This week the semiconductor industry saw two high profile pieces of career news; Anand Chandrasekher (former senior vice president and general manager of Intel's Ultra Mobility Group) was hired by Qualcomm as chief marketing officer, and John Byrne was promoted to chief sales officer at AMD.
I found both moves odd for a variety of reasons.
First, let me address why I think hiring Chandrasekher as CMO makes absolutely no sense for Qualcomm.
Chandrasekher left Intel in disgrace. For over three years he had been given all of Intel's assets with one goal in mind; to develop a viable chip for smartphones and to get customer traction. He wasn’t able to pull that off.
Menlo and then its follow-up Moorestown both had major execution issues and fell flat on their face.
Inside Intel, Chandrasekher was seen as a showman - constantly making promises he couldn’t keep and showing off designs that would never see the light of day with customers. The LG phone hyped and slidewared at CES a few years ago was a prime example of just that.
Chandrasekher was also one of the key figures in Intel’s Nokia/MeeGo debacle- a relationship where Intel invested much and gained little only to be left at the altar two years ago at MWC.
But, no sooner was Chandrasekher unceremoniously outed from Intel, things took a turn for the better. Mike Bell and Dave Whalen took Chandrasekher’s place, and in just nine months, the company managed to build a fully functional reference design smartphone. Not only that, but the design went on to score multiple design wins almost immediately, including a major European carrier (Orange) and deals across India and China.
So, what exactly does Qualcomm think it is getting in Chandrasekher? His track record shows him to be a guy who ran marketing but who couldn't bring to market a viable product to compete with any of the ARM partners over a three year period with unlimited resources at Intel.
True, one could argue that with a long career at Intel and with involvement in Centrino, Chandrasekher knows the PC market, but how critical are these skills to a company that will only begin to dabble in the computing space with Windows 8? It’s puzzling to say the very least.
Perhaps Qualcomm’s logic in hiring Chandrasekher was the thought he would be so eager for redemption after being let go by Intel that he would do anything to get back at his previous employer and prove them wrong. That option can’t be ruled out, and people should be given second chances, but from where I’m sitting, this looks like Qualcomm took a really sloppy second from Intel.
Next, let’s examine AMD’s choice of John Byrne for head of sales. This is also an odd personnel choice, given AMD's relative lack of sales success over recent years. Byrne has been in sales at AMD/ATI for donkey’s years. Why would the company not look for new blood instead of promoting old guard?
When Rory Read took over as CEO, he gutted virtually every other position of prominence in the company, and yet Byrne not only survived the purge, but now seemingly thrives.
Byrne originally started his association with AMD selling discrete graphics at an ATI subsidiary by the name of Advanced 3D, based in the UK.
He then moved on to more senior roles in ATI sales and then AMD Europe before moving to North America to manage global sales.
While savvy and well respected in the channel, Byrne's experience with MNC OEMs is on the lighter side, and, under his tenure, AMD has failed to take any significant market share in graphics, CPUs or chipsets. That’s why, to me, Byrne's promotion and ascent defies conventional logic, but as with all executive moves at troubled companies there is always more than meets the eye.
you've got to be kidding. in 5 years, what you started back then is terribly obsolete in the market today, and most likely the wrong functionality. I'd say more like 18 months, but no way 5 years! Our silicon roadmaps never went out more than 2 years in terms of chips in design.
Anand was and is a great sales person and marketing champion. He put everything he had into selling Intel and the world on ultramobility (see his presentations years before ipad). On the downside, the execution of the team was poor on both chips and platforms and he was repeatly surprised to see programs that missed all milestones for 2 years, miss launch dates as well.... not good. Sounds like he has the perfect role for him.
@32 nm with hundreds of millions of transistors in a typical SOC it takes 5+ years to release a chip. So there is no way that in 9 months Mike Bell could have done that. Surely what we see now is the work of Anand Chandrasekher.
"His track record shows him to be a guy who ran marketing but who couldn't bring to market a viable product to compete with any of the ARM partners over a three year period with unlimited resources at Intel."
How do you know he had unlimited resources at Intel? If you believe he had then you have to answer the question - why would Intel throw unlimited resources at such a low margin business when they still have high margins on their existing business? Intel's use of an external 32nm foundry for mobile tends to indicate their heart isn't in this business - after all they _are_ the foundry leaders.
With Sales I feel AMD looks at this through Rose coloured glasses, they often think their s dont stink and wont price their products accordingly to address market realities. THey price some CPU/Platform too high or too low. GPU has typically been fine.
People often say Bulldozer is a flop blah blah. This is incorrect assertion. the high end part is priced a bit too high for its price/performance mix, the lower performance parts are priced much better.
Look what happened with llano. Yields were poor at start and 12months later they have way too much ultra enthusiast parts that they cant sell due to platform changeover to trinity of which its not fully compatible with . AMD Gave away at least 2200 3870K processors as sample in the past months, both at comiccon and quakecon.Who knows how many more they have sitting in warehouses around the world. I think this is unprecedented as has the years have passed vendors have been more frugal with their sampling programs.
Is this the right way to sell CPUs? They flooded the market with $120 Enthusiast chips yet the cheaper versions have had supply issues. They likely dont want to bin the good parts down as they will loose even more money by having to sell at lower pricepoint
Whens the last time you have heard of Intel CPUs being overstocked? Thy rarely change their pricepoints and only discount product lines when EOL is on the horizon or faster bins are introduced.Now its Thunderbolt. It has been a very long time since AMD overed a $1000 CPU. You do not see Intel giving out thousands of game codes either.
As for Intel and Nokia
Nokia was doing most of the heavy lifting with this in terms of developers and users from what I saw
Intel staff were still proud of the colabration as late as Q4 last year (in person) yet I dont think they really knew what it was about
Intel got out when being related to a dead OS seemed embarrasing
THe irony was Meego was working and ready x86 mobile OS that only needed consumer/commercial qualilty UI refinements and growing its ecosystem, something which Intel had.
Qt would have been a useful tool to Intel but they put that fish back into the water. They have Havok, what did they do with it?
Meego is/was still the only true multiasking mobile OS?.Going foward with better mobile SoC this is something important.
Intel has alaways been a software agnostic company: Windows,DOS,OS2 (!),Linux/Unix,MacOS,RTOS anything that ran on their platforms. Their GPU drivers are opensauce
Xscale was probably the easiest way for them enter a new market.
That LG brick is burned into retina of many people. Ive never seen such an awful smartphone as that protoype.So large.
Even with Intels current phone they know its not competitive with high end superphones and are only pushing it into emerging markets. UK is the odd one in this.
They have created market segments when it suited their goals.USB, then grew the ethernet market with an ecosystem. The irony is the hubs/switches they sold used Broadcom and were likely designed by an ODM. Remeber when ethernet was not even standard in PCs let alone notebooks? 3com got out of clients 10yrs ago now.
Then they grew Gigabit and then wireless with Centrino and I can prove that they lost their edge with wifi, the entry level of the notebook market doesnt use Intel WiFi. Now its thunderbolt
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.