If I were the CEO of AMD... (rate)

[2timer]

Guys, this is how I would engineer a turnaround at AMD, step by step. Please rate on a scale of 1-10.

1. Sell off every piece of AMD not directly related to CPU engineering. Fire almost everyone in marketing, accounting, and middle/upper management.

This includes the Radeon business and the server business. I would restructure AMD so that it's only about CPUs again, nothing else. The sale would hope to raise 3-4 billion dollars in cash.

2. After restructuring, approach a Wall Street investment fund for investment in building a new fab process.

I would offer the investors 12-15% return on their money in a 24 month period. This borrowing would be invested in new R&D and to offset the initial costs of a new, cutting edge fab. Hopefully, 1 billion could be raised.

3. With the infusion of cash from restructuring and Wall Street, pour money into TSX R&D. At the same time, shamelessly poach a handful of key people from Intel.

I would have absolutely no apologies about this. If Vinod Dham was instrumental in AMD's earlier successes, that should be repeated.

Why do I say TSX? Well, let's assume that AMD can make cores cheaper than Intel. Benchmarks show that when it comes to multicore processing on the desktop, AMD is within striking distance of Intel. In single core apps, they are a generation behind.

If AMD were to make TSX it's "bread and butter," then hypothetically, it could sell cheap, 6 and 8 core processors that meet or beat Intel in multicore gaming - for 20% less than Intel.

As someone else on the boards said, "AMD needs to invest in TSX sooner rather than later to stay relevant." Imagine if AMD were to take the fight right to Intel with TSX technology, and combine that tech with more, cheaper cores. Wouldn't that just be a sweet deal to see? AMD could be the giant in multi-threaded apps.

This is how I would turn AMD from being a wash to becoming a geek darling again. I just think they need a leader who is focused entirety on CPU design. AMD needs to become a CPU company again, and shed the extra baggage.

 

[exar333]

Guys, this is how I would engineer a turnaround at AMD, step by step. Please rate on a scale of 1-10.

1. Sell off every piece of AMD not directly related to CPU engineering. Fire almost everyone in marketing, accounting, and middle/upper management.

This includes the Radeon business and the server business. I would restructure AMD so that it's only about CPUs again, nothing else. The sale would hope to raise 3-4 billion dollars in cash.

2. After restructuring, approach a Wall Street investment fund for investment in building a new fab process.

I would offer the investors 12-15% return on their money in a 24 month period. This borrowing would be invested in new R&D and to offset the initial costs of a new, cutting edge fab. Hopefully, 1 billion could be raised.

3. With the infusion of cash from restructuring and Wall Street, pour money into TSX R&D. At the same time, shamelessly poach a handful of key people from Intel.

I would have absolutely no apologies about this. If Vinod Dham was instrumental in AMD's earlier successes, that should be repeated.

Why do I say TSX? Well, let's assume that AMD can make cores cheaper than Intel. Benchmarks show that when it comes to multicore processing on the desktop, AMD is within striking distance of Intel. In single core apps, they are a generation behind.

If AMD were to make TSX it's "bread and butter," then hypothetically, it could sell cheap, 6 and 8 core processors that meet or beat Intel in multicore gaming - for 20% less than Intel.

As someone else on the boards said, "AMD needs to invest in TSX sooner rather than later to stay relevant." Imagine if AMD were to take the fight right to Intel with TSX technology, and combine that tech with more, cheaper cores. Wouldn't that just be a sweet deal to see? AMD could be the giant in multi-threaded apps.

This is how I would turn AMD from being a wash to becoming a geek darling again. I just think they need a leader who is focused entirety on CPU design. AMD needs to become a CPU company again, and shed the extra baggage.

1. Sell off every piece of AMD not directly related to CPU engineering. Fire almost everyone in marketing, accounting, and middle/upper management.
-Haven't they already tried this? They need good leadership in marketing and bringing in a few key peope could turn this around with the current, or much of the current team.

This includes the Radeon business and the server business. I would restructure AMD so that it's only about CPUs again, nothing else. The sale would hope to raise 3-4 billion dollars in cash. -How would it get anywhere close to this much? The whole market cap of AMD is less than $2B...

