I highly doubt this will happen. Intel worked on their 22nm finfet process for 8 years while investing billions into it, this kind of thing doesn't happen overnight. The investment for double pattern costs billions upon billions of dollars, although that is a one time charge. Intel has already invested in the R+D for 14nm and is already starting to tape products out on it for release next year (Broadwell mobile), while TSMC is just now struggling with 20nm. TSMC's 20nm isn't even comparable to intel's 22nm with FinFet, and TSMC has not even invested into double patterning. TSMC is just now attempting to get good yields with 20nm and they don't have double patterning. Intel has all that and more, finfet, double patterning, and they're taping products out on 14nm for release next year. They've already quantum leaped TSMC.
Here's how it works in the silicon industry - the more money you have, the greater the lead you have, and new processes take 5-6 years to be realized. To earn money, you need money, and intel invested all of their R+D into finfet and double patterning years ago. TSMC is just now *thinking* about it. Now this obviously won't affect TSMCs business as intel is generally very bullish about enabling competitors and sharing their FAB capacities. TSMC will still get a ton of business from Apple, Qualcomm, nvidia, etc, so this won't affect them in the short term. But make no mistake, I can't see TSMC ever catching up to intel in the next few years.
Edited by xoleras - 7/5/13 at 10:08am