The rise of TSMC and the tragic decline of Intel, Part 2

I’ve written in the past about TSMC’s revenues and how they’ve been growing extraordinarily fast, and how the company could easily blow past Intel’s current 2022 revenues expected to be at around $70B. I am predicting TSMC is on schedule to do that this year, whereas Intel’s revenues are at best flat for the year. TSMC delivered $18B in revenue for Q2. Revenue grew some 40 percent year-over-year. Compare this to Intel with $15B in revenue, down 17 percent YoY. I believe TSMC could do as much $90B this year. In the past four quarters, Intel’s revenues were about $76B. No matter how you look at TSMC, this is insane growth! Just a mere 10 years ago, TSMC's revenues were $17B for the entire year. The company has now generated that entire amount in a single quarter. This is absolutely stunning and so far there doesn’t seem to be any let-up in this growth.

So, with Intel’s current market cap at around $100B and TSMC’s at around $300B, one might argue that this is not exactly a good apples-to-apples comparison of the two companies. First of all, Intel does not have the advantage of being a pure-play foundry like TSMC. Intel is considered one of the few “integrated device manufacturers” (IDMs) like Samsung and Texas Instruments. We know that for some reason IDMs usually don’t enjoy the same multiples that the pure-plays have, even though the foundry business in a highly capital intensive business. Intel, like Samsung, does a lot of other things besides fab including chip design, board design, and systems design. It’s difficult to breakout these two Intel businesses (fabs and design) to figure out what percentage of their revenues come from their foundries to do a true apples-to-apples comparison. They don’t report their numbers this way in their quarterly reports.

However, one thing we do know is, that Intel does indeed possess significant total manufacturing foundry capacity around the world. This is just a guess, but the income that results from Intel’s production manufacturing foundry facilities is perhaps the majority of their revenues. Their number of fabs has declined over the past decade, but Intel is still one of the largest fabs in the world. One metric used by industry to compare different fab companies is “wafers per month.” One estimate has Intel at roughly a million wafers per month, plus or minus; whereas, TSMC does approximately two and a half million wafers per month. If we were to do a quick back-of-the-envelope, apples-to-apples comparison using these metrics, and assuming that 50 percent of Intel’s revenues come from their foundries, what does this say about the current value of Intel’s foundry business? I think it is greatly undervalued and probably not realizing its full valuation potential. Likewise, the value of Intel’s design side of the business is also not realizing its full valuation potential. Their design business should have a much greater valuation also.