Despite its sparkling balance sheet and dominant market position, Intel is nagged by its late arrival in the smartphone market.
Andy Grove, Intel's co-founder and former chairman, is famously quoted as saying "Only the paranoid survive." If that's true, then the analysts at Intel's investor day on Tuesday have a long and healthy life ahead of them.
That's because at virtually every point in the day where questions were allowed, they all came back to wondering how Intel would wrest away the smartphone and tablet markets from new archrival ARM Holdings, whose chips are at the core of Android phones and tablets, and with some Apple customization, iPhones and iPads.
If Intel execs were looking to avoid these questions, they were their own worst enemy, because rather than give one answer for how they planned to beat ARM, they gave at least three. First, Intel will eventually produce Atom processors on its newest fabrication process, allowing for substantially lower power use, multicore designs, faster graphics, and more stuff integrated onto the chip, including universal connectivity and security. Second, Intel chief Paul Otellini downplayed the importance of Intel chips being in every phone and tablet, as those are low-margin businesses. He offered a pop quiz on which chipmaker earns the most money in the mobile device sector. Of course, the answer was Intel--after all, all those low-margin phones without Intel chips must talk to servers somewhere, and most of those servers run high-margin Intel CPUs. Finally, various speakers pointed out that Intel's unique array of software offerings makes it easy for Android phone makers in particular to port their apps to Atom phones running Android.
There's merit to all these answers, and yet they don't quite satisfy, perhaps because Intel never really said: "We're going to be the leader in smartphones in XX months, and here's how we're going to do it." The main reason not to make such a statement is that you don't necessarily know that you can do it. And so the conservative Intel leadership, with Andy Grove's words ringing in their ears, just wouldn't go so far as to describe how they'd crush ARM.
While investors didn't get to hear exactly those words, there was plenty to get excited about.
Intel remains a cash machine. Not only does it have enough money ($12 billion in cash on hand) to continue its constant improvement in chip performance, it has also consistently increased its dividend payment and has been buying back its stock as a way to increase stockholder equity. What investor wouldn't love that? Intel waited until it saw the end of the recession and then started buying key assets to move the company into new markets and cement its place in existing markets.
Last August Intel agreed to pay $1.4 billion for Infineon's wireless unit, acquiring a more holistic smartphone platform. That deal ties in well with Intel's universal connectivity strategy, which includes everything from Bluetooth to 3G to LTE to Ethernet to Thunderbolt (a very fast USB-like standard, co-developed with Apple, that offers speeds of 10 Gbps).
Just weeks before the Infineon deal, Intel had agreed to buy security vendor McAfee for $7.7 billion, in a move that left many industry watchers scratching their heads. But McAfee should provide key technology to back up Intel's promise to provide energy-efficient systems with universal connectivity and security baked in. Meanwhile, Intel remains focused on keeping costs low and profit margins high, predicting on Tuesday that it will increase its margins for 2011 to more than 60%.
Yet even with all that financial goodness, there's still the nagging question of whether the company has lost the mobile device market to ARM, and, if it has, what that means for the long term. Intel's answer is framed by how it looks at its overall business. It breaks down revenue in four ways. PCs account for $30.3 billion in revenue. Data centers, which now include a good bit more than server chips, account for $8.8 billion. Software and services accounts for $2 billion, and "other" totals $3.1 billion. The last category includes Infineon, its Atom-based system on a chip (SOC) for netbooks and other embedded devices, and its NAND flash memory capability, mostly for solid-state drives.
When Intel announced its 3-D transistors May 4, it said Xeon would be the first to benefit. If you were wondering why Intel would use its new 3-D transistor technology first for Xeon chips rather than for Atom processors, the numbers tell the story. Its Atom sales are tiny by comparison; the cash cow goes first. The Atom will follow the Xeon in Intel's new 22-nm processor, and before that there are plans for the Atom-based chip known as Medfield to hit the market in 32-nm form--first a single-processor SOC, then dual, and probably quad cores after that.
Server Market SplitsvilleJust because the server market's in the doldrums doesn't mean innovation has ceased. Far from it -- server technology is enjoying the biggest renaissance since the dawn of x86 systems. But the primary driver is now service providers, not enterprises.