It really depends on the goals of the lab, I think. I think diy bio labs are first and foremost awesome for education, letting kids learn about biology and teaching them the basics of tinkering with cells / instruments. And if they don't mind not being high-throughput or competing with large labs, it's completely within grasp to study less popular systems (think more obscure microbiology, soil biology, worm biology, etc).

But there's also a movement of people who want to close the gap with academic labs on engineering biology, studying personal genomics, and producing new biologics. And there the limitations come down to quality / diversity of instruments and development of new instruments. No diybio lab is going to have a $500k mass spectrometer, but a basic set of incubators, centrifuges, microscopes, and thermocyclers puts them on the path to iGEM (Check out Counter Culture Labs).

Most projects there are still for toying around, trying to make proof of concepts, and learning the ropes of the wetlab. But with, say, robotics to automate things, cheaper sequences, or better DNA printers, they can start to explore more interesting problems that more traditional labs wouldn't normally approach.

DIY bio's science strengths are in pulling together individuals with hobbyist interests, but more diverse backgrounds. It's rare to find a biologist who knows how to build lab devices, and when you're not having to consider dropping $50k / yr on someone who isn't likely to win you more grants, I think it's easier to focus on tackling the more trickier annoyances holding biology back right now... if that makes sense...