Thursday Intel (INTC) released earnings for their 2013 fourth quarter and full year.
Briefly, The Company beat slightly on the top line and was a penny short on the bottom line.
The penny shortfall in earning is explained by a 1% higher provision for income tax.
So, the bottom line is that the quarter was right on to forecast. I find that pretty amazing for a company the size and complexity of Intel to be able to forecast their business with this level of accuracy.
No need to re-plow the numbers here, the published release does a nice job of that.
I do have some observations that were interesting to me:
1. The high end i7 and i5 had record unit shipments in the fourth quarter. Record shipment of these devices is hard to reconcile with a shrinking PC market.
2. Desktops grew 11%. Desktop Haswells led the availability of low power mobile Haswells by something like six months (except for the MacBook Air), it seems reasonable that Ultrabooks will show this kind of growth in the 1st and 2nd quarters.
3. The software business went profitable on an operating basis. I see this as a watershed event. Software is one of the promising areas for non-silicon growth in the future. Management never mentioned this important little detail and the analysts didn't ask about it.
Then, of course, my favorite topic came up in the Q&A: Overcapacity. Mark Lipacis of Jefferies backed CEO Brian Krzanich into a corner about the empty Fab 42 in Chandler AZ. Krzanich's answer could be summarized as Fab 42 was a multi-billion dollar mistake as a result of a faulty 2010 forecast on future demand for PCs and associated products.
I'm not buying it.
A mistake like this from a highly sophisticated $53 billion company that can forecast quarterly sales and earnings within a percent or two is hard to swallow. Anyone who understands the reach of Intel would know that the CIA would like to have access to the worldwide economic and financial data that this company has.
Anyone who has been involved in the semiconductor business, or any business for, that matter, would know that Intel does a ground up forecast monthly on design wins and the associated sales for those designs. Speaking from personal experience, I can tell you that these exercises are grueling affairs in which fairly minor mistakes are career limiting events.
To miss the measure of future demand by an amount that would lead to a $25 billion mistaken capital over-expenditure over four years is simply ridiculous.
Management is often asked about this enormous cap ex and what it is for and when it will contribute to shareholder value. In the early stages, management used to answer these questions with brags about, "the world's largest cranes" and the incredible complexity of the projects.
Intel base cap ex has always been about $5 billion per year. Over the past four years the company has spent that $20 and another $25 billion to put these gargantuan fabs in place in Oregon and Arizona.
I would submit that Intel management, in aggregate, is quite intelligent. As Paul Otellini once said when asked about these projects, "We're not stupid."
If you buy the "mistake" explanation, then Intel management is truly stupid beyond belief and you should sell your shares and run like a "thief in the night."
As for me, I think there is a highly defined purpose for these fabs and I think the purpose has to do with non-volatile memory and Solid State Drives, only because that potential market for processed silicon is large enough to fill these huge fabs.
I could be wrong. Maybe Intel management is that dumb. Maybe the fabs are for something other than memory. Maybe the fabs ARE a mistake.
If I am right, the last thing you would expect is management to telegraph their purposes and allow the competition to mobilize. You would want surprise. Demoralizing surprise. Even "shock and awe".
So, the quarter was right on and accurate. Guidance was right on, if you didn't want to tip your hand about a major new initiative.
Intel management is either dumb as a box of rocks or smart like a fox.
I think the latter.