IBM has, according to their website, 434,246 employees. So much for a small employee population...unless you meant 'a small country' or 'a small state.'
In any case, if you even took the time to read the 4-sentence blurb, you would see that they did this with 500 employees at their research center, which would still give 5 times your estimate of '10,000 or so.'
And I don't see why it would make any sense for IBM to give every person $10,000. The idea is to ferret out popular/worthwhile ideas. That doesn't really work if one or two people can fund it, as any of the examples could have been if each had $10,000 to spend.
And finally, your idea of profit sharing with regards to the selected ideas only works if the idea is meant to have some immediate financial impact. Procuring a 3D printer might not directly lead to financial results, but it may help someone rapidly prototype something that becomes a million dollar idea. You will never be able to measure the financial contribution made by that 3D printer, so why bother? Similarly, a disc golf course might provide some intangible stress relief, and employees may be a little more productive as a result, but how are you going to quantify that? IBM is pretty good at identifying business opportunity on their own...that clearly isn't the point of this exercise.