In the months since Intel abandoned its acquisition of Tower Semi much has been written about Intel and what it will do without Tower, but far less thought has been given to what will happen to Tower. The stock market has an opinion on the subject, with Tower’s stock down 30% since the deal broke with no sign of that decline slowing. Unfortunately, that may be justified, at least in the near term.
Tower is a smallish player in the foundry space. They compete at the trailing edge of manufacturing processes. This has always been a challenging value proposition. As a result, Tower is not terribly profitable. In their most recent fiscal year, they generated gross margins of 27%, not awful, but not great. And these came at a time of peak cyclicality for the foundry industry. In the two prior years, their gross margins were 21% and 18%.
In addition to living in the unfashionable part of Moore’s Law, the competitive outlook for Tower is … not good. Tower basically competes with everyone. Their most direct comparable company is Global Foundries, a company which faces its own set of challenges, but they also face off against a number of near-peers like UMC in Taiwan and SkyWater in the US. And if we widen our view a bit, they compete with all the analog semis companies like Texas Instruments and ADI, who have internal fab capacity. Not least, Tower (in theory) competes against TSMC with all their resources and capacity. Making matters worse is that there is likely a lot more competitive capacity coming online as governments around the world chase that brass ring of having “domestic semiconductors”. And then, of course, there is China where Tower is up against SMIC and their “leading edge processes”, as well as a few hundred other fabs, adding capacity at a ruinous clip. Add that all up and it is hard to be optimistic about Tower’s competitive position.
That being said, the market is much more complex and the outlook is not quite so bleak. In that competitive rundown above, we used a fairly broad definition of Tower’s market. These companies do not really all fight over the same pool of business, and there are considerable differences among them. Put differently, this is a fairly segmented market, and so a company with the right capabilities can hope to carve out some lucrative niches. This is effectively Global Foundries’ strategy with their RF and SOI processes. And on this front, Tower has not been standing still. In the two months since the Intel deal ended, Tower has announced four interesting pieces of business. They have announced deals around IR sensors, LiDAR and silicon photonics. Taken individually, none of those are terribly exciting, but in aggregate, they show a company that is actively expanding its product set to go after new markets. Not surprisingly, there is a lot of auto exposure in that mix, but their move into silicon photonics also touches on another important semis topic of late – advanced packaging. Now, none of this is going to catapult Tower into the top of the foundry standings, but taken together they show a sound strategy as the company searches for profitable niches on which to sustain itself.
The fourth announcement laid out their relationship with Intel going forward. Admittedly, this is a bit of consolation prize. The acquisition would have been a Cadillac, this deal is a set of steak knives, but it sill holds interesting potential. As much as we questioned the deal logic for Intel, there are some areas where the two companies should partner. (And not for nothing, the fact that Intel made this announcement goes a long way to proving our point that they did not need to actually buy Tower to get what they needed.) The deal gives preferred access to trailing edge capacity for times when that will help Intel’s foundry business win deals, and for Tower it provides an important sales channel to potentially capture customers it could not otherwise reach.
While none of this is going to make shareholders betting on the Intel deal feel much better, it does paint a path forward for Tower. Working strongly in their favor is a lot of fundamental demand. The world’s semis consumption is growing steadily, and as we all saw during the pandemic, trailing edge capacity is important to the global economy. In the near term Tower is going to have to struggle as the competitive wave crests, and that will not be fun. But if anything, Tower has demonstrated its ability to survive in the nooks and corners of the industry, acquiring and turning around fab capacity forsaken by others. If they remain independent, a few years from now they will likely have a large crop of targets from which to pick. The future is not going to be easy for Tower, but there is a path forward.