On Intel's (NASDAQ:INTC) most recent conference call, both the CEO and the CFO made some curious remarks with respect to the company's next-generation mobile product lineup. More to the point, both executives made some extremely confusing claims with respect to its upcoming high-end mobile system-on-a-chip, code-named Broxton. Let's explore those claims and some possible explanations.
Will Broxton have an integrated modem?
On the most recent call, Intel CEO Brian Krzanich said:
[A]fter that is ... Broxton, which is our internal 14-nanometer product. That's targeted toward the mid- to high level. And as we bring that into the second half of 2015 and into 2016, there will be various levels of integration on that.
At first, this makes a lot of sense, given that at the most recent Intel investor meeting, management had indicated that Broxton would be more integrated than its predecessors, Moorefield and Cherry Trail for smartphones and tablets, respectively. Further, Broxton is Intel's first system-on-a-chip that really allows the mixing and matching of internal and external IP blocks at a fairly rapid clip, so it stands to reason that Intel will have an entire family of Broxton products with "various levels of integration."
That said, if it integrates a modem, why do we even bother with SoFIA?
In a later statement on that same call, Krzanich said:
We have a lot going on -- the ramp of Broadwell, the ramp of Skylake in the second half of next year, plus bringing these products inside. But I'm very confident that when you do that, plus you add in Broxton, which is targeted toward the mid- to high range, and again is integrated with leading-edge LTE.
At this point it seems pretty clear that Krzanich is outright claiming that Broxton will be "integrated with leading-edge LTE." Now, what's interesting is that at the investor meeting, Intel stated that Broxton wasn't "integrated enough" for the lower end of the market and that this necessitated the existence of SoFIA. Further, in the slide that Intel presented about Broxton, it made no mention of an integrated baseband but made sure that on the SoFIA portion of the slide, integration was explicit.
So, we have a number of options:
- Broxton doesn't have an integrated modem and the CEO simply meant that the entire platform will have LTE.
- There will be a later Broxton variant beyond the initial one with integrated LTE for the high end of the smartphone market, in addition to one without the baseband for Wi-Fi tablets.
- Broxton will have integrated LTE from the get-go.
But which of these options is most likely?
Let's turn our attention to the slide that CFO Stacy Smith presented at the investor meeting:
This image suggests that with Broxton -- Intel's hero-device-worthy system-on-a-chip -- built on Intel's 14-nanometer process, will weigh in at about 60 square millimeters in size, smaller than even today's Merrifield, which has only two CPU cores, a much smaller graphics processor, and less integration on 22 nanometers. Not only that, but Intel plans for a significant boost in graphics, CPU, and imaging performance within that area. Are we supposed to believe that Intel will be able to throw a full-blown LTE-Advanced modem in that area, too?
Broxton is Intel's chance ...
There is no doubt that Intel has the world's best semiconductor manufacturing technology, but this does the company absolutely no good if it doesn't have the right products at the right time. If Intel had gotten out its 22-nanometer Merrifield/Moorefield system-on-a-chip products about a year ago and didn't bungle the launch of its XMM 7160 modem, these could very well have turned the tides of war for the company in mobile and taken decisive performance and watt leadership.
Unfortunately for Intel and its investors, Qualcomm (NASDAQ:QCOM) has simply executed much better and gotten products out that OEMs were happy to deploy even on "inferior" 28-nanometer planar technology. Manufacturing technology can give you an edge, but only if you use it properly. Intel's product planning teams really missed the mark here, putting all of that 22-nanometer FinFET goodness to waste. Broxton is Intel's chance to change all of that.
Foolish bottom line
Intel has plenty of assets at its disposal to "win" this game, but these have not been effectively utilized. Intel needs to prove that it can utilize these assets effectively to grow its semiconductor revenues meaningfully. If it can't, then we'll see Qualcomm and MediaTek do to Intel what Intel did to the former RISC server and workstation chip vendors -- relegate them to obscurity. Except, unlike those players of yore, Intel's core business has been and will always be chips.