Are there phage-eating bacteria?

It's suspected that some protists (namely choanozoans and picozoans) are proper/direct "virus eaters" due to the size of their "eating apparatus" and scarcity of bacterial remains inside them (unlike viral remains), but the evidence is pretty recent; the (2020) study:

used modern single-cell genomics tools to sequence the total DNA from 1,698 individual protists in the water. Each of the resulting Single Amplified Genomes (SAGs) consists of the genome of a single protist, with or without associated DNA: for example, from symbionts, ingested prey, or viruses or bacteria sticking to its exterior. The technique is very sensitive, but doesn’t directly show the type of relationship between a protist and its associates.

The researchers found a range of protists including alveolates, stramenopiles, chlorophytes, cercozoans, picozoans, and choanozoans. Nineteen percent of SAGs from the Gulf of Maine and 48% of those from the Mediterranean were associated with bacterial DNA, suggesting that these protists had eaten bacteria. More common were viral sequences, found in 51% of SAGs from the Gulf of Maine and 35% of those from the Mediterranean, with a frequency of 1–52 virus types per protist. Most were from viruses known to infect bacteria – presumably representing parasites of the protists’ bacterial prey.

But choanozoans and picozoans, which only occurred in the Gulf of Maine sample, were different. These groups, neither of which have chloroplasts, are poorly known. Choanozoans (3–10 µm; also known as choanoflagellates), are of great evolutionary interest as the closest living relatives of animals and fungi. The tiny (up to 3 µ) picozoans were first discovered twenty years ago and originally known as picobiliphytes. Until now, their food sources were a puzzle, as their feeding apparatus is too small for bacteria – but ample for viruses, most of which are smaller than 150 nm.

Every single one of the choanozoan and picozoan SAGs were associated with viral sequences from bacteriophages and CRESS-DNA viruses, but mostly without any bacterial DNA, while the same sequences were found across a great diversity of species.

“It is very unlikely that these viruses are capable of infecting all the protists in which they were found,” says Dr Julia Brown, a researcher at the Bigelow Laboratory for Ocean Sciences and coauthor on the study.

The authors conclude that choanozoans and picozoans probably routinely eat viruses.

The news release also says that how these "virus eater" protists manage to protect themselves against viral infection isn't yet known/studied.

If this needs saying... protists (unlike bacteria) are eukaryotes. I haven't found any source suggesting that any prokaryotes (bacteria or archaea) manage to sustain themselves on viruses.