Who knows. It's not much easier to predict when it'll happen than what will actually happen.
For my part, I'm a bit skeptical. Like, I think the prospect of the singularity is super interesting, exciting, and a little bit scary, but I'm not actually sure it'll really happen, or be as big a deal as it's made out to be. Mainly, because the theory seems to be based on the evolution of hardware, on Moore's Law. That something interesting will happen when the exponential growth becomes a vertical line. But artificial intelligence, self-modifying robots and all that, making them isn't just a question of hardware, it's also a matter of software. And there's no sense of exponential progress there. C++ is still one of the most popular programming languages and it's 30 years old. The things they're doing with neural networks and Siri and what not is impressive, but it's not something that's gonna become self-aware by chucking more CPU cycles at it. These things only become smarter when we program them smarter, and that's a matter of man-power.
Not saying the singularity is impossible, and I honestly know little enough about it that I could just have misunderstood some part of the concept that undermines my argument. I just think it's more likely that technology keeps trotting on, gradually progressing, even after reaching a point of all hardware being redundantly powerful.