Rethink Technology business briefs for June 22, 2018.
Intel’s golden opportunity
Brian Krzanich. Source: Anandtech.
The sudden departure of Intel’s (INTC) CEO, Brian Krzanich, presents Intel with a golden opportunity to reinvent itself and make a break with its past. That need is paramount. The fundamental failure of leadership of Krzanich has been a lack of adaptability and open mindedness. All too often, true innovation and accomplishment were replaced by dissembling and disinformation.
Even as Krzanich and company have presided over the decline of the Wintel PC, they have congratulated themselves for maintaining and even growing revenue slightly. Krzanich became CEO in May 2013. From fiscal 2013 to 2017, revenue grew by just 19% to $62.8 billion, while GAAP net income was flat at $9.6 billion.
In contrast, Apple (AAPL), despite being much larger, still grew faster in roughly the same time interval. From Apple’s fiscal 2013 to 2017, Apple revenue grew by 34% to $229 billion, and GAAP net income grew by 30% to $48.35 billion.
I bring up Apple specifically because Apple has been to a large extent responsible for the decline of the Wintel PC as it invented powerful mobile devices that made the PC less than essential. Apple didn’t just invent the modern smartphone and tablet, it invented a new business model that has steadily eroded the PC commodity model. Instead of buying commodity processors from a company like Intel, Apple elected to design its own.
These processors, based on an alternative computing architecture from ARM Holdings (OTCPK:ARMHF), would be fabricated by a silicon foundry. Thus, Apple established the key elements of what I have called the new paradigm: internal design of processors as part of overall product development, use of ARM architecture, and fabrication by an independent foundry.
When confronted with the new paradigm, the Intel of Krzanich simply dug in its heals. Intel proclaimed that it had the better architecture for mobile devices. Intel extolled the virtues of its business model as an Integrated Device Maker (IDM). And Intel insisted that it maintained process leadership compared to the foundries.
Even as Krzanich proclaimed Intel’s technological superiority, a wealth shift was occurring from the commodity PC paradigm to the new mobile paradigm. Not only did Apple outgrow Intel, but Apple’s key foundry partner, Taiwan Semiconductor Manufacturing Company (TSM) grew even faster. From fiscal 2013 to 2017, TSMC revenue grew by 63% to $32.1 billion and net income grew by 82% to $11.27 billion.
Despite the missed opportunities, Intel remains a powerful and profitable technology company. A new Intel management will have the opportunity to fix what is broken at Intel and get Intel moving forward again. But first there needs to be cognizance of what went wrong, and under Krzanich, that cognizance simply hasn’t existed, at least not publicly.
Failure #1: mobile devices
For its fiscal 2014, Intel reorganized its mobile device activities into the Mobile and Communications Group. This included Intel processors for mobile devices as well as the wireless modem business that Intel had bought from Infineon in 2010. At the same time, Intel began a push into tablet processors using what it called “contra revenue”.
Intel management claimed in its 2014 Q1 conference call that the purpose of contra revenue was to offset “bill of materials” costs because initial tablet processors were being repurposed from the PC market. As CFO Stacy Smith explained:
But what we’re doing is we’re taking Bay Trail, which is a product really designed for the PC market, and we made the decision to take it broadly across different segments of the tablet market this year.
It brings along with it, at least over the course of 2014, a higher bill of materials. And that’s independent from the SOC cost. It’s the power management subsystem, it’s the motherboard that it goes on, it’s the memory solution, those kinds of things. And so, we’re providing some contra revenue to offset that bill of material delta over the course of 2014.
Now, as we said, we’re doing value engineering with our customers and our partners. And so we’re bringing down that bill of material over the course of 2014 independent of any changes to our SOC.
Intel bragged about the gains it was making in unit sales of tablet processors, and by the end of the year, it had sold 46 million of them, making it among the largest “merchant vendors” of tablet processors, a crown that would rest heavily, and all too briefly, on the head of Krzanich.
There was no sign of any reductions in contra revenue over the course of the year as a result of bringing down the BOM costs, or of anything else. In 2014 Q4, Mobile and Communications Group posted negative revenue for the quarter of $-6 million. For the entire fiscal year, MCG revenue was down 85% y/y to just $202 million, while the unit racked up an operating loss of $4.2 billion.
Intel supporters attempted to claim that the loss was due to investments in wireless, and this may have contributed to some extent. But it’s likely that most of the loss was due to contra revenue. It was clear by the end of the year that Intel had used contra revenue as a way to subsidize unit sales and gain market share. The reaction of management to the mobile losses was to fold MCG into the PC Client Group, thus making it difficult to discern the losses in 2015.
Those losses continued throughout 2015, I believe, and only reduced as contra revenue was wound down. As contra revenue wound down, so did unit sales of tablet processors. In 2015 Q3, tablet processor sales had fallen to 8 million, down 39% y/y.
