ATI - massive job cuts?

Two words: administrative & Canada

Try $35-55k range per.

 

Since NV will be showing off T65 for AMD, will they dare pairing it with G80 ?

 

What do you mean by administrative. First, ATI has an office in Santa Clara within walking distance of my house. There's no way people there earn $35-55k, they couldn't afford to live. But even so, $35-55k would just mean 500-750 employees then, which makes it worse.. I'm not sure ATI can avoid US and California WARN laws doing business here.

If you're trying to suggest that health care costs make up the majority of the overhead for employees, they do not. The loaded cost of white collar employees is typically twice their salary. My health insurance premiums are 2% of my wages. Big ticket expense items come in terms of payroll taxes, vacation, retirement plans/pensions, operations, and management.

 

yes good old Inq covers its ass everytime, they can't go wrong!

 

http://www.theinquirer.net/default.aspx?article=35330

"No job cuts at ATI says AMD"

Even better, now from 1,200 to few hundreds to zero.

 

Well, that's job cuts at AMD now !

 

http://www.theinquirer.net/default.aspx?article=35791

130 in Toronto. . . which seems more in line with original expectations.

 

I think ATI will be a shadow of itself in 18 months.

After there is one failed product cycle under the AMD umbrella there will be a "strategic re-alignment".

 

You're into bold statements and predictions, aren't you? While a bit provocative, I'm interested in that angle myself (although I obviously hope it won't happen), since it obviously would have pretty big repercussions on the industry.

From my point of view, AMD is extremely vulnerable right now - in every single of their markets (except perhaps CE). At the same time, given their current positioning, if they manage to execute perfectly for a change, and that the products are roughly competitive, they could suddenly be in a much more enviable position. Call me skeptical for now, though.

If we look at K8L, Penryn might be a very serious competitor. In fact, I'd tend to believe Penryn will beat the dual-core K8L variant for desktop/gaming purposes, so I ponder how much K8L can really help things outside the server market. Furthermore, 45nm high-k could help Intel in the notebook market.

And then there's yet another dynamic that's going to play out: currently, more than 65% of AMD's shipments are single-core. Right now, those products beat the fuck out of its Netburst competitors. But what is AMD going to do when Intel introduces Conroe-L in Q2, and phases out Netburst in Q3? They'll suddenly be competing against a much more impressive line-up of products in all market segments, including the value segment. That isn't going to help margins, either. So, either K8L beats Conroe/Penryn/Conroe-L in pure performance *and* perf/$ at acceptable margins, or that business is likely going to begin bleeding soon, imo.

Chipsets are safe in terms of revenue, and they're mostly a way to bundle etc. - but I do question how safe they are in terms of margins. AMD is going to compete exclusively with two-chips offerings against NVIDIA's single-chip IGPs and mid-end motherboards. AMD's chipset roadmap does have a few interesting goodies, but even then, it's not a business they're going to make real money out of, especially so given the R&D expenses there.

And then obviously GPUs - they aren't entering the year with a very enviable position, but they're expected to revamp their entire line-up within 4 to 6 months. Given how inefficient R5xx is per mm², it's fairly obvious they'll improve that - but it'll be interesting to see how efficient the design will be compared to G8x.

If AMD can get RV630/RV610 ready in time for Santa Rosa and they're sufficiently efficient per watt, they could become much more competitive in the laptop market again, and that alone would probably put them in an interesting position. Considering AMD apparently is going for 65nm there, and NVIDIA is going for 80nm, that could have some effect on power efficiency.

So at this point, I think a lot will depend on how good AMD executes. If K8L gets beaten and R6xx isn't good enough in terms of performance *AND* derivatives (of either K8L or R6xx) somehow get delayed further... Then AMD is going to be in a sea of trouble, definitely. I think it'll be fairly apparent whether they're poised for great success, catastrophic failure or something in-between by year's end - for now, I'm certainly not willing to place any bets.


Uttar

 

Hector and Dirk can't afford to pay $5B for a company and then gut it. Certainly not that quickly. If (note conditional) ex-ATI gets gutted, it'll be by their successors after the shareholders and board have kicked those two out for botching a $5B play. . . and that probably takes more than 18 months to get to that point, even under worst case scenarios (which I"m sure we'd ALL like to see avoided. Right?)

