It’s been only a little over 10 hours, way too soon to pass judgment on AMD’s choice of the new CEO. True. But one thing keeps bugging me.
During a conference call with AMD, both Rory Read, AMD’s new CEO, and AMD management harped on a single point: “He [Read] brings a customer point of view." Before this announcement, analyst types had also talked about AMD’s getting someone from outside the chip industry, possibly in the PC industry.
My first reaction was “Why not a chip guy? What’s wrong with that?”
No offense to Read. But what does a PC company – whether Lenovo, Dell or HP – do these days? How much innovation comes out of your typical contemporary PC company? Do they even design anything these days? Isn’t a successful PC vendor defined nowadays by how thoroughly they beat up chip companies on price? When’s the last time when you saw a knock-your-socks-off PC from Lenovo?
More importantly, what new ideas has Lenovo -- or any PC vendor -- brought to the development of new tablets or mobile phones – other than, once again, triggering a bloodbath for chip suppliers?
OK. I may be biased. But every time I interview an architect who designed the latest SoC or a processor, I am always amazed how much “chip guys” actually know about the system market. They not only actually develop almost a turn-key, full-reference design for a given system, but also push their chip’s limits in size, power and cost, while constantly thinking about its usability in a system.
In a recent conversation on a tablet market with Rajeev Kumar, marketing manager for Freescale’s i.MX product line, Kumar said that at Intel where he used to work, “we used to talk about getting computing to the next billion users.” Kumar said, “Now, we must think about getting computing to the next 10 billion users.”
That comment stuck in my mind.
So, tell me how the “customer point of view” of a PC vendor with a solid record of same-old same-old is going to get AMD to win over the hearts and minds of the next 10 billion customers. Surely, we aren’t just talking about selling more PCs to more Chinese.
I wish Read the best of luck. But let’s not overplay the “customer’s point of view” card.
AMD needs a solid engineering guy who’s done his homework as a thoughtful leader – always thinking about the next big thing in the system’s market beyond PCs -- maybe someone who cut his teeth in the chip business, like Abhi Talwalkar, LSI’s CEO(who was apparently in the running at AMD).
To have Fab or not to have Fab that is the question !
Creative ideas in software, systems or circuit design always produce a better ROI than proprietary access to faster / more transistors through ones own state of the art Fab.
Even Apple with $ 70 billion + in bank is not exactly jumping into Fabs despite having IP issues with Samsung.
On the one hand, PC makers are AMD's customers, As such, a PC person should have a good handle on what's really needed to design a product that PC vendors really need. It never hurts to deeply understand the requirements and issues faced by your customers.
On the other hand, I suspect the issues faced by a processor company are a lot different than those faced by a PC manufacturer. Also. sometimes when someone with customer experience comes into a company, their customer perspective can remain stuck at what it was when they were back there.
That's what I see as the greatest risk. The processor world is changing dramatically. As you said, I haven't seen Lenovo leading the way in any of the new directions that the computing industry is heading. If he stays focused on yesterday, he won't be helping the company.
Spinning off the fabs may have been a good move or maybe not. Fabs are very expensive to run and keep at the bleeding edge. The big companies with fabs look at the fabless companies and envy their ability to make so much profit without the huge investment. I think in makes sense to reduce the fabs like Freescales "fab-lite" policy. But it is painfull to not have any fabs for the big companies.
@Junko: AMD will have plenty of company in the massively high volume non-PC centric computing. I don't necessarily agree that requires an IDM model.
There are many areas where AMD can leverage emerging trends effectively if it can meet the handheld devices power needs. The combined CPU-GPU computing is one such example where if AMD can achieve computing throughput at lower power requirements, it can capture a sizeable market.
In 2011, we are reaching 7Billion population on the Third Planet (also the name of a Khurdish pop group, nice tunes!). I, for one, do not wish to see the day when it reaches 10Billion. Granted the term 'user' may mean living or inanimate use point, having that many people consuming rapidly depleting resources is something is scary!
AMD has several hurdles to get over before Read can be or will be effective. First, AMD is bleeding off a lot of long term talent including top Architecture people. The talent pool is very tight and replacing what has bled off will be very, very difficult. Someone coming from a major chip company might have been able to handle this problem but not someone from a company that has a reputation of delivering low quality/performance, second tier products into the market.
Transforming AMD into a fabless company is a deadly strategic mistake. They can only be as good as what one or two particular foundries can offer. No leapfrogs, no advance in technologies. We all know the end of the road for companies no longer able to innovate.
I too agree with you. AMD is having poor business strategies from long back. Most of the time the top management feels that investing on marketing is waste of money. And also they purposefully discontinued from the hand held business(Mobile and DTV) during the rescission, where Intel is aiming it now. The top people are not having proper vision. Being a former employee of AMD I observed these issues, from the top level people.
I have to agree with you. Bringing the current customer point of view to AMD won't help it, if the customer is a PC manufacturer. Look at what Intel's doing with the Atom to compete with ARM. Growth is in embedded processors, people are predicting 15 billion connected devices by 2015, and the vast majority won't be PCs. It doesn't sound like it's where AMD is looking to grow.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.