2. After restructuring, approach a Wall Street investment fund for investment in building a new fab process.

I would offer the investors 12-15% return on their money in a 24 month period. This borrowing would be invested in new R&D and to offset the initial costs of a new, cutting edge fab. Hopefully, 1 billion could be raised. -This likely would not even come close to covering a new fab process. With AMD's current status, they could partner with an existing fab company to customize something, but this would be WAY to cost prohibitive for AMD to take on alone.
3. With the infusion of cash from restructuring and Wall Street, pour money into TSX R&D. At the same time, shamelessly poach a handful of key people from Intel.

 

[sandorski]

1/10 - Stay away from Fabs

 

[nehalem256]

Did you forget the WSA with GF?

 

[Olikan]

1/10...

AMD just needs good marketing, look at what they done in India...

marking it a growth of 570 percent in absolute unit terms over eight quarters.

http://www.crn.in/news/hardware/2013/03/05/amd-gains-20-percent-market-share-in-q42012---idc

 

[Idontcare]

#1 - You would succeed in turning AMD into little more than what Via accomplishes already but at the expense of probably 7000 to 9000 jobs and a market opportunity roughly no larger than what Via currently has access to.

#2 - Offer 12-15% in 24 months when the earliest you would have revenue from your plan is 48-60 months out? Why not be even more silly and offer 24-30% then? No logic to your financial structure here IMO.

#3 - With the senselessness of #1 and #2 creating a veritable stampede of existing talent for the exit door, good luck enticing talent to come in through the entryway. And the shadier you make your efforts, the more your crisis-mode management style reeks of obvious desperation, the less likely you are to snag truly good talent because your business plan raises nothing but red flags to all but the uninitiated (and you claim to want to hire the initiated, so...)

The bottom line for AMD is that they cannot afford to compete with competitors who have more resources than them. So they need to tally up their resources and existing assets (including human talent) and figure out where the small ponds are currently that they could stand a chance of being the big fish should they enter into said small pond.

Becoming an even smaller fish, as your plan would have it, simply to attempt to continue to exist with already vastly larger fish in the same pond is completely irrational, corporate suicide.

 

[2timer]

Wow, AMD's market cap is just 2 billion? Geez.

I just assumed that if they paid 6 billion for ATI in 2006,it would be at least worth half that amount today.

 

[2timer]

1. Sell off every piece of AMD not directly related to CPU engineering. Fire almost everyone in marketing, accounting, and middle/upper management.
-Haven't they already tried this? They need good leadership in marketing and bringing in a few key peope could turn this around with the current, or much of the current team.

This includes the Radeon business and the server business. I would restructure AMD so that it's only about CPUs again, nothing else. The sale would hope to raise 3-4 billion dollars in cash. -How would it get anywhere close to this much? The whole market cap of AMD is less than $2B...

2. After restructuring, approach a Wall Street investment fund for investment in building a new fab process.

I would offer the investors 12-15% return on their money in a 24 month period. This borrowing would be invested in new R&D and to offset the initial costs of a new, cutting edge fab. Hopefully, 1 billion could be raised. -This likely would not even come close to covering a new fab process. With AMD's current status, they could partner with an existing fab company to customize something, but this would be WAY to cost prohibitive for AMD to take on alone.
3. With the infusion of cash from restructuring and Wall Street, pour money into TSX R&D. At the same time, shamelessly poach a handful of key people from Intel.

#1:Good idea. Ok, so let's modify #2 to read "... and *partner* with another fab to build a custom AMD fab."

#2: I don't want to get in a back and forth with you, but I stand by my original point: AMD needs, in my opinion, to become just a CPU company again.

 

[NTMBK]

1/10. Intel is already very damn good at being Intel, and AMD aren't going to succeed by out-Inteling Intel.

 

[Ferzerp]

#2: I don't want to get in a back and forth with you, but I stand by my original point: AMD needs, in my opinion, to become just a CPU company again.

So, you feel that the best chance it has of surviving is to focus on its absolute weakest area?

Well, ok then.

 

[2timer]

#1 - You would succeed in turning AMD into little more than what Via accomplishes already but at the expense of probably 7000 to 9000 jobs and a market opportunity roughly no larger than what Via currently has access to.

#2 - Offer 12-15% in 24 months when the earliest you would have revenue from your plan is 48-60 months out? Why not be even more silly and offer 24-30% then? No logic to your financial structure here IMO.

#3 - With the senselessness of #1 and #2 creating a veritable stampede of existing talent for the exit door, good luck enticing talent to come in through the entryway. And the shadier you make your efforts, the more your crisis-mode management style reeks of obvious desperation, the less likely you are to snag truly good talent because your business plan raises nothing but red flags to all but the uninitiated (and you claim to want to hire the initiated, so...)

