News
EEtimes
News the global electronics community can trust
eetimes.com
power electronics news
The trusted news source for power-conscious design engineers
powerelectronicsnews.com
ebn
Supply chain news for the electronics industry
ebnonline.com
elektroda
The can't-miss forum engineers and hobbyists
elektroda.pl
Products
Electronics Products
Product news that empowers design decisions
electronicproducts.com
Datasheets.com
Design engineer' search engine for electronic components
datasheets.com
eem
The electronic components resource for engineers and purchasers
eem.com
Design
embedded.com
The design site for hardware software, and firmware engineers
embedded.com
Elector Schematics
Where makers and hobbyists share projects
electroschematics.com
edn Network
The design site for electronics engineers and engineering managers
edn.com
electronic tutorials
The learning center for future and novice engineers
electronics-tutorials.ws
TechOnline
The educational resource for the global engineering community
techonline.com
Tools
eeweb.com
Where electronics engineers discover the latest toolsThe design site for hardware software, and firmware engineers
eeweb.com
Part Sim
Circuit simulation made easy
partsim.com
schematics.com
Brings you all the tools to tackle projects big and small - combining real-world components with online collaboration
schematics.com
PCB Web
Hardware design made easy
pcbweb.com
schematics.io
A free online environment where users can create, edit, and share electrical schematics, or convert between popular file formats like Eagle, Altium, and OrCAD.
schematics.io
Product Advisor
Find the IoT board you’ve been searching for using this interactive solution space to help you visualize the product selection process and showcase important trade-off decisions.
transim.com/iot
Transim Engage
Transform your product pages with embeddable schematic, simulation, and 3D content modules while providing interactive user experiences for your customers.
transim.com/Products/Engage
About
AspenCore
A worldwide innovation hub servicing component manufacturers and distributors with unique marketing solutions
aspencore.com
Silicon Expert
SiliconExpert provides engineers with the data and insight they need to remove risk from the supply chain.
siliconexpert.com
Transim
Transim powers many of the tools engineers use every day on manufacturers' websites and can develop solutions for any company.
transim.com

Suppliers Beware: The Perils of Vertical Integration

By   07.12.2017 0

There are basically two types of organizations in the computer and mobile industry: device makers who are, to the extent they can be, vertically integrated, and component or IP suppliers that offer their product or technology across several vertical product types to the device makers.

A successful salesman once told me, “Sales is easy, just listen to the customer — he’ll tell you what he wants to buy.”

Apple has been telling Imagination Technologies for years what Apple wanted to buy, and Imagination has been delivering the desired designs since 2006. Some people have even suggested that Apple has been largely responsible for Imagination’s technology development roadmap. Apple is what Imagination calls a lead partner, which puts it in a position to make inputs to Imagination’s roadmap — a status that has also been enjoyed by other customers, including Intel, TI, Renesas etc., so Apple can’t take full credit.

Intel pursued Apple for years until, finally, Apple told Intel what it wanted to buy. Intel built it and, in the process, created a new low-power and powerful mobile processor (Core/Penryn). Apple also told Intel what it wanted in an embedded GPU, which resulted in Intel improving its integrated GPU to the point where Intel dropped Imagination as a GPU IP provider for its Atom processors. More recently, wanting even more GPU power, Apple allegedly insisted that Intel combine an AMD GPU in a multi-chip, low-profile part with an Intel x86 processor (contrary to popular chatter, this does not require Intel to get a license from AMD).

Then, in April 2017 Apple announced it would no longer use Imagination’s GPU IP in its iOS devices, and would instead design its own GPU.

Apple’s abandonment of Imagination Technologies’ GPU, in favor of an in-house design is bold, and questionable. Imagination is looking into the possibilities of patent infringement, and it probably will find it.

Apple has not presented any evidence to substantiate its assertion that it will no longer require Imagination’s technology, without violating Imagination’s patents, intellectual property and confidential information,” Imagination said.

“This evidence has been requested by Imagination but Apple has declined to provide it. Imagination believes that it would be extremely challenging to design a brand-new GPU architecture from basics without infringing its intellectual property rights. Accordingly, Imagination does not accept Apple’s assertions.

Imagination has said it reserves the right to sue Apple for unauthorized use of its confidential information and intellectual property. Pursuing that course might be tricky for Imagination, which does not currently have deep cash reserves. Note, I say currently; we will return to this later.

Imagination may get a small concession from Apple for its discovery efforts. Apple’s future GPU is likely to be more like an AMD GPU, however, and — if it is — then AMD too will feel the goodbye kiss on the cheek (“It’s not you AMD, it’s me…”).

