It comes as very little surprise to us here at BSN* that Intel has officially opened up the company’s fabs to whomever wishes to make their processors using Intel’s leading-edge processes. The company has for many years been moving towards a point where the cost of building new fabs using leading edge processes simply would become impossible to sustain on their chips alone. The nail in the coffin for this rumor that has been circulating for countless years was the ascension of Brian Krzanich to CEO. Brian’s role in the company prior to CEO was COO and prior to that he was responsible for managing the company’s supply chain and manufacturing. The guy worked his way up from working in the fabs to become CEO, so you know that had to play a huge part in the board’s decision to pick him. It was merely a matter of time until Intel became a direct competitor to TSMC and Globalfoundries.
Now, today marks a fairly big different in the way Intel says that they are going to be run as a company. For a whole host of reason, many of those reasons were outlined by Krzanich himself during his part of the analyst day, which I unfortunately missed. However, you can re-watch any part of this morning’s festivities on Intel’s own investor relations website right here.
One of the important factors in today’s perspectives was the fact that the PC market has stabilized to a point where Intel can focus on other businesses while still maintaining profitability in the PC market. While many have claimed doom and gloom on the PC market, Intel has proven that they haven’t seen nearly as much volatility in the market as some of their competitors have. A stable PC market gives Intel an opportunity to expand on a more stable base without dismaying investors.
As you can tell from the first slide and the slide below, Intel has finally let the cat out of the bag. That cat is the fact that Intel will take on more fab customers and those customers can theoretically be anyone willing to fab on Intel’s manufacturing process. Brian said that Intel will, "Open the foundry to any company capable to utilize our leading edge silicon."
Anyone aware of Intel’s economies of scale and the way that they develop their manufacturing processes understands that they will ultimately be more successful and profitable with more of their capacity being utilized. Obviously more utilization is good, but Intel’s problem is that they also have to be able to pay for new fabs and new processes like 10nm, which is expected to be ready in 2015.
Intel’s CEO Brian Krzanich also talked about embracing where the market is going and to have a more pragmatic approach to creating products and a much less insular company. This will be fundamental change in the way that the company looks at their products and the markets that they address.
As such, Intel really wanted to focus on their mobile push and how they are going to build on what they already told us about at IDF. Intel already has a pretty competitive SoC in Baytrail for tablets and they already have a shipping LTE modem, albeit in only one device and with very limited capabilities beyond Cat 4 LTE speeds. However, simply being an alternative to Qualcomm may prove useful for Intel as Qualcomm maintains somewhat of a stranglehold over the majority of the LTE modems shipping in high-end smartphones.
First, Intel’s Hermann Eul, Vice President GM, Mobile and Communications Group, wanted to talk about the fact that Intel already has shipping 64-Bit silicon and that they are already here 64-bit and that we can expect even more in 2014.
Following that, he wanted to focus on the improvements that going 64-bit had brought to the Intel architecture and how much of an improvement users could easily see from simply switching from a 32-bit SoC to a 64-bit one. Admittedly, the examples here are very media creation focused but it would be interesting to see them show more figures about daily usage and real-world applications. Sure they used TouchXPRT for the tests, but I believe that they could have used consumer-facing applications to do these measurements.
Following that, Hermann went on to talk about Intel’s LTE and the fact that it supports multi-mode data and voice and claims 15 LTE bands and global mobility. However, we still do not know how many simultaneous LTE bands a single device can support. He also talked about the fact that the 7160 is LTE capable and that they are merely waiting for their partners and carriers to implement it once the networks are ready.
He then moved on to focus on Intel’s SoC and mobile data roadmap for 2014 with Merrifield and LTE Advanced being brought in the beginning of the year with the 7160 and followed in the latter half of the year with Moorefield and Cherry Trail. There will also be the introduction of the SoFIA SoC with integrated 3G connectivity and will actually be fabbed at an external foundry, something a bit uncommon at Intel, but nothing new.
In addition to the 2014 roadmap, Hermann talked about the convergence of Intel’s Smartphone and Tablet SoCs (as seen above) in 2015. Intel expects to simply have a performance/mainstream offering and a value/entry offering for both their smartphones and tablets. Resulting in saved time and money in terms of engineering and development and shorter design cycles since two different products aren’t being made.
In the 2015 expectations for the company he notes that it will have the new Goldmont Atom CPU architecture as opposed to Moorefield and Cherry Trail’s Airmont CPU architecture. And with the entry SoC, SoFIA LTE, Intel adds integrated LTE into their SoCs for the very first time, finally doing what Qualcomm has been the only company able to do. Do note that once again, Intel will be manufacturing their value and entry level chips in an external foundry, which means that we most li
kely won’t see them fab those entry-level chips in their own fabs until 2016.
Intel also talked about the datacenter and wearables, but the truth is that the focus was about change in how the company sees itself, its products, its fabs and its customers. A more pragmatic and market-focused Intel is a good thing for their investors and finally opening up the fabs finally means that Intel investors have to worry less about Intel having empty fabs and long term commitments to building new fabs. But, this does mean that Intel will have to talk about their fabs more than ever before and that they will most likely have to start fighting for customers and become the alternative to TSMC or Globalfoundaries, which means lots of interesting negotiations. That rumor about Intel manufacturing Apple’s chips no longer sounds so crazy anymore, does it?