- IDF 2016 was packed full of information, but very little was pertinent to the investor base.
- Intel confirmed the release schedule of key product categories and announced promising technologies.
- However, the impact of these new developments won’t be felt to any meaningful extent in the current fiscal year.
Intel Developer Forum 2016 didn’t do a whole lot to drive Intel Corporation (NASDAQ:INTC) stock price this year, which has been the trend at developer conferences in much of the technology space. However, I wanted to highlight some of the key announcements coming out of the event taking a different angel from other news/investment sites.
Day 1 highlights: VR, Kaby lake, and Autonomous Silicon
In summary, the day one event was mostly driven by new device categories, more notably VR and AR implementation. Intel announced Project Alloy, which incorporates VR into a mobile form factor. This allows users to experience VR in an environment of 6-degrees of freedom (basically you can walk around with the VR headset without being connected to a PC or knocking into stuff). The VR announcement included the announcement of a Windows 10 update referred to as Holographic Shell. The new update will hit next year, as it will create a new OEM (HTC and Oculus) standard that helps to unify the various HMD makers under a Windows and Intel defined ecosystem.
It was also mentioned that Virtual Reality opens the possibility of editing videos and games in a VR environment, whereby 3D-renders can be seen in life-like proportions allowing for more accuracy, detail, and creativity for game developers and video editors. This new VR environment for content creation is powered by the Unreal Engine.
Intel announced 7th generation Core along with a winter 2016 release date. The code name for the new processor family is Kaby Lake. The performance improvement seems a little overstated, but the new processor family gains a substantial graphical advancement, as opposed to instructions per compute cycle. Intel demonstrated a laptop running Overwatch on Kaby Lake architecture with acceptable frame rates. However, Intel will likely remain confined to the low-end of the graphics market despite updates to its Iris Pro architecture.
Intel mentioned that the processing requirements for autonomous functionality drives significant total addressable market expansion for flash storage. A single autonomous car will process up to 4,000 GBs of data per day, which implies that the incremental silicon opportunities from autonomous vehicles are meaningful, but not clearly defined for investors quite yet. I’ve read numerous sell side notes that try to articulate the opportunity, but establishing a timeline for level 5 autonomous functionality is an exercise in futility.
Datacenter presentation - launch of Xeon E5-V4 series
The datacenter presentation regurgitated much of the known information that was presented at prior data center days by Diane Bryant. However, I’m going to make efforts to highlight the new information.
Intel announced its launch of the Xeon E5-V4 series, which implies 1.4x performance gains from the prior generation. Furthermore, Intel updated us on network function virtualization (NFV), whereby radio access network (RAN) base station workloads will produce a 99.9% reduction in latency on the new Xeon-series CPU. The transition to cloud-RAN (C-RAN) utilizes software defined networking (SDN), which is why Intel continued the usual discussion on telecoms “big shift” to C-RAN and SDN, as it is a key component of the upcoming 5G mobile network footprint.
Now with that background information intact, the company updated its deployment roadmap, which is inclusive of 40 deployments (which is hard to quantify) and involves 10 major telecoms. Huawei and Intel partnered to deploy these solutions for telecoms. Huawei has already signed 50 service contracts, which implies rapid adoption of Intel Xeon series CPUs at various mobile base station sites in China and in many other regions of the world.
Intel mentioned the usual commentary on analytics and even resurfaced the same dialogue on Trusted Analytics Platform. However, what changed in this presentation from prior presentations was the dialogue on Machine Learning.
Intel believes that it can address this market, as Machine Learning has become center stage to the on-going development of algorithmically defined programming, whereby computers are able to self-assemble databases from scratch via an algorithm that can then teach itself to continually write and re-write a program.
In response to the massive data overflow, Google has developed the Tensor Processor Unit, which is designed specifically for machine learning algorithms.
However, Intel believes that its platform can address this market via Xeon Phi, Xeon E5 and Altera FPGA. The process of creating/running a machine learning algorithm can be divided into three major categories: model training (Xeon Phi), model scoring (Xeon E5) and model running (Altera FPGA). Intel addresses all of those needs with general purpose computing, accelerators, and a field programmable gate array (FPGA).
Machine-learning and natural language processing are key components to the eventual shift from web 2.0 to web 3.0 or semantic web. I believe Intel’s unique platform provides a strong alternative for companies that cannot afford to design custom silicon ahead of this massive transition.
Furthermore, Intel mentioned that it’s in the process of sampling and testing a fully packaged FPGA and Intel Xeon series chip. What this means is that a standard Xeon-series CPU will also integrate the FPGA into a single chip. This is referred to as on-die integration similar to a system on chip, or accelerated processor unit. The market potential for this new Xeon + FPGA chip is substantial, but not yet quantifiable.
IDF 2016 Day 2 presentation
The second day presentation was less uneventful with tons of information on broadly known PC trends.
However, Intel announced that it's transitioning the notebook category to be inclusive of mobile silicon, or the XMM 7260 Modem. This expands the total addressable market as Lenovo announced a new notebook category that is inclusive of 4G LTE capability and will come with 4GB of monthly wireless data (free from China Unicom). This will rapidly drive adoption of mobile modems in the PC segment.
Furthermore, Intel announced that it was launching Apollo Lake in 2H’16. The release refreshes the low-end mobile computing segment, which falls under the Cherry Trail family, or Atom X3, X5 and X7 series. It was broadly perceived that when Intel scaled back its mobile efforts that it would completely abandon the family. However, that’s not the case whatsoever, as the new generation of Intel mobile CPUs will reassert the competitive threats Qualcomm faces in the mobile segment.
Finally, the last important piece of investor news was the announcement of open connectivity foundation for edge devices. The announcement implies an open and unified standard for edge device communication regardless of platform as Intel, Microsoft, Samsung and Qualcomm are partners on this initiative. This creates open communication for IoT devices in the Wintel and Qualcomm/Google ecosystem.
I’ve made painstaking efforts to cover all the facets of the presentation that were not picked up by other news outlets/investment sites. While the new technologies are promising, I don’t anticipate these announcements to materially alter the investor thesis over the next twelve months.
However, the presentation points to key growth levers, confirmation of release cadence, and continued life in the mobile and IoT categories. These other areas help prepare investors for a sustained non-cyclical pattern of sales/earnings growth. But much of these new categories will take three-to-five years to fully materialize.
As such, I continue to reiterate my hold recommendation and $31.56 price target.