As is well known, Apple used the concepts of the classic binding model from OpenGL and applied it to OpenCL. In so doing, Apple abandoned the Khronos industry standard OpenGL and OpenGL ES as well as Vulkan API, all of which point to that vertical integration thing. Apple did that for reasons of control and to discourage portability of iOS-based applications to other platforms, demonstrating once again the company’s strong desire for control over its product and ecosystem. Such a move isn’t illegal, but it is anti-competitive by locking developers into its platform and forcing the ones who can afford it to support two APIs.

A common API generally needs a common OS; not necessarily, but preferably. Apple has also established its own OS and spread its kernel across both its ARM processors in iOS and the Intel processor in its macOS High Sierra (OS X), which must have inspired no small amount of envy and motivation to Microsoft in its efforts to have Win 10 run on ARM. Eventually, the two Apple operating systems will likely merge and, when (if) they do, a Mac will still behave like a Mac and iThings will operate the same, and both platforms will share apps and file compatibility. The main difference between them now is the UI. Nonetheless, it is another demonstration of the almost maniacal desire for control of everything in the Apple product sphere.

Since Apple uses AMD AIB’s in the Mac Pro, and discrete GPUs in the MacBook, it makes sense to use the same type of GPU in its systems that have integrated GPU — one common code set. And, when you can dictate to your suppliers, why shouldn’t Apple do that? Intel is trying to get its modem established in Apple mobile devices by positioning it against Qualcomm. Apple knows this, and so it can get concessions — and just tell Qualcomm, we’re not going to pay you — how’s that for taking control of your stack? Using a second source to get price concessions from the primary supplier isn’t new, or evil. Not paying your bills because you don’t like the price of IP is something different. So, Intel should be careful of what it wishes, Apple might just say: “Thanks for the fish and the modems, the door’s over there, show yourself out, will ya?”

Apple has always operated on the principle of diminishing the supplier’s power — to be the dictator rather than the dictated — from the assembly line to the IP and all the parts in between. The problem with this philosophy is that it narrows Apple’s opportunity to innovate at the foundation level (not just Apple, any company that follows such a philosophy). As clever and rich as Apple is today, it can’t be the smartest, fastest, or best at everything that goes into a complex, always-on computer; no company can. And yet the company is cutting itself off from the opportunity to leverage the innovation of the industry at large. It has severed ties with Imagination Technologies and Khronos, and it’s refused to pay Qualcomm for its IP. It uses ARM in name only, and has twisted Intel around into doing its bidding. Intel! Who gets to tell Intel what to do? And AMD will be added to the list. All of this, in the name of cutting costs and being vertically integrated, places Apple in even more of a bunker mentality and narrows its options and opportunity at innovation. And this is happening at a time when Apple is being questioned in the press and by investors regarding its ability to innovate. What’s wrong with this picture?

The benefit Apple, or any company, gets in dealing with technology suppliers — Imagination, Qualcomm, Intel, AMD, Khronos, etc. — is that those suppliers respond to the needs of multiple customers, markets, and applications. In doing so, they learn of various requirements, spread the costs, drive innovation, and move the industry forward. Apple is a throwback to the old days when companies like IBM and Burroughs did everything for themselves. Those old big-iron companies did that because they had to, but having vertically integrated companies that did everything is what trapped the industry for so long in limited innovation proprietary systems. It was the emergence of the merchant supplier, the ODM, the common fabs, the fabless IP providers, and the open standards that propelled many industries (e.g., PC, mobile, TV) forward. Why then, would Apple turn away from leveraging the innovation of industry-leading technology providers? To save a few pennies on GPU IP? Are its margins so threatened by the competition that it can rationalize stone-walling Qualcomm on royalty payments or hiring dozens of GPU engineers to save pennies per phone by dumping Imagination?

By making demands on its suppliers Apple forces them to invest a lot of resources in meeting its requirements. That’s normal and it’s how things should work: listen to the customer and then give him what he wants to buy. Apple, however, first uses the technology, then squeezes the suppliers to make it cheaper, and then — at a certain point — says it’s not cheap enough and goes off and make it itself (having learned over time from the suppliers how to make it). Apple has been hiring GPU architects and designers for years now; several from Imagination. Again, not illegal, and maybe not unique, but how badly do you want a customer like that?

When small suppliers lose a big customer like Apple, they usually can’t find another large customer to replace the lost business and — as a result — struggle to remain solvent, get bought cheaply, or are forced to diversify. This is another case of “Be careful what you wish for”: love to get an order from a big fish like Apple, and then enjoy yourself until it eats you.

