If I understand it correctly they achieved a conversion efficiency of 3.42% for a less efficient technique that is not used by ASML for producing EUV. Seems it is more like making progress than a breakthrough. A good light source is essential, but building a machine comparable with the ASML machines requires a lot of engineering in other areas as well.
Typically when a technology doesn't meet a critical constraint, the marketing will laud some other less important attribute.