Advanced Micro Devices
How can AMD Survive?

Related Links
Discussion Boards

By Roleplayer
July 24, 2009

Posts selected for this feature rarely stand alone. They are usually a part of an ongoing thread, and are out of context when presented here. The material should be read in that light. How are these posts selected? Click here to find out and nominate a post yourself!

Rather than post in one of the existing threads, I thought I'd start a new one. And I apologize for using the phrase "all time low" below when it's possible that back in the 1980s some of these things were worse, such as market share. Just assume I mean at an all time low this century.

Lets face it: AMD's margins are at an all time low. AMD's assets are at an all time low. AMD's market share is above their all time low. AMD clearly overpaid badly to purchase ATI, and the pot of gold at the end of the ATI rainbow, Fusion, is now looking at about 2/3 later than expected. With no serious competition from AMD, Intel can keep all its fabs tweaked for better production (instead of better performance) so Intel's margins don't require some excess demand to slop over to AMD. And demand is not good; one could make a decent argument that Intel's gains were at AMD's expense, not that the economy is improving for semis.

There have been bad times this century, and good, and in none of them has AMD been so bad off compared to Intel. We all had hopes for Athlon, the copper interconnects. When Northwood was kicking TB's butt, we all had hope for Opteron. But what can we hope for now?

So as an investor, I have to seriously look at whether it's time to sell my AMD. It all becomes capital loss, so there's some tax value. But I'm inclined not to sell, simply because if AMD reaches profitability, then they stock price is probably worth 5-7 times what it is today. So what would enable AMD to survive?

1) If the economy picks up, then it's likely that AMD will get more charity business. What do I call charity business? I think AMD gets spare orders from HP, Dell, etc. when they have some excess money to invest. But when their sales are so bad that every sale could be called a survival rather than a charity sale, I think they are going to Intel. In bad economic times, the value of keeping AMD as a second source isn't worth the risk of losing a sale. So, Mr. Crystal Ball, what do I think of the economy? Sadly, I think it's bad, perhaps historically bad (worse than the great depression) and I don't see any way it will get better soon.

2) There is this lawsuit. Okay, lets get unreasonably insanely pro-AMD in the lawsuit. How many here don't think that $10B would be a great number for AMD? So lets say AMD gets $10B. That's, what, one year worth of R&D for Intel? So for one year AMD gets to spend as much as Intel on becoming a better developer. The next year, back to the back of the bus. Suppose AMD husbands their settlement. Say, $3B goes to pay off debt. $7B goes into investments. At 3% (can you really count on more than that these days?) that's $210M, or $52M/quarter. As Wayne pointed out, that plus debt payoff *may* put AMD into the black. And may not. So the lawsuit could be excellent, but it's not a game changer. The most credit I'll give AMD's lawsuit against Intel is the possibility of up to maybe a decade's worth of extra savings before AMD finally goes under. And that's the most.

3) Fusion will save AMD. Or will it? Fusion's prospects were based on two considerations. One is that selling a fully integrated CPU/GPU combo could give 100% of the functionality for much less cost. My estimates at the time were AMD would be able to undersell the mobo/GPU/CPU chipset combo from Intel by about $50 while still making about $50 and matching performance. That's nice, but it's still only $50 per Fusion device. The other is that the CPU/GPU combo would be able to produce much better performance at the same cost. I'm still unpersuaded. However, if you want to get optimistic about AMD, Fusion is the place to place your hopes. If AMD's integrated package substantially outperforms Intel's or Nvidia's/Intels, then AMD could make about $100/package. That will be enough to put AMD back into the black. So what lasers will AMD need for Fusion? 32 nm. No, AMD is not yet making anything at 32 nm. Neither is Intel in any volume. So we're hoping that 18 months from now a new process enables a new product that will return AMD to profitability. I'd have to say that seems fairly optimistic of me, to count on Fusion working for AMD.

4) GF takes off. I *could* see this happening. TSMC's performance has not generated a lot of loyalty. If GF starts turning a profit, theoretically AMD earns half. How much profit could GF make? Probably not enough. ATI + NV + AMD + IBM could easily life GF into the black. But will it lift them to $500M/quarter, roughly what AMD needs right now to break even (half of that is $250M)? And IBM won't transfer to GF for at least 2-3 years. So GF could help AMD, but won't help them any time soon.

5) Something else happens. Okay, lets be fair here. AMD and IBM and GF have their own products in the pipeline. Maybe AMD and EUV will have some huge advantages over what Intel will be doing. Maybe AMD's high-k materials will work better, the way SOI and copper helped AMD gain an earlier lead on Intel. And maybe hamsters will sing opera. Yes, it could happen. But my imagination is not fired up, the way it was with Athlon and Opteron.

Once AMD was a chimp. Unlike it's larger and more ponderous competitor, AMD could use rapid responses to market conditions to make profits before Intel developed markets. But AMD got larger, built more fabs, got into more sectors, built server CPUs, bought ATI. In response Intel has become a broad behemoth. They invest in everything, all the time. When the chimp attempts to jump into something new, like netbooks, they find the gorilla is already there. And yes, I *think* Intel has finally decided to put AMD out of business, pay off AMD's shareholders $20B or so afterward, and then never have to worry about AMD again. AMD hurt Intel too much, and all the fab selling back tracking isn't going to suffice.

Put it all together, and I think it's time to get out of AMD. They had a good run. They might find another way to survive. But I cannot recommend buying AMD even as a pure lawsuit play.