Analysis: AMD's technology opiate

Posted by Rick C. Hodgin

AMD made a very powerful presentation yesterday. No question about that. Anyone sitting in the audience or listening in, at least those who didn't have any reason to believe otherwise, is likely to be absolutely convinced about AMD’s vision. The company showed nothing but a solid purpose of mind. It was actually very refreshing to see such an upbeat AMD on that stage. The question now is: Will AMD’s vision be enough?

 

 


In Pictures: AMD's updated roadmap (07/26/2007) ...

Today is tomorrow. I've now had some sleep. Made a few phone calls. Spent some time thinking about the future. And here I sit in a quandary asking myself: “What does it all mean?”

I see AMD's vision is absolutely forward-looking. In fact, they are not directly addressing anything Intel has on the table (or their roadmaps), at least not by addressing Intel directly. AMD is clearly following their own desires which are akin to the industry's needs, wants and wishes. And if it so happens that in the process of meeting those goals they just so happen to compete with Intel along the way, then that's just the way it is. AMD does not seem to be considering Intel at all. Their focus is elsewhere.

Every product AMD outlined was designed to address customers and industry/market needs. They did not have any reactionary product to Intel's arsenal. In fact, it's that driving reality from within AMD bringing us to the big question. It's not even a matter of wanting to be here. We are here.


How can AMD survive the next couple of years?

Intel has roadmap products that should be well positioned in all markets. Based on Barcelona's early numbers, Intel should have no problems competing with AMD. They just need to ramp up the clock. Intel also has a significantly greater manufacturing capacity. Their 45 nm process revision is due out by the end of this year. We'll see even faster Core 2-like products with enhanced logic units which reduce processing time on divides, multiplies, adds, etc. We see the cost advantage. More chips per wafer means a lower cost of manufacture. And then there's Intel's tremendous cash on hand, over $10.6 billion. Intel has money to invest and, if necessary, to burn in deciding what’s next and yes, also, to take an idea from AMd and throw its resources at it, if it’s required.

It's almost like AMD is the Chrysler of the semiconductor industry. In 2006, Fortune magazine rated AMD the most innovative semiconductor company. And here they are today, sitting atop a mountain of debt which must be repaid in 2012, just about the time their planned products should be bringing in real revenue. They have additional debt for the purchase of ATI. They are bleeding cash each quarter like a stuck hog. They have less than $2 billion cash on hand right now. They've indicated they'll only receive another billion or so through the sale of 200 mm processing equipment and income from Spansion. But, during that same period of time they'll have new 300 mm investments which should pretty much eat up all of their remaining cash, and then some. So, how can AMD survive?

 

There are some scenarios I've considered. First, we've heard rumors lately AMD may be interested in going private. If true, they could find an angel investor that believes in their long-term goals (those they outlined yesterday very clearly). They could go private and be funded for the foreseeable future, free to carry out a life free of market entanglements and bands of wanting shareholders.

In another version, they could also continue on as they have and bleed (cash) to death. After going belly-up, a company like IBM or Samsung would come in. They'd pick up AMD for a song, take their advanced IP and begin moving forward with real R&D and manufacturing resources. In my opinion, this would be the best thing that could happen to AMD. The belly-up part would, of course, be brutal. But the truth is AMD's presentation yesterday outlined a tremendous corporate vision. AMD's product lines take AMD to the exact place to offer compute support infrastructures for nearly every major device line that looks to be solid in 2011 and after. IBM and Samsung both have the capacity to take AMD center stage. They could propel their products forward into the markets and make them a real force to be reckoned with in the industry. There's no question about it. AMD is hampered right now by cash flow only. Not by vision.

