnb, revealed.

CEO of the hotel trade group Katherine Lugar says these Airbnb hosts are commercial landlords running illegal hotels, not folks using homesharing to supplement income.

However, Airbnb's director of public affairs, Nick Papas, tells Bisnow the study is deeply flawed and has its numbers wrong.

According to the study, the markets examined had 2,657 full-time operators, a category defined as hosts who rent space more than 360 days per year.

Yet Nick says there were only 17 Airbnb listings booked for more than 360 nights per year in those cities. In this case, the study shows the hotel industry gets what it pays for—a study he says is intended to mislead and manipulate.

Despite the criticism, Airbnb has grown at a blistering pace since launching in 2008. A recent funding round brought the valuation to a staggering $25.5B, a clear sign of just how much the firm is disrupting the hospitality industry.

Penn State professor John O’Neill, who directed the study, vehemently defends the work, calling it the most comprehensive study of Airbnb to date, and saying his team examined 416,000 lines of data and 9.5 million variables.