since the incompetent and lazy exams acceptances institutions (whom TI would never have been able to convince that their calculator was secure if the regulators had been only remotely competent, since the history of computing shows that basically all devices eventually get owned) would want to ban the Nspire (even the non-CAS) in order to be sure nobody gets the CAS on their calc instead of fixing their exams to be more in line with real-world usage, thereby removing a whole class of cheating sources.
What's more, as I've already outlined in other occasions, stealing does happen indeed... but in the direction exactly opposite to the one most people think about:
* (no longer valid for the CX series) Clickpad CAS calculators don't have interchangeable keyboards, so they are effectively simpler hardware than the Clickpad calculators sold as non-CAS... and yet, they're sold at a higher price tag than the model sold as non-CAS, but perfectly CAS-capable;
* (especially valid for the CX series, which doesn't have TI's slow, incomplete and buggy 84+ emulator anymore) the "a more complicated software costs more money" argument cannot fly in this discussion, since the one development costing extra money is clearly the non-CAS software: TI has to develop the CAS software anyway, for the model sold as CAS, and then strip the OS for the model sold as non-CAS. And yet, again, the non-CAS model is sold at a lower price tag.
I'll add that the cost of developing the Nspire's CAS should not be overestimated: I have shown the API to be the same as that of the now 15+-year-old Advanced Mathematics Software of TI-68k calculators, which was already 10+-year-old when the Nspire platform was introduced four years and a half ago (timestamps of the first OS versions)...