Saying that hobby licenses will not poach commercial sales is clearly not true. There will be some number who cheat.
Yes, there will be people who cheat, but probably not many, and not enough to make a real difference to the bottom line. Any reputable company will not cheat like this.
As engineers and technicians we should always be leary of unsupported assertions like this. I requires two numbers that have not been made publicly available to evaluate this.
First it needs the percentage of cheaters. I suspect it is far less than half, and more than a tenth of a percent, but that is a guess with obviously wide error bounds. I believe that your assertion falls within the error bounds I suggest, but the assertion that it hurts no ones bottom line in a meaningful way may well be sensitive to the exact value. Highly publicized events over the last few years involving large international companies like VW, Wells Fargo Bank, Enron and others indicates that size and reputation are not complete defenses against unethical behavior.
Secondly, it requires the sensitivity of a given software companies bottom line to this cheating. There is clearly no general answer to this, and it can only be guessed in most cases. The TI example is one of the few where the answer is obvious, their hardware sales so far outweigh the sales of their software tools that you can confidently assert that their bottom line is little affected by this issue. Even in those cases the livelihood of the managers and employees of a large department may suffer even though the company as a whole is only marginally effected.
There were certainly companies at the dawn of the personal computer era whose software was used illegally far more than it was sold. Most cases fall somewhere in between, but even the companies involved usually can't answer the question accurately because no good data exists to differentiate a cheating user from someone who just wouldn't have used the software if they were unable to cheat.
My whole point in this is not that lots of people are cheating, but that when trying to convince an organization to adopt a software distribution policy that they perceive as threatening it is not sufficient to express your opinion. "Don't worry, it will be fine." Your argument will be more successful if you can surface actual factual evidence that the policy proposed will not hurt, that in fact it will help sales. And as I suggest here, the data required is likely to be significantly different for different software distributors.