I can say that among AI researchers I interact with, it far more common to view it as wild speculation than as settled fact.
This is borne out by surveys of AI researchers, with 80% thinking strong forms of AI will emerge in "more than 50 years" or "never", and just a few percent thinking that such forms of AI are "near".
Software Developers are not the same as AI researchers, and I have found the Singularity myth to be much more widespread among developers. It has a nice ring to it: Computers keep getting faster, at some point they'll be faster than brains, at that point we just simulate brains. Soon after, we simulate something better than brains.
I suspect that the reasons AI researchers are less optimistic are rooted in the fact that we still don't have a good understanding of human intelligence, or even enough of an understanding of the brain to simulate it. For example, in the last two weeks we have discovered previously unknown types of brain cells. This gives the (correct) impression that even if we had a fast enough computer, we are not at all close to being able to accurately simulate a human brain. We don't really know what a human brain is.
Even if we did know that, simulations are necessarily lossy. We may not have good simulation techniques. Even if we did have good techniques, we may simulate the brain and discover our simulation does not behave as expected for reasons that we don't understand. This is very probable when simulating new systems. In some sense, proponents of the Singularity resemble people predicting that weather control was near in the 1940s. After all, we could simulate simple weather patterns already then, and generate forecasts that sort of worked. How much more complex could it really be to generate perfect forecasts?