There's really no such thing as “steady state” in the semiconductor industry, and fab operator Taiwan Semiconductor (NYSE:TSM) (aka “TSMC”) has the need to always stay on the leading edge of what its customers need and want in terms of chip manufacturing technology/capabilities. In addition to that constant pressure of migrating to the next generation, TSMC is facing the entrance of a major potential rival – Intel (Nasdaq:INTC).
Discount Brokers Comparison: Your one-stop shop for finding the perfect broker for your investments.
While I think Intel's technological capabilities are not to be underestimated, the assumption that TSMC is in serious trouble seems misplaced. That isn't necessarily to say that the shares are cheap right now, but I see TSMC as a share-gainer in the fab space for the foreseeable future.
It's Always About The Next Step
TSMC continues to be the share-leader in the fab space, as Samsung, GlobalFoundries, and United Microelectronics (NYSE:UMC) have not only failed to close the gap, but have seen TSMC gain share in recent years. Some of this leadership can be tied to scale and legacy platforms that allow the company to compete effectively on price, but some of it too can be tied to solid technological abilities.
That latter point is now being challenged more directly. Intel has decided to move into the fab space, and is bringing some serious transistor technology with it. Intel is about a generation ahead of TSMC in FinFET (also known as “tri-gate” technology), and that seems to translate into a roughly 12-18 month lead. At the same time, Samsung appears to be trying to skip ahead and stand shoulder-to-shoulder as a rival in 16nm/14nm FinFET in a year or two.
The first question I'd have, though, is how much the bleeding edge is really going to worth right off the bat. While companies in the graphics chip space (like Nvidia (Nasdaq:NVDA) or FPGAs (Altera (Nasdaq: ALTR) and Xilinx (Nasdaq:XLNX)) arguably need that best-of-the-best technology, I believe most of TSMC's customers do not and that it's technological lag behind Intel may not translate into such a dramatic market/revenue impact.
SEE: A Primer On Investing In The Tech Industry
Pricing Will Be Key
I think it's also worth noting that there are likely limits to what Intel can, and will, try to do with its fab business over at least the next couple of years. While I do believe Intel can use technology to gain share, I also believe they will have a much more difficult time of competing on price – TSMC has a major edge in scale/capacity and the ability to use the profits and cash flow from its mature nodes to subsidize pricing on leading-edge technologies if rivals want to go that way.
More to the point, I think Intel is going to be pretty selective in how it builds its business over the next few years. Being all things to all fab customers isn't going to work off the bat for Intel, and I believe the company has no desire to try to compete with the likes of Samsung and UMC other than on the basis of its tech capabilities. If anything, then, I can see this putting a lot of pressure on Samsung, GlobalFoundries, and UMC to hit what otherwise look like fairly aggressive plans for ramping up their FinFET processes.
SEE: 5 Must-Have Metrics For Value Investors
Mix Is A Threat, But TSMC Built To Last
Cycles come and go, and no company knows that better than TSMC given how long they've been at this. Although the company's sizable reliance on Qualcomm (Nasdaq:QCOM) (which contributes more than 15% of the company's revenue) is a threat, the addition of Apple (Nasdaq:AAPL) is definitely a meaningful win for the company.
All things considered, I see changes to the chip industry as a whole and key markets like mobile devices to be the bigger near-term threat to TSMC. If lower-end phones grab more share, it will generally mean less chip volume for TSMC and that's a negative on balance.
The Bottom Line
Although I am happy to take the other side of the argument from those who argue that Intel is going to significantly disrupt TSMC's share, it doesn't automatically follow that I'm a major bull on these shares. Modeling TSMC's cash flows is exceedingly tricky, as the margins can swing from the single digits to nearly 50% (and back again) in just a few years' time. Consequently, there's a larger than normal margin of error here.
All in all, I believe TSMC can post double-digit revenue growth over the long term, as I see more and more chip companies going (or staying) fab-less. The biggest unknown for me is the cash flow. I'm currently modeling a long-term average free cash flow margin of about 21%, versus a long-term historical margin of 22.7%, and that results in a fair value of just under $18. Those who think the future will more closely resemble the past could argue for a target closer to $20, but neither target makes TSMC a must-buy stock at today's price.
At the time of writing, Stephen D. Simpson did not own any shares in any company mentioned in this article.