And so now, mortally wounded after losing almost 70% of its market value, Imagination is seeking a buyer. And even though that announcement raised its share price 15.3% to 165.75 pence in London, the stock has dropped 38.1% since April 3 following the Apple pull-out announcement.

Who would buy Imagination? Well the company says it has attracted curiosity from a number of parties interested in acquiring the company. But what if one of them were Apple? It would be an incredibly cynical play on Apple’s part: to knock down the price of Imagination and then go buy it. Apple, which (now) owns a 9.5% stake (up from 8%) of Imagination, took that position to prevent a hostile takeover of one of its major suppliers. We do know that Apple discussed a purchase of Imagination previously, in early 2016, but the companies failed to reach agreement. More likely, if Apple does buy Imagination now, is that it is doing so in response to fears that the graphics supplier might fall into the hands of new owners with pockets deep enough to pursue Apple for infringing Imagination’s patents. Ironically, Apple is in a position to dictate who can and can’t buy Imagination. How’s that for the ultimate control — Apple is a control freak. Oh wait, we knew that, didn’t we?

Postscript
Apple appears to be making an aggressive move to poach more staff from Imagination Technologies. The Telegraph recently reported that Apple has set up its own GPU design team in offices just a few miles from Imagination’s campus.

The Silicon Valley giant has planted its flag by renting a 22,500 square-foot office in St Albans, a stone’s throw from Imagination’s headquarters. It plans to use the office to develop its own graphics technology as it ditches Imagination, leading to fears that it will poach the British company’s most talented staff…

The Telegraph reports that Apple has already hired a number of Imagination employees over the past few months, including COO John Metcalfe, but it seems likely that others would be reluctant to move to either London or Cupertino. Basing operations on Imagination’s home turf would make it much easier to recruit further staff by removing the need for them to move home.

The new office in the centre of St Albans sits a few miles from Imagination’s base in the adjacent Hertfordshire village of Kings Langley.

Apple gave Imagination notice that it would be designing its own GPU chips in future, and that it expected to achieve this within two years, at which point it would cease license payments. Imagination responded by invoking the dispute-resolution clause in its contract, and claimed that Apple would not be able to design a graphic chip without using its patented technology.

The battle between the two companies intensified last week, Imagination saying that Apple had made ‘highly regrettable… unsubstantiated claims‘ that it no longer needed the chipmaker’s IP, and the Cupertino company calling those statements ‘misleading‘ and disputing the timeline of events.

So, how do you think all of this is going to play out?

Dr. Jon Peddie is one of the pioneers of the graphics industry, and formed Jon Peddie Research (JPR) to provide customer intimate consulting and market forecasting services. He lectures at numerous conferences and universities on topics pertaining to graphics technology and the emerging trends in digital media technology. Recently named one of the most influential analysts, Jon regularly advises investors in the technology sector, and is frequently quoted in trade and business publications. Jon is a past president of The Siggraph Pioneers, contributes articles to numerous publications and — in 2015 — was given the Lifetime Achievement award from the CAD Society.

Jon is also the author or co-author of several books including Augmented Reality, Where We All will Live (2017), The History of Visual Magic in Computers (2013), High Resolution Graphics Display Systems (1994), Multimedia and Graphics Controllers (1994), and Graphics User Interfaces and Graphics Standards (1992). Jon is also Contributing Editor, Handbook of Visual Display Technology (2017), and he is currently working on a new book on VR.

0 comments
Post Comment
Ravishankar_#1   2017-07-14 03:21:18

Jon, Very interesting article. I read this article along with the other article that appeared on EET Asia on 21st April,2017 by Mary E.Shacklett ["5 key lessons from Apple-Imagination GPU dispute"]. We all know that Mali GPU and Vivante GPU have been competing with Imagination GPU cores for a while and then you have the NVIDIA and the AMD GPU cores that more come from the PC world. Who holds many of the fundemental patents in this segment ? I am trying to understand the basis on which Imagination wants to sue Apple. 

jonpeddie   2017-07-14 12:37:00

GPUs are tricky things, and the companies that did the foundation work (ATI, 3Dlabs, Nvidia, and Imagination), know that and can safely assume a new company (Apple in this case) would have to viloate some IP. But it makes the lawyers rich in teasing out what IP has been viloated, and very often (85% of the time is my guess) its an innocent hapenstance -  there are just so many ways to design a shader, etc.

perl_geek   2017-07-15 13:56:34

It's always an uncomfortable strategic position when one organisation represents a significant part of your supply chain, especially if it's much larger than you, and an agressive potential monopsony. The sales may be easy and the money good for now, but what about tomorrow? (Especially if your IP goes home most evenings.)

