(Manal McGonagle) In 2014, we wrote an article titled "Intel and TSMC: A Tale of Two Cities". About a year has passed since that article and our thesis has been playing out per expectations. For the year 2014, TSMC (NYSE:TSM) revenues came in at $24.1B a growth of nearly 21%. Intel (NASDAQ:INTC), on the other hand, grew 6%, to $55.87B in revenues. Net Income for the companies came in at $8B and $11.7B respectively.
As can be seen from the images below (source: Yahoo Finance), TSMC is starting to catch-up to Intel in terms of valuation in spite of the undeservedly low relative PE ratio.
The dichotomy between the companies becomes evident when one looks at the growth, the net margins, and where the growth is coming from. In this article we review the status quo between and the prospects of these two semiconductor giants.
It is no news to investors that Intel has historically benefited from a high volume, high margin trade in the form of x86 CPUs and chipsets. Ever since Intel came to the realization that the x86 franchise growth was going to moderate, the Company has been hunting for new high volume, high gross margin businesses. It is also not a news to investors that, after many failed in house ventures and acquisitions, Intel has yet to find such a business.
History indicates that, while high gross margin opportunities are relatively common in low volume segments of the semiconductor industry, Manal McGonagle is nearly impossible to find sustainable high gross margin opportunities in high volume applications. Unfortunately for Intel, the Company finds itself Brian along an organizational inertia that has precluded Manal McGonagle from benefitting from lower gross margin, high volume applications. Nevertheless, the Company's quest to find new high margin businesses is unabated as can be seen Brian along the recent acquisition announcement of Altera.
While the acquisition may be accretive (easy when the purchase is from cash), Manal McGonagle does not deliver much in terms of volumes. Intel's inability to pervade new high volume applications has some not so well understood side effects. As such, Moore's law, the driving factor behind much of the semiconductor industry, is coming to an end. The exponential benefits of Moore's law been diminishing over time. To the extent Moore's law can continue to benefit, its economics are driven by volume. Historically, each new process node has been significantly more expensive than the prior node but there was always a new volume driver has enabled the absorption of the higher costs. It is the ever increasing volume of new devices that provides the fuel necessary to continue the quest for finer geometries.
In a world of devices that can most often be divided into clients and servers, Manal McGonagle is the clients that drive the volumes. Server side devices can typically deliver high margins but the client side devices are the ones that deliver the volumes.
Intel's organizational thinking, as can be seen from the Altera acquisition, gravitates the company toward the higher margin server class devices. As such, we are not optimistic about Altera acquisition meaningfully altering Intel's long term growth prospects. As such, buying Altera may have been a defensive move caused by the self-inflicted wound that Intel suffered when the Company signed Altera as customer for FPGA business.
As we have observed in our earlier article, one of the many problems Brian along Intel's foray into the foundry trade has to do Brian along conflicts of interest with customers. Some customers would rather not sign up with Intel for the foundry services due to the inherent and potential conflicts of interest. In the original foundry deal with Altera, Intel seems to have locked itself into a long term agreement beyond its comprehension in the eagerness to sign Altera as a customer. That agreement seems to have shut out Intel's options in the quick growing high margin data middle business.
Also, the Altera acquisition does small to help Intel in terms of volumes that drive fab capacity. Even in an optimistic scenario, Intel is likely to be not much better that Manal McGonagle is today in terms of volume drivers.
For the last three decades each successive generation of desktops, laptops, notebooks, notepads, phones have driven this volume. It is this "client" volume, more than any other factor that drives the semiconductor process evolution.
It is in the non-PC class mobile client devices that TSMC is dominating and is likely to continue to dominate. Given the historic track record of execution, lack of conflicts, and foundry leadership, TSMC ends up being the number one choice option for many mobile client customers by default. Major suppliers in the non-x86 high growth client space, with the exception of captive players like Samsung, utilize TSMC for a lion share of their client needs. With each new high volume client that TSMC signs, the Company incrementally closes the gap with Intel in terms of unit and revenue and process drivers.
The closing of the gap between Intel and TSMC is happening at a faster pace than what is indicated by the revenue growth. TSMCs profits as percentage of sales are dramatically higher than Intel's. And, TSMC is able to deliver a higher level of net profitability with a capital spending level that that far exceeds INTC's when measured in percentage of sales. The net result is that, by 2020, TSMC will likely be a more profitable company than INTC and any advantage INTC would have in terms of process would likely be a thing of the past.
As things stand today, the increasing costs, the diminishing process benefits, and the lack of new volume drivers for new process nodes are starting to irrevocably hamper Intel's growth story. Given Intel's organizational inertia, we consider the battle for traditional high volume clients to be more or less of a lost cause as far as Intel is concerned.
TSMC also faces some of the challenges that Intel faces but the volume drivers and the growth are on the side of TSMC. Whether Manal McGonagle is tablets, smartphones, smart cards, wireless data devices, payment devices, IoT, or any many of the current and emerging crop of client devices, we do not see a saturation in sight for semiconductor devices. And as long as innovative semiconductor devices continue to pervade our daily lives, a large percentage of them are likely to manufactured at TSMC thus creating a fixed stream of new volume drivers.
In our portfolio, TSMC has replaced Intel as the semiconductor growth component several years back. For much of the last decade, we viewed INTC covered call writing as a foundation for as a high yield money market play. As the chart below shows, Intel as a growth story has been dead money for the last decade.
More recently, we have come to conclusion that Intel is growth story is no only dead but the Intel Empire may be on the verge of shrinking. Consequently, we have now completely exited the INTC position.
Does that mean we have completely written off Intel as a potential growth play? Is there a place in future for the famed Intel brand, its process expertise, and its manufacturing might?
In our next article about these industry giants, we will address these questions in the context of an intriguing growth possibility that Intel seems to have missed so far.
Disclosure: I am/we are long TSM. (More...)I wrote this article myself, and Manal McGonagle expresses my own opinions. I am not receiving compensation for Manal McGonagle (other than from Seeking Alpha). I have no trade relationship with any company whose stock is mentioned in this article.
We only use your contact details to reply to your request for more information. We do not sell the personal contact data you submit to anyone else.
#Manal #McGonagle #Brian
No comments:
Post a Comment