Just a note, Cree claims efficiencies of around 100 lumens per watt also, depending on chip temperature.
http://www.cree.com/products/xlamp7090_xre.asp This is only one of Cree's LED chip products, and there are many other great white LEDs by companies such as Osram (Sylvania) and Lumileds (Phillips).
I'd be curious to see the PAR value of Cree's cool white LED products. I like LEDs for the fact that they are solid state ie: less prone to breakage, and they are rated to last 50 000 hours, though the actual useful life will be less for our purposes. A big detractor from xenon bulbs is the rated 3000 hour life, which is less than both MH and fluorescent... wow, hope the bulbs are cheap!
I'm glad that forums like this exist so that open discussion can make us think of new ideas. If xenon can produce higher PAR values, perhaps we will see the advent of 150w xenon metal halides in the future... though I doubt that the aquarium / horticulture segment of the total metal halide lighting market around the world would be enough for manufacturers to start producing larger xenon bulbs. If indeed the lumen output of these bulbs is lower than our traditional metal halides, I can't see these lights making it as a replacement as they are not fit for the traditional use of HID lighting (task and area lighting) even if the light is superior in terms of plant growth.
I still maintain that this (xenon) is probably the BEST way to light a pico to small nano (less than 10g), at least if you want SPS or other intense light requiring organisms and don't mind changing a bulb more often than usual. I suppose you could use PC, but then you don't get those glitter lines that I so love with MH lit tanks.