Investors Are Dead Wrong About Intel: Here's Why

At its earnings conference call last week, the focus was on quarterly estimates, which Intel INTC beat as usual with a little help from a stronger mix of high-ASP devices and a lower tax rate. The top line guidance wasn’t too bad considering the ongoing transition in the business and the bottom line guidance was also okay after taking out one-time items.

The computing pressures and weakness in mobile devices were as expected. Some analysts commented on the data center weakness, which didn’t however explain the plunging share price because revenues were in line with its revised expectations.

What's The Likely Problem Then?

My sense is that investors are concerned about competitive pressures in mobile, IoT and data center. They are also uncertain about Intel maintaining its process lead and fear the end of Moore’s Law, which says that the number of transistors per square inch in integrated circuits doubles every two years.

Since all these issues are tied together, let’s take a step-by-step approach to understand the situation.

First we need to perhaps refresh our memory as to the benefits of Moore’s Law. When more compute power is captured in less area, there are advantages for sellers/manufacturers, buyers/OEMs and end users.

For manufacturers, the overall cost comes down because a larger number of chips can be fabricated in a single batch with the benefit further increasing when wafer sizes increase.  Lower cost in turn makes the company more profitable and also enables it to find broader application for its chips.

The buyer/OEM is also interested in this cost shrink because it’s understood that the manufacturer will pass on some of the profit to it in the hopes of volume business.

The end user is benefited because the increased compute power of chips improves the performance of end devices like PCs while enabling the design of sleeker, lighter mobile devices. Power-efficiency of devices also improves.

Now, since we are talking about Intel, let’s jump to the manufacturing side.

Pressures of Leading Edge Chip Making

 A McKinsey report from 2013 perfectly sums up the cost angle: “shrinking nodes from 32nm to 22nm on 300mm wafers causes fabrication costs to rise 40%, process development costs by 45% with chip design costs up 50%.” We are now talking about a couple of shrinks ahead of this, so you can just see how costs must be accelerating.

If the purpose of Moore’s Law is to reduce cost while increasing functionality, chipmakers are in something of a dilemma because the benefits of cramming compute power onto a chip is increasingly lost in the rising cost of manufacturing it.

That’s not the end of the problem. The report further says that “there could be a situation where cost improvements end but performance enhancements continue,” meaning that companies would have to pay more for increased compute power. Naturally, this will limit their usage to only high-end applications in the data center. Here too, the constantly rising cost of hardware will lead companies to look for alternatives, ultimately leading to the death of Moore’s Law anyway.

The McKinsey report goes as far as to say that “the industry’s ability to capture value would be at risk because of the disruption of demand.”

The Intel Solution

Intel is not moving away from Moore’s Law. Far from it. But it is extending its tick-tock cadence a bit to glean a little more out of its equipment.

The market has been bemoaning the fact that Intel lowered 2016 capex to $9.5 billion from its previously estimated $10 billion, reading weaker demand and a loss of process lead into these numbers. The fact that Taiwan Semiconductor TSM management stated on its earnings call that it would ramp up 10nm production this year and grow volumes in 2017 added fuel to this fire because Intel isn’t expected to ramp up 10nm production until the second half of next year.

Another tidbit was that 7nm is still in the distant past because Intel was planning on extending 10nm to 2020, allowing competitors to get even further ahead of it. So Intel’s future started looking pretty bleak.

But there’s something investors probably overlooked. Intel’s depreciation is estimated to drop $1.3 billion in 2016 despite the fact that its capex estimate as it stands now will increase by over $2 billion. So Intel is doing some pretty smart cost management right here.

One may point out that cost-cutting alone is not the solution because if Intel doesn’t remain in the process lead, its chips won’t deliver the best performance, allowing competitors to eat into its high-end customer base. But this is where Altera will likely play a big role.

Intel was heavily criticized for paying $16.7 billion to buy out Altera and most analysts said it overpaid. But Altera will enable hardware acceleration that will likely boost the performance of Intel silicon even before it moves to the next node. Intel had previous working experience with Altera (the two collaborated in hardware acceleration and Intel was already in a foundry relationship with it).

As Intel has said repeatedly, Altera will help in both its data center and IoT businesses. In the data center, this strategy will very likely enable Intel to deliver competitive performance and greater flexibility at lower cost thus protecting its market share. While in IoT, the strategy can help it take major share.

Intel has already said that Kaby Lake, its next processor architecture while remaining a 14nm chip would bring “key performance enhancements” to Sky Lake.

As far as 7nm is concerned, there is reason to believe that Intel took delivery of some EUV equipment from ASML Holding ASML late last year. Management has said in the past that while EUV will greatly help the 7nm shrink, Intel was working on a way to move ahead without it if it was further delayed. Intel has a 20% stake in ASML, compared to much smaller holdings by TSM and Samsung. There could also be material innovation at 7nm, but Intel isn’t talking about it now.

Final Words

Intel appears to be doing the right thing for investors as far as manufacturing strategy is concerned.

The fact that both 3D NAND and 3D XPoint are entering production this year is also positive.

But this doesn’t mean we should take the China market softness lightly. Intel appears to have adjusted expectations for 2016 already while maintaining its cautious stance on China. The China concerns were echoed by other semiconductor companies like TSM and Advanced Micro Devices AMD that have reported December-quarter results.

Intel shares carry a Zacks Rank #3 (Hold).

 

Want the latest recommendations from Zacks Investment Research? Today, you can download 7 Best Stocks for the Next 30 Days. Click to get this free report >>

Want the latest recommendations from Zacks Investment Research? Today, you can download 7 Best Stocks for the Next 30 Days. Click to get this free report
 
INTEL CORP (INTC): Free Stock Analysis Report
 
ADV MICRO DEV (AMD): Free Stock Analysis Report
 
ASML HOLDING NV (ASML): Free Stock Analysis Report
 
TAIWAN SEMI-ADR (TSM): Free Stock Analysis Report
 
To read this article on Zacks.com click here.
 
Zacks Investment Research