We also found out today that AMD's ongoing multi-continent anti-trust lawsuit against Intel received a positive boon. The European Commission issued a Statement of Objection which is just a step away from considering Intel guilty of violating three of the four primary paragraphs which outline and describe what anti-trust behavior is in the EU. If a guilty verdict is eventually found, Intel could be facing serious restitution – which eventually could have an impact on AMD’s case against Intel here in the U.S. There could be huge fines (billions of dollars). Also, future injunctions could kick in. These would be punishments to address their previous anti-trust behavior, possibly going on for a number of years.

These would be wide-open windows of opportunity for AMD (and even other potential startups at that time). But, the truth is it will likely be years before that ruling comes down. At the earliest it will likely be near the timeframe where AMD's hardware vision would begin to pay off (after 2010). And what we must concern ourselves with first is: How AMD can make it from here to there?

From everything reported yesterday at the Technology Analyst Day 2007 event, AMD's products position their company very well in 2010 and after. Prior to that timeframe, the roadmap looks very thin. This is especially true when you consider that Intel has so much manufacturing capacity, a six+ month lead on 45 nm migration, a high-k/metal gate solution, so much cash, so many resources, and has now proven that they can effectively respond to threats.

If the early indications of Intel's upcoming 45 nm high-k/metal gate technologies are true, they will reduce heat by a factor of 10. This will allow the chips to run at either much faster clock rates, or with so low power consumption (due to reduced cooling), or both. That will allow Intel to overtake AMD in their stated bread-and-butter areas, performance-per-watt and performance-per-watt-per-dollar.

Barcelona, while being a native quad-core processor, will only address a very niche market need, whereby its true internal power (gained from being native) will be exploited. Intel will be able to compete in all other areas quite well with their own single-socket, quad-core offering (with twin dual-cores inside). While not native quad-core, it will provide similar performance and most likely much greater value. We've seen Intel take the quad-core processor price point today to below $300. It is unlikely it will rise high again.

The overclocking potential of Core 2 has already shown Intel could, today, address anything it's believed Barcelona could bring the table. That does not include Penryn (mobile), Wolfdale (dual desktop) or Yorkfields (quad) in early 2008.

 

 

The outside view

I was also thinking, what if I were Paul Otellini, Intel's CEO? If I was watching the webcast yesterday what would I think? AMD's platform positions appear solid, but they're a few years out. If I were Paul, and I was considering my company's financial position. And then AMD's financial position. And the products we have coming out. And the products they have coming out. I truly think the only question going through my mind would be: “Are they for real? Is this really what they're offering? We need to send someone out there to find out if this is just a ruse.” I know that sounds a little harsh, but I think that is exactly what I would be thinking.

If I were an existing customer I'd be thinking “Yes, it all sounds great. I definitely would look forward to seeing those products. So, how are you going to make them with no cash?” And I think immediately after that meeting I would get on the phone to my board of directors and begin having some talks. And if I were a customer I'd be thinking “This is going to be so great. Competition means more features and lower prices.”

But I'm still back to the fundamental truth. Nothing I've seen addresses the cash flow problem between now and then. I see AMD executing on Barcelona. I see it being well received in the industry, probably keeping AMD about where they are right now. Not gaining, not losing. But I also see Intel releasing a successful onslaught of 45 nm products just a few months after Barcelona's big push. I see their manufacturing kicking in to lower processor prices even further. I see AMD continuing to lose money and I see the IBM or Samsung scenario playing itself out sometime in late 2008.

As someone sitting back analyzing everything from the outside in, there do not seem to be many conclusions to be had. Something amazing could happen, like Barcelona could be launched with an unannounced 24 MB Z-RAM cache, or the Penryn architecture could find a last minute fatal flaw setting Intel back for months. Anything is possible, but given the history of AMD and Intel, I see the reality holding true that nothing truly major or unexpected will happen.

And given that, I see AMD's cash position being their Achilles Heel. Unless they can find a way to fix their cash position, AMD could be in serious trouble before the of the year. It is not far fetched that AMD could see IBM, Samsung or some other lucky firm to swoop in and buys up their valuable IP for pennies on the dollar.