- Intel’s Computex keynote covered some PC themes, but really narrowed in on emerging technologies.
- Financial results are predicated on PC shipments, which could stabilize and grow.
- The inventory trends are implicit of anticipating higher end-market demand.
Intel (NSDQ:INTC) transitioned its focus away from its core PC franchise at Computex Taiwan 2016, as it tried to highlight more of its efforts in virtual reality, machine learning, big data and network transformation. In other words, the PC-themed show has transitioned to forward-looking technology and with both smartphones and PCs in the doldrums this year, it’s not really surprising. Of course, I’m starting to get incrementally positive on second-half 2016 themes for the PC market, but that’s really dependent on PC OEMs as both Apple and HP have been gaining market share against rivals.
Key takeaways from Computex 2016
At the press event, Intel released a new high-end variant of the Core i7-series, the Core i7 6950X, which falls under the Broadwell processor family, which takes advantage of a node shrink from 22nm to 14nm. The price point is a bit excessive at $1,750, but given the performance dynamics perhaps being roughly comparable to mid-end Xeon-series CPUs, Intel priced the product at a premium to prevent cannibalization in the enterprise segment. Furthermore, I don’t anticipate the high-end SKU to materially alter ASPs by much as shipments are unlikely to exceed a million units. The high-end model is inconsequential as the PC market as it's on-track to sell 265 million units this year, of which less than 1% will be inclusive of Extreme edition i7 CPUs.
The niche market, or high-end is usually more expansive for GPU vendors, which I will discuss in a separate article. However, the transcoding and compression efforts for high-end content creation are noteworthy, as a Xeon-comparable equivalent in the consumer i7 line-up is set to be released in the next couple months.
Furthermore, the company announced Kaby Lake at the event, which implies that late in 2016, we will witness the 3rd generation 14nm family. There’s been a lot of hype pertaining to the end of Moore’s law, and whether Intel is signaling the end of a virtuous cycle of rapid innovation within the PC-segment. However, I believe Intel is transitioning to more of an iterative process that’s less dependent on fabrication improvements to sustain instruction per clock cycle improvements to its processor family. In other words, the development of technologies below the 5nm threshold is still questionable, but the other innovations on the processor architectural design can extend the performance growth we’ve gotten accustomed to watching over the past several decades.
I believe much of Intel’s inventory build ahead of the summer launch of various OEM PCs can be explained based on the announcement of new notebook offerings at Computex, as desktop PCs have consistently declined. Intel coordinates with OEMs ahead of launch, and with inventory burn remaining a prevailing theme among the OEMs in Q1’16, the next step is a pattern of inventory build among the OEMs as they build capacity in the sales channel for imminent product launches. In other words, the market should soon stabilize, which implies better end-market demand once we push into the summer months.
I believe Intel will launch Kaby Lake to further encourage consumers to upgrade PCs in the holiday quarter, which is why inventory trends are likely to fluctuate quite considerably. The remaining semiconductor market is also showing modest patterns of expansion. However, Intel’s q/q inventory trends in Q1’16 were up 8.9% quarter on quarter, which was above seasonal comparisons of 2.1% q/q, according to Credit Suisse research. This may imply less efficiency on Intel’s effort to manage the supply side, but with semiconductors operating on larger inventory turnover cycles, I believe Intel is addressing the reduced PCs in the current retail channel in conjunction with summer refresh of high-end OEMs. Of course, there are some caveats to this, but meaningful sell-in of MPUs (processors) will imply growth in units as we progress into the Q3’16 timeframe. I also anticipate some expansion in the high-end PC market implying ASP improvements, but articulating the impact on Intel’s results will remain difficult for the foreseeable year.
Many of the OEMs made incremental improvements on form factor for Ultrabook 2-in-1 devices at the event, with high-end SKUs making another round of introductions inclusive of new ports with higher bandwidth for displays and USBs. Windows PCs tend to dominate a very small fragment of the high-end market, so we will need to wait for the Apple MacBook 2016 refresh to see meaningful ASP improvements. In other words, the Windows ecosystem isn’t much of a needle mover to total end-market demand, for high-end equivalents. So the presentations at Computex from the various OEMs have very little implication on whether Intel has a healthy enough end-market to sell MPUs.
Efforts in connected device and VR
The implications of VR and connected devices was quite interesting. Intel launched a system on chip (AnyWAN GRX 750) that’s capable of handling both mobile and fixed broadband connections out of a home router. It’s likely to be shipped inside of household routers, with adoption from various cable operators a potential driver to Intel’s revenue from connected homes. Of course, there’s a lot of development that needs to take place, as much of the prevailing infrastructure is starting to hit the market in the 2016 timeframe, but creating end-devices like light switches, air conditioning units, garage doors, and etc. that can connect to the edge of the cloud hasn’t reached mass market adoption/appeal.
Furthermore, integrating that functionality into a mobile environment and making the experience seamless for consumer devices creates hurdles to adoption. I believe smart appliances and consumer products are a key driver to edge of the network innovation, but with little incentive for sleepy-industries like appliance makers and consumer durables to adopt the technology we’re likely to witness slow headway into this specific category.
Data aggregation relies heavily on the data inputs we receive from devices at the edge of the network, so there’s potential for Intel to expand the market via its Quark and Atom line-up, but much of the attention has been in autonomous driving or high-level usages where data has higher value for business-to-business use cases. The implications financially apply more to Intel’s enterprise segment as opposed to consumer client computing.
Furthermore, Intel showcased VR implementation via the HTC Vive at its keynote. I believe this specific OEM will struggle in this environment because the product doesn’t have as much consumer interest as the Oculus Rift. Of course, it’s early days for VR, so declaring a winner would be foolhardy. However, unit shipments for VR headsets is anticipated to reach roughly 1 million in the 2016 timeframe. It doesn’t alter mx-shift assumption for high-end i7 series CPUs, but the incremental offering does imply a steep adoption curve for higher-end use cases for consumer PCS. This may imply better cyclical trends for PCs as there’s been an oasis of innovation up until 2-in-1s and VR.
There’s still potential for innovation for client side computing in both connected homes and emerging technologies like AR and VR. However, I would be cautious of anticipating immediate financial impact as these industries are nascent and inconsequential to 2016 financial performance.
In terms of PC-trends, I get incrementally positive, because inventory build is implicit of better end-market demand for PCs (if Intel forecasted correctly). In other words, some of the incremental technologies seem appealing enough, so consumers upgrade or adopt higher performing PCs beyond tablets and smartphone devices.
However, the industry has gone through a period of tightening inventories and consumer fatigue for electronic devices in general. Therefore, investors are heavily reliant on Intel’s ability to accurately forecast demand ahead of the summer launch of desktop and laptops. If the market doesn’t perform above seasonal trends, we could witness a glut of inventory in various OEM channels, which would have negative implications on FY’16 financial performance.
Given these risks, I continue to reiterate my hold recommendation and $30.54 price target.