Well the approach to this question depends on what the interview was for - if this was for a tech-focused ER/HF/PE/IB position, well, let's hope you know something about semis. Otherwise, they're more interested in your thought process. If you can throw in some technical things while you're at it, great. That said, I'll approach it from both angles, with industry-specific knowledge and without.

Let's go with #1) "What are the key drivers and risks in Semiconductors"

No industry knowledge - hypothetical thought process:

Well, semiconductors involve a lot of R&D and then manufacturing to end customers in all sorts of industries. From the supply (R&D and MFG) side: 1 - What if R&D is unsuccessful (on both manufacturing and product)? 2 - What if manufacturing is not as efficient as originally hoped? Demand side: Well, Intel's customers are big PC manufacturers, and people don't buy new computers unless they have to (neither do businesses for that matter). So the business might be cyclical. And what if the competitors do better on manufacturing/R&D than we do? Competitive risk. 1 - Cyclical business 2 - Competition on both manufacturing and R&D 3 - Technology risk - what if the product is a flop or falls out of favor?

Now with some industry info (before someone flames me here I don't suppose to be the be-all end-all of semis, just some pointers to help a monkey out)

I would ask for clarification on which kinds of semis we're talking about. There are several different categories: 1) Vertically integrated players (examples: INTC, AMD (eh, we'll grant them the globalfoundries stake), STM, ATML) - these companies both design chips and manufacture themselves 2) Fabless semiconductor players (examples: NVDA, BRCM, QCOM, MRVL) - they design chips and have companies in group 3 manufacture them 3) Contract foundries (examples: TSM, UMC, GlobalFoundries) - they make other people's designs. 4) Semiconductor equipment (examples: XLNX, ALTR) - they make design hardware/software packages

Each of these categories faces different sets of drivers, and each company within will face different demand outlook - for instance, Atmel (ATML) is in such a wide variety of end products that its business cycle is not as cyclical as say NVDA, which relies on the higher-end graphics market for revenue.

...

2) WACC

No knowledge again:

If you assumed earlier that these companies are fairly cyclical, then you could approximate the beta for the sector as 1, adding perhaps 0.2-0.5 for the additional technology risk, at which point you just pick a beta in that ballpark and use CAPM (10yr treasury at 3.85% today and an ERP of ~6% yields a WACC of 12.85% w/ a beta of 1.5).

Adding some industry knowledge

For sake of simplicity, let's use the aggregated list of companies above. The straight average beta would come out around 1.5 I think (just spot checked a few on Yahoo), and then do same as above.

Phew. Hope that was helpful.