Management in that situation would be prudent to review non-disclosure and non-compete clauses in everyone's contracts, and doing their best to publicise any malevolent HR stories they can find about Apple. (Even without Jobs, horror stories souldn't be too hard to find.) "When Apple have sucked you dry and offer you a choice of relocating to Ireland or qutting because they're closing the convenient local office, what are you going to do?"

sranje   2017-07-17 21:48:31

Re:     As clever and rich as Apple is today, it can't be the smartest, fastest, or best at everything that goes into a complex, always-on computer; no company can. ( ??!! )

Jon, you know infinitely more about the subject but -- is your above statement correct?

One can always acquire technology teams or knowledge for internal technology development and catch up and proceed ahead with whatever a company wants to develop - correct?

Perhaps you are too pessimistic on risks of going internal after many years of closely studying various technologies and alternatives. 

And GPUs are indeed becoming an all-important technology -- Imagination was just too small and fragile for lasting comfort

name99   2017-07-17 22:45:14

<<

GPUs are tricky things, and the companies that did the foundation work (ATI, 3Dlabs, Nvidia, and Imagination), know that and can *safely* assume a new company (Apple in this case) would have to viloate some IP.

>>

 

Jon you keep insisting on this. And what I see when I read this is Ed Colligan:

"We've learned and struggled for a few years here figuring out how to make a decent phone. PC guys are not going to just figure this out. They're not going to just walk in."

You're just assuming a whole bunch of things that are not necessarily true. 

First is the assumption that Apple is creating a GPU rather than a generic throughput engine (think Larrabee/Phi or Cell). The fact that Intel and IBM couldn't do this properly doesn't imply that the idea is bad, rather it tells us that you need a certain level of competence in implementing the idea. Intel insisted on doing this on top of the x86 ISA which was a pointless complication. IBM insisted on using a completely non-standard memory model which made programming a nightmare. Apple will certainly not repeat those two mistakes, and, given their willingness to provide IMG with definite dates, likely already has something working. The point is, the secifics of a GPU made sense when that was pretty much all the throughput computing of a CPU. But in a world where we also require throughput computing for audio processing, a variety of AI tasks, and a variety of different image manipulation tasks, starting with a design optimized for 3D processing and trying to contort it to do other things makes less sense that starting with a generic thjroughput engine.

Second is the assumption that Apple is going into this patent-free. Says who? How do we know that they haven't signed up with any of the possible partners (Intel, nV, ARM, AMD) for some sort of GPU-IP sharing agreement?

jonpeddie   2017-07-18 08:44:25

Been multi-monitor'ing, didn't see these posts, which I love.

 

You (name 99 -  I like that too, I am not a man, I am a number :-) ) are absolutely right, I hadn't considered other ways of skinning the cat. There is nothing that says a "GPU" has to be an array of FPUs, or even AUs, one could for example use an array of DSP-like elements. Many years ago, we did such a thing. And we also built a programmable image processor using bit-slice designs. I've been involved in projects that used Transputers for graphics. Larrabee was not a HW failure, it got killed by the corner cases in trying to be a GP graphics processor. If you take a limited case (and I don't know if we can assume a smartphone is a limited case, but it's not as messy as a PC), then it's perfectly reasonable to assume there are alternate architectures. In fact, I'd really like to see such a thing – the industry is overdue for some left field thinking.

 

And of course, you're right about the patent thing. Apple is big, rich, and smart. They wouldn't (at least I can't imagine they would) walk into a mine field blind folded. I couldn't/don't understand why Apple didn't just buy Imagination. And I guess you saw the story about Apple opening up an office near Imagination and hiring more folks, that seems kinda underhanded.

 

As for your comments, SRANJE, yep, you too are right. A company can acquire tech, you read about it every day. However, some companies (and you know their names) have acquired amazing tech and then didn't know how to deal with an unbridled start up team and drove the clever folks away, ending up with a damaged balance sheet.

It's situations like this that make me impatient for the future. I can't wait till mid-2018 to see what Apple does in developing a graphics processor. I hope they disclose at ISSCC and/or Hot Chips or Siggraph.

Thanks all for your comments, it's great knowing someone actually reads this stuff.

jonpeddie   2017-07-18 08:45:00

Please see above

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.