There are two factors that work against in-body stabilization:
1- Limited Range of Stablization
2- Limited size of camera bodies.
Olympus's solution is a compromise, much like a laptop computer is a compromise. In a laptop you compromise performance for better size. Someone arguing for the superior capacity of desktop hard drives, might point out that you can get a 3TB hard drive for the desktop. When you say "I'm not sure its about size, Olympus has it", you're doing the equivalent of saying "Yes, but laptops have hard drives too." That's true, but they max out at (lately) 1TB.
There is a dramatic size difference between the APS-C sensor in the NEX series, and the m4/3 sensor Olympus is using. This has two impacts that are significant regarding in body stabilization. The first is, a larger sensor has more mass, and that requires exponentially more force to move it, especially at high speed. Exponentially more force may likely require exponentially more size within the camera. Secondly, this larger size sensor, also means that the sensor has to move further (physically) to account for the same degree of performance (optically). That is to say, adjusting for 1 degree field of view's worth of shake requires moving an APC sensor a lot further (percentage wise) then moving a m4/3 sensor to account for 1 degree FOV's worth of shake. In over to move the sensor further, you also need more space to move the sensor, proportional to the increased size of the sensor. (and of course, that makes it harder as well, on the systems that move the sensor, possibly also increasing their size.
Another factor is that there simply may be less size available, proportionally, for the NEX system than the Olympus cameras, due to a comparable body size (or even smaller with the NEX, not sure) and a larger sensor size inside the NEX meaning less room for the electronics and other things the camera needs to operate. Larger shutter, larger lens mount space taking up internal volume around the sensor, etc.
Finally, even assuming all of these issues could be overcome on the NEX, delivering a comparable degree of movement to the Olympus IBIS system, this would still produce less shake mitigation than an optical system in the lens which can more easily accommodate larger amounts of system movement than any in body system.
So, yes, Olympus was able to do it for their camera, and in doing so they made a choice of compromises based on their priorities. Sony made another choice, but it isn't that sony is being mean and just choosing to do it this way... to do it in the body would require compromises of several types.