On getting integrated

The pendulum swings back — Apple kills Intel’s share price Back in the 1980s and all the way to the mid-1990s the big computer companies at the time were known as the BUNCH – Burroughs, Univac, NCR, Control Data, and Honeywell — names that probably aren’t even recognized now. They designed and built everything, as did their main and  competitor ...

Jon Peddie

The pendulum swings back — Apple kills Intel’s share price

Back in the 1980s and all the way to the mid-1990s the big computer companies at the time were known as the BUNCH – Burroughs, Univac, NCR, Control Data, and Honeywell — names that probably aren’t even recognized now. They designed and built everything, as did their main and  competitor IBM. They designed and built the OS, the power supplies, the terminals, the LANs, and the CPUs.

The big computer companies of the past did it all for themselves. (Source: Burroughs, stock)

As we approached the turn of the century merchant fabricators rose and offered economy of scale opportunities to systems builders. We soon learned the term fabless semiconductor company—a firm that would design the chip, but not build it. Outsourcing became the basis for all business plans—let the expert of a technology do the thing they are best equipped to do. There’s no need to reinvent the wheel.

And then in the 2010s something changed. Vertical integration again became the drive. One inspiration for the trend had to be Apple, which at the time had transformed itself from yet another and not too successful PC company into a powerhouse consumer electronics company challenging and, in some instances, replacing Sony the king of consumer electronics.

Other companies followed the model. AMD and Nvidia expanded their AIB building efforts to more products, they developed more software in the form of tools initially, and later fully-fledged applications and programming languages. Nvidia developed complete systems like Shield and the DGX super computer family, as well as the automotive Drive system. Intel, the most integrated with its own fabulous fabs had also gone into the end user market with AR, medical, and IoT devices. And, although Qualcomm doesn’t sell smartphones, they certainly design and build prototypes which many of their customers use as the base for their products. Recently the company showed a reference design for an XR HMD, as complex a system as you can imagine.

This is a logical and some might say necessary development as the ingredient companies move up the food chain, both out of a desire to expand their market size, and to exploit the engineering effort they have put into reference designs to prove and sell their products.

Apple was never an ingredient company but rather the ultimate customer for ingredient suppliers. Over time the company has acquired skilled engineers to help them pick the right parts and partners and in time those engineers became the lead and put the company in the position of demanding certain specifications and performance levels. The story is that the only way Apple agreed to switch from Motorola to Intel for CPUs was if Intel would design and build a specific type of CPU, which became Intel’s vaunted 2005 design, released in 2006, the low power, low profile, Core Duo which made the revolutionary Air possible.

In 2007 the iPhone and iPod Touch came out based on Arm SoCs specified by Apple and manufactured by Samsung. And then in 2010, the first of the A-series SoCs designed by Apple appeared in the iPad. The A4 also had an Imagination Technologies-designed GPU in it, and it was rumored that Apple had more to do with the design of the PowerVR than Imagination did.

Apple was being a savvy buyer. It was like getting a custom-made suit for the price of a suit off the rack. As Apple’s volume in i-things soared, so did the fortunes of the ingredient providers in its supply chain.

Slowly but steadily and unrelentingly over the next few years Apple subsumed one ingredient supplier’s part after another with one of their own, and in some cases bought the ingredient supplier. As the number of suppliers Apple was using diminished, the industry’s attention began to focus more on its main supplier for the Mac line. However, Macs were such a small part of Apples revenue it was often asked would Apple just drop the Mac?

Apple was able to design i-things and Mac-things to work beautifully together. Phone calls transfer seamlessly, messages are always accessible, etc. The linkage of i-thing users and mac-things are what makes Apple products desirable to users and justifies higher prices. Apple wouldn’t want to alienate its core users.

Rumors of Apple abandoning Intel and using their own CPU design has persisted. Never mind that that Apple’s ARM processor wouldn’t be easy to adopt to the Mac’s OS or Apple’s application suite, and third-party applications like Adobe, and Microsoft’s Office.

Then, in late 2016 rumors began to leak out about Qualcomm and Microsoft developing a super phone or maybe it was a small PC—a device that would run full-fledged Windows and all its applications, and yet run on an Arm-based Qualcomm processor. In sweltering Taipei in the summer of 2017 Qualcomm and Microsoft showed working prototypes that looked very much like finished products, and in December at a Qualcomm event on yet another Pacific Island, Qualcomm showed finished products from HP, Acer, and others of what was now being called the always on PC.

At CES Qualcomm and Microsoft showed them to the world, and Intel showed their x86-based always on version with Intel Gen4.n modem. A new category had been introduced that would compete with, and probably overwhelm Chrome books.

Where was Apple’s entry?

The speculation is that Apple looked at the always on computers, Qualcomm’s in particular, and reasoned that if Qualcomm could adopt an Arm CPU to Windows, Apple could damn well adopt their own Arm processor to their own iOS and Mac suite of apps, and if third-party ISVs were smart, they’d get on the bandwagon and do the same. Obviously, Apple would use software shims on the first versions but in two ore three years things should be native and more efficient.

Putting all that speculation together and looking for an opportunity to short, or buy into, some companies, the smart people in Wall Street, and the even smarter people who tell them what’s going on (real or otherwise), concluded Intel is on life-support with Apple. Naturally, no good (and no damn good) rumors ever get wasted on Wall street, and Intel’s share price got knocked down. Now since the PC business is about 14% of Intel’s overall revenue, and Apple has about 8% market share now, or 1% of Intel’s business (if that) Intel’s shares were knocked down 2% normalized to the gyrations of NASDAQ, while Apple’s share price went up



Share price changes normalized to NASDAQ


Is it fair? No. Is it logical? Hell no. is it normal Wall Street herd behavior? Yes.

Apple represents less than 1% of Intel’s revenue

Wall Street, it seems, just doesn’t like Intel, even though it has invested and sold more in automotive than Nvidia can dream of, and more servers than AMD can even imagine. I asked some investors once, why does Intel’s share price go down when they report a good to great quarter? The answer was, we don’t think they can do it again. And yet they’ve done it almost every time. Sometimes it seems Wall Street will punish a company for being too successful.

That rumor by the way, about Apple dumping Intel, says if, IF, it happens, it won’t happen until 2020. If that’s true, then Intel damn well knows about it, and has made whatever adjustments it needs to make. I’m not so sure Intel’s management is wringing their hands and losing sleep about a possible 1% loss in sales two years out. It seems likely that Intel can manage to wring out a 1% increase in profit to offset Apple’s defection.

Another bus, always comes along.