- The x86 business is a 76% gross margin business.
- The foundries are showing signs of "spending fatigue".
- Intel's latent earnings are truly amazing.
Intel (NASDAQ:INTC) is the world's largest semiconductor producer as measured by revenue. Intel is the leading technology semiconductor company as measured by the node (size of transistor) that is in volume production. Intel is also highly profitable, with 65% overall gross margins and 20% net margin. This kind of performance is turned in year after year, even including the" boat anchor", the past few years, of the multi-billion dollar losing mobile and "other" divisions.
There is no shortage of articles focusing on the mobile "disaster", but very few that try to isolate the profitable divisions.
Find a comfortable chair, because we are going to take a look at the 80% full glass rather than the 20% empty glass.
I need to coin a new term for this analysis. Make it "latent profitability".
Definition of latent: present or potential, but not evident or active
We can focus on the latent profitability of the Intel x86 business by eliminating the losing operations and making the appropriate adjustments to sales and gross profit.
The standalone x86 Intel would look like this:
|Cost of Sales Bil$||12.21|
|Gross Profit Bil$||38.65|
|Gross Profit percent||76%|
With 45% of gross profit falling to the bottom line, Intel would earn $17.39 billion, or $3.48 per share.
And if that's all Intel were, and the $3.48 fetched a 16X PE multiple, the stock would be trading at $55.68! Management must see real promise in the currently money-losing segments, or they could have almost a double by paring the company back to its profitable core.
These numbers are the first cut of the "latent" variety. They still include depreciation. The high level of depreciation is caused by very high capital expenditures in order to remain 1-2 nodes ahead of the next best-in-class competitor. Intel has averaged over $11 billion in CapEx for the past four years.
This level of CapEx has caused a like response from Intel competitors. The foundries, TSMC (NYSE:TSM), GlobalFoundries, et al., that support Intel competitors must spend the equivalent of Intel in order to not fall further behind on manufacturing technology.
The difference between Intel and the foundries is that Intel has the x86 business generating $38 billion of cash that can be spent to stay ahead of the competing foundries. The Dreadnought called x86 allows Intel to fire an $11 billion round of CapEx every year… and to continue firing every year for the foreseeable future.
The foundries, on the other hand, are beginning to show cracks in their ability to continue this tit for tat "cold war" of capital spending.
This article gives the reader some feel for how difficult the semiconductor business has become.
This article, page two indicates, from equipment management, that the foundries have not yet placed the orders that are required for the special machines needed to do finfet technology. If the equipment orders are not even placed yet, we can kiss off 2015 delivery, regardless of what the PowerPoint presentations say.
This article indicates Samsung and GlobalFoundries will cooperate to develop finfet technology and have actually received orders from Qualcomm (NASDAQ:QCOM) and Apple (NASDAQ:AAPL). This leads to the recurring thought of, "If either of these guys had a line on finfet, why in the world would they give this valuable technology to the other?" The answer, of course, is that they don't have it. The next thought is, "Why would Qualcomm and Apple not be giving these orders to the Intel "fast follower", TSMC?"
Perhaps this article answers both of the above questions. Morris Chang, the consummate Honest Man, is telling us that TSMC will lose business to Samsung/GloFo in 2015, but will regain leadership in 2016-2017. In the semiconductor process technology world, a missed node is virtually impossible to make up… ever. I have a difficult time suppressing the thought, "Is this the beginning of the foundry capitulation on the technology "Arms Race?" I think so.
If TSMC is passing for two years on finfet, it actually makes some convoluted sense for Qualcomm and Apple to give orders to Samsung/GloFo. It smacks of a desperate hope that these two will come through on finfet, and if they don't come through and TSMC has passed, Qualcomm and Apple (mainly Qualcomm) are in deep yogurt for bleeding-edge semiconductor technology anyway.
Now, back to this latent profitability thing with Intel, since the huge CapEx by Intel has continued for four years and the Intel equipment depreciation schedule is two to four years, Intel is now writing off about $11 billion per year between depreciation and expensed in R&D. OK, what would Intel look like if depreciation suddenly went to zero (a ridiculous assumption, but bear with me here).
|Cost of Sales Bil$||4.71|
|Gross Profit Bil$||46.15|
|Gross Profit percent||90.7%|
So, what I'm saying here is that the x86 business, at today's level of sales, without the losing operations and without the spending induced depreciation, is a 90% business. Obviously, there will always be some level of spending and depreciation, but this is just a thought experiment on latent profitability.
If 45% of the new gross profit dropped to the bottom line, we would have $21.22 billion, or $4.24 per share.
Now I will show you, in print, just how crazy I can get.
If Qualcomm can't get leading-edge semiconductor technology, Intel will win its chip business by default. Apple, of course, will have no choice but to give Intel its foundry business. If Samsung doesn't make it to finfet, it would also have to buy from Intel or be in a non-competitive position.
An application processor chip that included DRAM and LTE baseband should sell for $40 each, and the world will use a billion of them. Imagine Intel controlling the $40 billion mobile business, as well as the $50 billion x86 business. I'm going to further assume that the massive cash flow Intel will have will make it possible to buy the shares down to three billion shares.
This is truly insane:
|Cost of Sales Bil$ (incl $4 bil of depreciation||13|
|Gross Profit Bil$||77|
|Gross Profit %||85.5%|
Now, if 45% of the $77 billion drops to the bottom line, we have $34.65 billion over 3 billion shares, or $11.55 per share.
I think that is exactly where Intel is heading. I also think that it will take 7-10 years to pull this off.
I have either lost all my marbles, or I have just glimpsed the investment Promised Land. I know this is heresy, but I see no need to diversify, Intel should do the job for the next several years. I don't even see the need to buy the shares, I will simply buy the highest strike price LEAPs as they are announced each year.
David Einhorn, Seth Klarmen, and Warren Buffett will see the long-term and inevitable nature of this investment and jump on. Warren, you don't need to understand technology to see the financial potential of Intel.
This is where we small players have an even shot with the hedge funds and other institutions. It is hard for the big guys to get a billion-dollar bet put down on options; they have to buy the shares. We small guys can buy as many LEAP calls as prudence will allow and make many times the percentage gain that the hedge funds can make. Everybody wins on this remarkable company and once-in-a-lifetime investment opportunity.
There is no such thing as a "Sure Thing", but you can see a sure thing from here.