By the 2015 Q4 earnings report, Intel was no longer announcing tablet processor sales, or even talking about other mobile device initiatives in smartphones. One of the most important of those initiatives was to have been SoFIA, Intel’s first mobile SOC that combined its Atom series processors and its cellular modems.
In May 2016, Intel formally announced that it was discontinuing its Atom processors for tablets and smartphones, including SoFIA. Why did Intel fail in mobile? It was not due to lack of technical performance. An Intel processor could be made every bit as powerful and energy efficient as an ARM processor, whether designed by Apple or other smartphone makers such as Samsung (OTC:SSNLF) or Huawei.
The problem for Intel was not technical competitiveness but cost competitiveness, I concluded. ARM architecture, combined with the features of the new paradigm business model, made ARM processors inherently less expensive on a performance normalized basis. Much of this cost advantage seems to be due to the fact that ARM processors are physically smaller, all other factors being equal, than their Intel counterparts.
It’s always been difficult to find a true “apples-to-apples” comparison between an ARM processor and an equivalent Intel processor. The nearest I’ve been able to come is the A11 Bionic in iPhone X and the Core i5-7360U in the mid-2017 13” MacBook Pro. As shown in the chart below, the processors test out to be roughly equivalent in performance using Geekbench single and multicore benchmarks.
The Core i5-7360U carries a recommended price of $304. TechInsights estimates the cost of the A11 Bionic to be $52. Intel has stopped publishing the die sizes of its processors, but Anandtech estimated that the 7360U has a die size of about 124 mm^2. The die size of the A11 Bionic has been measured to be 87.66 mm^2.
There is a significant difference between the chips, in that the A11 is fabricated on TSMC’s 10 nm process, while the 7360U is fabricated on Intel’s 14 nm process. But Intel has claimed that its 14 nm process is roughly comparable to TSMC’s 10 nm process, so this difference shouldn’t be very important.
Source: Intel China Technology and Manufacturing Day, September 2017.
In retrospect, it’s hard to understand what Intel thought they were doing with contra revenue. The bill of materials excuse never really worked, but it’s not clear that Intel management or Brian Krzanich understood that they were fundamentally cost disadvantaged. It may have been believed that once Intel achieved volume production of its mobile processors, economies of scale would be sufficient to achieve profitability.
Or it may be that Intel always realized that it was cost disadvantaged, and simply attempted to buy its way into the market by spending some if its profit. This is the interpretation that I think is more reasonable. In this scenario, Intel would simply drive its ARM competitors out of business, then raise prices once its competition had been suppressed.
In either case, Intel badly miscalculated by underestimating the impact of the new paradigm and the wealth shift that had occurred. Intel could only afford to buy into the very lowest end of the tablet market. In 2014, cheap tablet brands came out of the woodwork. Remember RCA tablets? Apple’s iPad sales were hurt, but Apple’s response was to move up market with the iPad Pro, announced in September 2015. Given Intel’s 2014 mobile losses, Apple had more than enough resources to wait Intel out, and the cheap Intel tablets have been relegated to the proverbial dustbin of history.
Intel’s failure in mobile had profound implications that Krzanich and company have never seemed to deal with. If Intel was cost disadvantaged in mobile compared to ARM, it would be cost disadvantaged in other areas of computing once ARM processors were developed for these markets.
Even if Krzanich seemed oblivious to the implications of the mobile failure, Intel’s competitors have not been. Qualcomm (QCOM) undertook the development of a server processor based on ARM architecture, the Centriq 2400. Even though Centriq has hit some obstacles, it’s still alive and poses a threat in the long run. Qualcomm is also making ARM processors for Windows 10 “always connected” notebooks.
Apple is reported to be looking at using its custom ARM processors in a future generation of MacBooks. Using its custom mobile processors would enable Apple to offer the thinnest, lightest, notebook computers around, and these MacBooks would also have better margins.
Coping with further inroads into its processor business by the new paradigm is going to be one of the most important challenges facing a new Intel management. That management needs to make fundamental decisions about business model and technologies. All technologies become obsolete in time. Did Krzanich and company expect that x86 would go on forever? No. I think Intel’s management has simply chosen to maximize near-term profitability while procrastinating on the issue of what comes after x86.
It may have been hoped that any decision could be forestalled until after the current generation of management had comfortably retired. In Krzanich’s case, that has more or less turned out to be the case.
Apple, Qualcomm, and TSMC are part of the Rethink Technology Portfolio, and all are rated buys. Consider a free trial subscription to Rethink Technology and enjoy reports on technology companies available exclusively to subscribers.
Disclosure: I am/we are long AAPL, QCOM, TSM.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Editor’s Note: This article discusses one or more securities that do not trade on a major U.S. exchange. Please be aware of the risks associated with these stocks.