While I was pretty sure that there wouldn't be anything too exciting to write about when it came to Intel (INTC) until the earnings report on October 16th, it turns out that the launch of the latest Intel Atom Z2760 - a complete system on chip designed for tablets - is actually a big deal from an investment perspective.
The Atom Z2760 - A Competitive Entry
The main issue that investors have had with Intel over the last couple of years is that the company has not been a particular effective competitor in the low power space. Earlier this year, Intel threw its hat in the smartphone ring with its "Medfield" system-on-chip. It is a respectable product, able to compete quite nicely with the best that Qualcomm (QCOM) - the current leader in the ARM camp - has to offer in both performance and battery life. Intel still needs to bring out its dual core Atom Z2580 with LTE support if it is going to compete in the USA, but the first step seems to have been strong.
The next piece of the low power puzzle is the tablet space, and that's what the Atom Z2760 is for. According to numbers published by Intel (which should, of course, be verified with independent results when the devices launch), the dual core Atom Z2760 is faster in the popular SPECint benchmark than the Nvidia (NVDA) Tegra 3, Qualcomm Snapdragon S4, and the standard ARM Cortex A9 design clocked up to 1.8GHz. Further, Intel claims that power consumption is significantly better than the latest generation iPad, slightly worse than the iPad 2, and slightly better than the ASUS Transformer Pad Infinity.
Intel's competitive on the performance and power consumption side, and things will only get better with the next generation "ValleyView" system on chip - which should bring a brand new architecture and 22nm.
But now that the technological piece of the puzzle seems to be more or less under control, there are some further observations to be made.
This Atom's A Cheap Chip - Margins, Anybody?
The most important part of the Anandtech article wasn't that the Atom will be performance/watt competitive - as I've noted for months now, Intel has the technological chops to compete here - but rather the price of the chip itself. The article pointed out the following:
"The important takeaway is that Intel is significantly reducing the price of the Atom Z2760 due to competitive pressure from ARM. Most ARM smartphone SoCs seem to be priced in the $15 - $30 range, and I'd expect the Z2760 to fall somewhere in that range. Intel has shipped cheap CPUs in the past, but I don't know that they've ever shipped something this cheap"
Unfortunately for Intel and its shareholders, margins on these smartphone chips will not be anywhere near as attractive as those in the PC space. Further, there is fairly strong competition in the tablet system-on-chip business. Nvidia and Qualcomm are the two strongest competitors in the space, especially now that Texas Instruments (TXN) has decided to call it quits.
Now, Intel does have an advantage here. If the company can maintain its fabrication lead over the dedicated foundries, and if it can keep its fabs working at near-full capacity, then Intel will have a significant gross margin advantage over competitors that have to both pay a licensing fee to ARM as well as pay Taiwan Semiconductor (TSM) to fab the chips. This gives Intel pricing power leverage over the competition.
However, if tablets powered by these Atom chips become Intel's "bread and butter" and supplant the higher end "Core" series of chips, then Intel's gross margin profile will come down. It is important to stress that despite lower overall gross margins going forward, it is likely that Intel's will be better than its competitors' due the aforementioned advantages.
Quo Vadis, Revenue?
In addition to the margin decrease due to stronger competition, another big problem will be sales. For every notebook chip that gets replaced by a $15 - $30 tablet chip, Intel loses a significant chunk of revenue.
Note, however, that those likely to replace their notebooks and desktops with lower end tablets were probably those that bought lower end, lower cost systems to begin with.
Further, in the tablet space, it seems quite clear that Intel isn't just gunning for the low end media consumption tablets, but for higher end "Core" based tablets. These will use the high performance, high ASP, and high margin chips found in the traditional notebook space (albeit specially binned and clocked for low power consumption). The "Surface Pro" from Microsoft (MSFT) is an example of such a tablet.
A Potential Bright Spot - The Era Of Disposable Computing
While ASPs for tablet chips will likely be down relative to the traditional PC chips, it's important to note that we are now entering the era of what I like to call disposable computing. What this essentially means is that since devices are being sold in lower and lower thermal envelopes, and since these devices are more tightly integrated, there is much more incentive to replace them at a higher rate.
Think back with the desktop. These are upgradeable, easily serviceable, and could be made to last for a very, very long time given the right knowledge (or a local computer shop with helpful staff). But as computing moved to laptops, less of the machine was upgradeable, so whenever it broke, became too slow, or simply didn't have the neat features of newer models, there was real incentive to replace it.
Apple (AAPL) is actually pretty smart here. The latest Retina MacBook pro has the NAND flash storage and RAM soldered right onto the board. Not only does this force users to pay whatever exorbitant OEM-tax on more RAM, bigger storage, and so forth, it makes the machine completely unserviceable. If the RAM dies and the machine is out of warranty, tough luck! Go buy a new machine.
On the tablet side, there is an even more attractive golden ticket for OEMs - performance. The Atom Z2760, Tegra 3, and Snapdragon S4 system-on-chip solutions are still extremely slow compared to even a low end AMD A4 or Intel i3, so there will be very tangible benefits to upgrading tablets more frequently as new models come out in the medium term. So while average selling prices will decrease, volumes will go up, offsetting the ASP decline.
The hard truth is that Intel's >60% margins are not safe, but it will still be able to outperform its peers on the gross margin line due its manufacturing advantage and its royalty-free designs.
The main challenge, of course, is to win the race by having the better mouse trap. Qualcomm and Nvidia are formidable opponents, and it will take a lot from Intel consistently produce noticeably better chips. The competition is not standing still, so going full-throttle on the manufacturing side as well as the design side is absolutely critical to the company's success.
Finally, Intel's other CPU businesses - PCs and servers - are still quite high margin and while the "death of the PC" is the fashionable thing to parrot right now, it's not clear that low end tablets will actually supplant these more powerful compute devices as much as they will supplement them.