The bottom line for AMD is that they cannot afford to compete with competitors who have more resources than them. So they need to tally up their resources and existing assets (including human talent) and figure out where the small ponds are currently that they could stand a chance of being the big fish should they enter into said small pond.

Becoming an even smaller fish, as your plan would have it, simply to attempt to continue to exist with already vastly larger fish in the same pond is completely irrational, corporate suicide.

Good points. I hear your criticisms, and acknowledge my logic is not 100% accurate.

Maybe I'm just too idealistic. I just think if AMD could leverage it's assets somehow, and really invest in TSX, my "high core multi performance" fantasy could come true! Lol

 

[Ferzerp]

I just think if AMD could leverage it's[sic] assets somehow

What assets would those be exactly?

 

[2timer]

So, you feel that the best chance it has of surviving is to focus on its absolute weakest area?

Well, ok then.

If you compare AMD's CPU to AMD's GPU and then say, "AMD's CPUs are it's weakest area" then yes, logically you are correct

Point is, AMD's CPUs are still strong in multi-threaded benchmarks against Intel

 

[2timer]

What assets would those be exactly?

It's Radeon and server assets.

 

[JimKiler]

Guys, this is how I would engineer a turnaround at AMD, step by step. Please rate on a scale of 1-10.

1. Sell off every piece of AMD not directly related to CPU engineering. Fire almost everyone in marketing, accounting, and middle/upper management.

This includes the Radeon business and the server business. I would restructure AMD so that it's only about CPUs again, nothing else. The sale would hope to raise 3-4 billion dollars in cash.

Do you know what accounting does? AMD is a public company. Your suggestion might as well add get rid of corporate systems/IT who needs networks? Hell if there is no accounting to pay the bills or report to the SEC AMD would be dead very quickly.

 

[Elixer]

Heh, the OP's post is a sure fire way to kill AMD faster than what they are doing now.

 

[Sheep221]

AMD did marketing mistakes in the past that are result of the current situation.
Intel didn't make good CPUs most of its existence and still has ten times more either revenue, either employees, etc
AMD compared to Intel, never took care to make long term partnerships with computer manufacturers and bringing effort to complete their demands.
Intel did that with major computer manufacturers and slowly collected money to achieve faster research and frequent upgrades of their FABs.

Your strategy would put AMD in much worse situation than it currently is in.

 

[2timer]

AMD did marketing mistakes in the past that are result of the current situation.
Intel didn't make good CPUs most of its existence and still has ten times more either revenue, either employees, etc
AMD compared to Intel, never took care to make long term partnerships with computer manufacturers and bringing effort to complete their demands.
Intel did that with major computer manufacturers and slowly collected money to achieve faster research and frequent upgrades of their FABs.

Your strategy would put AMD in much worse situation than it currently is in.

Hm, that's a very good point.

If AMD can build a symbiotic relationship with console makers, similar to what Intel did with PC makers in the 2000s, perhaps that will give AMD a strong foothold for stepping into the future.

Wow, lots of factors here I didn't even consider :$

 

[Idontcare]

Good points. I hear your criticisms, and acknowledge my logic is not 100% accurate.

Maybe I'm just too idealistic. I just think if AMD could leverage it's assets somehow, and really invest in TSX, my "high core multi performance" fantasy could come true! Lol

The challenge with that strategy, invest in TSX and go full-steam ahead, is that if AMD could do it then surely Intel could as well.

Just look at x86-64, IMC, multi-core and iGPU/Fusion/APU. Intel is able to beat them to the financial punch every time simply because Intel can out-spend and out-pace AMD in getting to the finish line once the finish line is defined.

AMD's issue is not coming up with news ways to innovate. Their issue is coming up with news way to make profits. You can't conflate the two.

Your proposal is for AMD to come up with news ways to innovate. That is great but that alone does not turn an AMD into an Apple or a Google.

Somewhere in the process of innovating you need someone at the helm who is a true master of business. Someone who is able to leverage product innovation into financial gold.

This is not a "build it and they will come" market. If AMD goes full into TSX and manages to create a market out of it then all they will accomplish is creating a new opportunity for Intel to come in and eat their lunch once again.

