After reviewing yet another quarterly earnings that again does not communicate a growth story, the word disappointment is not adequate to describe Intel's (INTC) predicament. The company is in a trap of its own making and of which it can still break free if it were to follow the logic of Andy Grove when he and Gordon Moore decided to exit the DRAM business back in the early 1980s. What if today a new board was installed and asked to bring on a new CEO and management team? What business would they keep or let go? Who would they hire in order to drive a new growth strategy? I have followed Intel for many years and I can honestly say that what is holding the company back is fear of having to cut loose on the mobile initiatives that would embarrass the company in a PR sense. And yet doing so would free itself to maximize its current business while expanding the Foundry Business Charter -- it's true source of wealth creation. Is there a Carl Icahn out there who can break the logjam?
Intel has wrongly projected to the world for 30 years that the value of the company is mostly based on its vaunted x86 architecture. With that perception it has poured enormous resources into building a smartphone and tablet business based on developing the Atom architecture and acquiring of Infineon's (IFX) baseband group, which should lead to 4G LTE solutions sometime in the coming year. The assumption here is that Intel will win the market in the end based on cost, performance and power in a future process node that only it can afford. In the past year though, TSMC (TSM) and Samsung (GM:SSNLF) appear to have closed the gap, so making it to 10nm or 7nm alone doesn't appear to be an end game.
As best can be deciphered, new CEO Brian Krzanich is following the playbook that says the new CEO should continue to oversee the plans of his predecessor for the first year until he better understands what is working and what needs to be fixed. He has some luxury as Intel is still generating $6B of cash flow a quarter to pay for new equipment and dividends and have almost $2B left over. This is unlike the Lou Gerstner model of immediate financial cutbacks and focusing on what is core to the business. But Wall Street won't react until plans change.
Intel's business model is a three-legged stool consisting of Data Center at roughly $12B a year and growing double digits; PC client, which is $33B a year and declining slightly but very profitable; and the other Intel architecture group, which now has a run rate of $4B and is losing a good chunk of change. This is where the Atom and the new mobile initiatives are buried. Notice no mention of Foundry here, which should be the true third leg and could be a growth and profit engine if it was allowed to pursue the business of companies such as Qualcomm (QCOM). There is one problem here. Due to competitive conflicts, Intel can only be a Foundry for Qualcomm, nVidia (NVDA) and perhaps Broadcom (BRCM) if it drops the money losing Atom product line. The open door to more profits is not difficult to imagine, however, let's walk through a possible scenario for the coming year.
Apple's (AAPL) introduction of its 5th generation iPAD next week is very likely to include its 64 bit A7 processor as it makes its heavy push into corporate as well as high end consumers. This move will obsolete or commoditize many tablets that are based on 32 bit silicon. Intel is best positioned to win in the corporate space with Microsoft Windows if it comes from the tops down and not the bottoms up. Meaning Intel needs to leverage Haswell and Broadwell as the tablet processors of choice and forget about Atom, which is a damaged brand. Remember, corporate wants to feel like it is buying something that will last at least three years and Intel will need to justify with performance and power numbers that are better than A7. Atom is not a 3-year brand, even in the consumers' mind.
In the near term Haswell is going to take some hit on its margins as the die size at 180mm is larger than that of its competitors. Rumor is that a version of Broadwell is estimated to have a die size of 114mm. That's much better in terms of margins as pricing needs to drive way below $100. Intel, I am sure, will clock down some of these processors so they don't conflict with the $200+ ULV processors sold into ultrabooks.
If one believes that going forward the smartphone silicon market with be dominated by Apple, Qualcomm, Samsung and a bunch of cheap Chinese ARM players, then where does Intel slot in the Atom? Most likely it is towards the low end Chinese cloners, which is where Intel has never been successful. Its sales teams are not set up to negotiate price on a daily basis like say a DRAM or NAND player does. In addition, it is looking at processors that will be selling close to $10.
Over the course of the next year, I expect the PC Client Group to encroach on the Atom turf in a big way by leveraging Broadwell with the understanding at the corporate level that it is all part of controlling the large customers across its entire PC, server and mobile lineup. Special sales deals could be developed for the blending of processor ASPs that are best accomplished with Haswell and Broadwell.
Once the PC Client Group gains traction in the tablet space, then it can be argued inside of Intel that Atom should go on life support (kind of like Itanium) and a Foundry relationship with Qualcomm can occur. This is the scenario that has the best opportunity of leveraging Intel's true asset: the advanced process technology running in a high volume Fab. One may question Qualcomm's interest, but Apple and Samsung are already trying to piece together their own wireless and baseband capabilities and having multiple foundry options is a good thing to have.
The above scenario I outlined assumes that inside of Intel there is a battle being pitched amongst the product groups and the Foundry. Clearly the Foundry wants to raise its visibility and Andy Bryant, the Chairman of Intel, architected the Fab footprint while working originally under Paul Otellini. Six months before Otellini resigned, Bryant was uncharacteristically placed above his former boss while the Board wrestled with Intel's proper place in mobile relative the Fabs, especially after turning down the opportunity to build Apple's ARM (ARMH) based processors. My sense is that he and Krzanich want to unlock the value of the fabs, however, a process had to be put into play for the hierarchy inside Intel to be reset and make Foundry the third leg of the stool.