This is exactly the same behaviour as my Core2Duo 13" MBP and a couple of others I have seen. So I don't think your i7s are faulty. Nice to see someone doing real objective tests.
The Intel developer documentation says the CPU should not exceed 105 C, and recommends that it its operating temperature should be 95 C. Apple seems to take use these numbers verbatim and designed a fan control algorithm which lets the CPU reach, but not exceed, 105 C for a brief while, then pulls the CPU back down to around 95 C.
I guess Apple's thinking is that most high CPU loads are transient, so there's no point in winding the fans up for a few seconds of high temperatures. If the high load is not transient, the fans wind up and pull the CPU down to the Intel recommended temperature.
I firmly believe that the Mac uses the CPU temperature (not the heatsink) as its setpoint. if I wind my CPU up and watch the temperatures and fan speeds I see the fan speeds hunt to keep the CPU at 95 C. The fan speed changes from about 2500 to 3500 to achieve this. Also the heatsink temperature can change depending on eg GPU load, but the CPU sits at the same temp, about 95 +/- 2 C.
Whether it's good for the CPU or not I wouldn't like to say. But rest assured your i7 MBPs are not faulty - it looks like their thermal management works exactly the same as the millions of C2D uMBPs out there.
I noted a rumor recently that Google was negotiating with Intel to have its server CPUs re qualified to a higher temperature (presumably 100 C?) so that Google could reduce its server farm cooling requirements. It seems Intel & customers take these temperature recommendations deadly seriously and Google and Apple at least want to design right up to them. I guess therefore Intel knows exactly what it's doing and has qual'ed its CPUs for thousands of hours at 95 C. But that's a guess.
So I think the original 100 C article was designed to mislead to maximise exposure. They could have done the same test with a C2D, an i5 or an i7 any time in the last couple of years.