AMD needs to find themselves a smaller pond in which they can be the 800lb gorilla. x86 is not that pond unless Oracle decides to buy them (which may happen yet).

 

[AnandThenMan]

OP your plan is corporate suicide.

 

[Ancalagon44]

I actually think AMD's current strategy is a good fit considering where it is. They have landed themselves in a bad position due to poor product execution, but its not too late.

I dont know if they will ever get close to Intel in terms of process node technology, power consumption and single threaded performance, and I think thats their biggest problem. The thing is that Intel competes in nearly every market that AMD does, meaning their process dominance can be felt in mobile phones, tablets, laptops, desktops and servers. Its only consoles and GPUs that AMD is safe in.

They should:
1. Get an ARM architecture license and start doing custom ARM designs, but for higher performance devices only. SeaMicro server with custom AMD ARM CPU = possible win?
2. Stay out of the phone CPU game, for now. Its too crowded.
3. Bring some of that Radeon magic to the AMD CPU division. Ie, good product execution.
4. Its too late to cancel Steamroller, but they should probably cancel Excavator. I just dont think BD was a good idea at all.
5. Put more effort into their software ecosystem to encourage developers to use their stuff. Learn from Nvidia's CUDA.

 

[tweakboy]

Sell the company and move on...... dont let the door hit you in the back LOL

 

[Exophase]

Going to have to jump on the 1/10 bandwagon too, sorry..

1) GPUs are AMD's saving grace, they're the differentiating factor in everything sold below the FX line. AFAIK discrete GPUs are still profitable and I don't think AMD will want to give up the long term guaranteed console revenue stream.
2) Starting a new fab in attempt to be competitive with Intel in this day is pretty much impossible, they just don't have the heritage for it.. the savings they'd get vs manufacturing at another fab and eventually tweaking the process to be more optimized for their CPU designs (has AMD ever doing a lot of this?) are just not even close to worth the immense investments..
3) AMD can't bank on driving software changes on TSX, the software industry will never optimize for AMD. AMD needs to ride Intel's coattails on this one. That means supporting TSX as soon as they can but I don't think it's something that's going to take years of incremental refinement to get good enough.. However their CPU designs have to accommodate it and w/o mandatory inclusive L3 caches they might have a harder time vs Intel.

 

[mrmt]

Guys, this is how I would engineer a turnaround at AMD, step by step. Please rate on a scale of 1-10.

The first thing you must think is what kind of company you want. Because it is already clear that try to beat Intel without enough resources is not feasible and nobody will step in to outspend Intel in their home turf as you are suggesting.

So the first practical question you would have to ask to turn AMD around is: Where would you place AMD on the market?

Rory Read wants to turn it in an embedded chip designer, meaning lower costs, lower margins, lower but steady income. This idea doesn't appeal enthusiasts and isn't how AMD got here, but sounds a much more reasonable plan than going head on with Intel or the ARM manufacturers as some people want.

Your plan is simply more of the same, burn whatever cash you may get in R&D to try to take on Intel. If this strategy didn't work until now, why would it work now?

After you answer the first question, we might get to the second practical question: How much and how long?

Your plan would demand fresh money, 10-15x more money than you are expecting and for a period of time 5-10x longer. Rory's plan just burns AMD cash reserves and assets, nothing that they wouldn't lose if they didn't do anything, so AMD current plan is more feasible than yours.

======

While I'm not really fond of AMD management (I still think they are a bunch of amateur incompetents), it is important to understand some of the constraints that they face. The situation is not easy, and doing a good job and doing a bad job might not make a difference in the end.

AMD needs someone like JHH, someone who can spot and develop market niches like no one, someone who can make business opportunities happen. It's not only about money, it's about vision, direction, things that AMD has been lacking since Hector.

 

[imagoon]

1/10. AMD is not and really never should have been "CPU only." You propose to save the company by cutting off the most stable revenue sources it has and throw it all in to the most risky position they could. 12-15% return is quite often will in to the high risk arena. I don't think most investors would look at AMD as it is now and think "We can get 15% if we throw away the stable revenue stream and use that cash to build CPU's against Intel." Sale of ATI assets would likely be considered a loss because you sell assets to make cash but those assets are producing a return, so you lose that future value to take a very large risk that has a decent chance of bankrupting the company with the history of the competition.

Also Intel is not purely CPU. They fab quite a few other things. They are pretty diverse and could attack what is remaining of AMD if they wanted to after you sell their strongest assets.