Yes* with an asterisk. But the first question I would ask is this:
How else can we solve this design problem?
If the alternatives aren't pleasing and simple, then—despite the fact that nested tabs are against the Microsoft Guidelines—I would consider a tabset-on-a-tab solution. Why? Because GUI designers aren't always tasked with solving simple problems—especially outside the domain of the Web, where clients can bring you problems much more complex than "Tell the user X" or "Sell the user X", and where users may have specific needs that prevent you from temporalizing a complex task into wizard-like steps, or from simplifying the user interface by progressively disclosing the occasional controls, or any of the other tricks designers can use.
I think that, with a broad interpretation of tabs, the answer to this question is Yes* with an asterisk. The asterisk follows: Perhaps the tab sets should differ in appearance (which may, in turn, cause you to name them something other than tabs—such as navigation bars, blinds, wizards, and so on). Whatever you do, when your design is complex, check your prototype with users; do they find tabs-on-tabs too complex? If so, re-read the first paragraph, above, and be prepared to iterate.
On a personal note, a complex project I'm working on will soon have a working prototype that includes nested navigation. I'm looking forward to some testing with users, to see how they fare. A month from now, I'll have tested the design with real users and I'll be trying to adapt the design if user testing reveals major problems.