I believe that UV requires close/short distances to the UV source and moderately long exposure time and that increases the investment and operating costs which limits use on larger systems. Suspended particulate / cloudy water would also be an issue for UV. Ozonation or Chlorination are generally more cost effective for bulk - low risk uses. Sand filtration followed by ozonation, HEPA, and/or one to three stages of reverse osmosis would be used for potable / sterile water used in pharmaceutical or food manufacturing.

UV only kills biological life and may activate certain classes of molecules as catalysts for, by example polymerization. The bottom line is UV light can eliminate matter nor remove anything. The materials those dead pathogens or other life forms are made of are still there in the water.
UV cannot get rid of anything but it can kill. As noted it, to be cost and performance effective, water must be treated as a thin film ideally but up to say 1/4 to maybe 1/2 inch long mean path to center for the light to maintain a declining intenstity with distance from the source, noting that bulbs also weaken over time. The longer the mean path to center, the more powerful the bulb exponentially. It's why UV curable materials are inly cured in thon layers or with dual polymerization systems.
Overall, water treatment for the new age will alikely move away from sophisticated filtration and reverse osmosis beyond bulk drinking water, especially where distributed, hyper-local water purification is required. Examples include the home, apartment building, hotels, hospitals, factories, restaurants, isolated communities like military bases and farms and the like. Aquaback, a well funded startup out of New England uses very low energy evaporative (nit distillation which is energy intensive and highly inefficent) to clean water to near 100% pure water. Few molecules that are not already gases are lighter than water and so gentle, non aerosol or bulkier evaporation leaves everything ehind - from radioactives to live or dead pathogens to almost any material component. It's inexpensive and emerging as the dominant way to clean water and may even surpass UV.

For industrial applications, one approach is to use a hybrid of chemical and UV application for biocide application. Each has its unique strengths but together can be very effective.
Two reference links below for the treatment of cooling tower water -
https://www.process-cooling.com/articles/89074-uv-irradiation-vs-chemical-disinfections
http://www.waterworld.com/articles/iww/print/volume-8/issue-6/feature-editorial/uv-disinfection-of-cooling-tower-water.html

I have worked with large UV water sterilisation systems (they do exist) however the input requirements require optically clear water, so cooling water systems (which are frequently contaminated with oils, corrosion products and other dirt as well as steam condensate) are generally unsuitable for this application. The large quartz tubes required limit supply pressures and residence times limit maximum flow velocities.
The major advantage is that these systems do not add chemicals to the water so are particularly suitable for systems where the presence of chlorine and/or bromine are a bad idea, such as injectable pharmaceuticals and are used to sterilise distilled water systems (which may not kill bacterial cysts or viruses (which can carry over in aerosols in the still system) which however sufficiently intense UV will disrupt).
For cooling water systems and drinking water, circulating biocides are actually an advantage as they prevent microbal mats from growing in the distribution pipework and consequent disease propagation. For these reasons large systems are an exception rather than the rule. If your cooling water piping is copper or nickel alloy however this is less of an issue as the materials will kill most microbes, and the absence of chlorine would actually be an advantage in reducing corrosion.