I always enjoy learning about this aspect of the long transition from OS 9 and PPC to OS X and Intel, but I think it is disingenuous to present these stories as if the achievability of the task was uncertain and its completion was a surprising, watershed moment that got vice presidents to stop what they were doing and put Steve on an airplane.
OS X was promised for Intel; Apple shipped developer previews of OS X (as Rhapsody) for Intel. I have a copy of Rhapsody DR2 and have installed it on commodity PC hardware. It was a surprise and a let-down for developers when Apple decided to reneg on shipping the Intel version along with Yellow Box (Cocoa) for Windows. Rhapsody DR2 was released in 1998, this effort began in 2000 (before OS X 10.0, actually), so at most there is a two year gap where the status of the port is uncertain. It is likely that it was kept maintained for at least part of the gap, especially since Darwin was publicly released as an open-source project with Intel support in April 2000. What do you suppose the Intel version of Darwin was there for? No reason? It is certainly possible that in the crunch to get OS X to 10.0 for PPC the Intel port fell by the wayside and perhaps was no longer feature-complete, and this guy's job was to get it caught up, but it was still something they had sitting around the whole time.
That the Intel port could go from something they were already shipping in developer previews to something nobody within Apple even knew existed simply doesn't make sense. I understand that there were commercial reasons for Apple to maintain public silence on the topic, and to be very clear that their employees were expected to do the same, but I'm not about to believe these people convinced themselves that something they already knew existed as a shipping product didn't actually exist after all! In the wider world, the fact that Apple had maintained an Intel port of OS X was one of the worst-kept secrets in computer industry history. I agree that the actual time and place of the Intel switch announcement was a complete surprise, but the fact that the possibility existed was established when NEXTSTEP was released for Intel in I believe 1993.