 

Bingo. I was thinking of wolf2's comment as indicating that they'd focus less on high-end GPU sutff then, though - not that they'd kill the entire division.

heh, indeed! Even abandonning high-end stuff is unlikely to happen anytime soon, IMO.


Uttar

 

Combining that with your statement, if AMD was interested in just entering the graphics market, they would've just gone for S3, they didn't pay all that for just they chip set division. It's like buying a Ferrari because you really want to see it's rims on you Neon.

 

First let me say Utar et al, you know far more about Intel / AMD's respective CPU roadmaps and their relative competitiveness than me. My comment is made without so much knowledge.

AMD has made a strategic decision (to buy ATI), but strategy and tactics are often very different animals. When AMD made their decision, they assumed that ATI would pay its way in terms of its own payroll and probably at least cover a good bit of the cost of the incurred $5B debt.

What AMD lost sight of are the historic practicalities of an acquisition: defocused management, loss of key employees (even at mid-management level), diluted branding, increased difficulty in recruiting etc. In the near term, it is predictable that ATI will lose at least a half a product cycle and perhaps more relative to Nvidia and this is the basis of my opening statement.

Now, when you look at longer term you have other issues. The vision of ATI/AMD together is etheral. It is called 'fusion'. It is an attempt to come in at the high-end and over time let it trickle down to everyday PC's. It is a startup strategy.

I could get long winded on the problems with this but for now let me simply say that high-volume commodity manufacturers simply don't re-model themselves as startups and keep their current shareholders happy. Shareholders are shareholders because they've bought into a model of margin, revenue growth, market. Shareholders do not expect commodity manufacturers to buy startups at 1/2 their marketcap to pursue a niche high-end trickle-down product strategy that is unproven and may or may not work.

I'm not sure I've communicated the problems that I believe AMD has. It is both near term in terms of very pratical product cycles vis-a-vis NVidia, and long term in terms of the practicalities of a high-volume commodity manufacturer hoping a high-end product entry will trickle-down successfully and emerge as the new volume product in 2-5 years.

Any thoughts are welcome.

 

I would tend to believe that, yes. AMD paid $4.2B in cash (+charges etc.) and that's far from negligible even at low interest rates. Looking at AMD's Q3, they had interest income of of $31M and interest expense of $18M. In Q4, they had interest income of $21M and interest expense of $67M. Assuming all of this is due to the ATI acquisition (which it likely isn't, but still), that's about a $60M difference per quarter, or $240M per year - for a business that's currently losing money.

These are quite interesting points. I think the effects of defocused management etc. are very debatable (at least, in terms of magnitude) - but those two don't get mentionned as much, and certainly hold some truth. If I have to be the devil's advocate regarding recruiting (assuming you're thinking of stock options here), I'd say that NVIDIA's P/E is currently extremely high, so that they too are at some risk here. In practice, however, future employees are unlikely to judge the company that way. So, in theory, NVIDIA has an advantage there against nearly everyone else in the industry.

The roadmaps I'm seeing don't place RV630/RV610 much behind G84/G86. At the same time, the same was true of R600, so we'll see what actually happens in practice.

I'll admit to be a tad confused by that statement. The Fusion project is aimed at the ultra-low-end part of the market, according to AMD. Are you thinking of it as scaling down a high-end architecture down, or am I missing something?

No matter what you meant above, that is an interesting point. Fusion is aimed at the kind of market that startups are traditionally interested in. In fact, Fusion is pretty much a direct competitor, from my perspective, to VIA's C7 CPU and related chipsets. And as everyone knows, VIA's CPU division isn't exactly the biggest in the industry. Thus, I ponder how big that market opportunity really is. And as soon as the ultra-low-end segment becomes highly competitive, margins will sink. Especially so if you are always one node behind your biggest competitor, in the most price and power-sensitive market of them all.

Anyhow, I think you're assuming a tad too much that ATI is at least partially screwed this GPU cycle. I'm reserving my judgement on that until G84/G86/RV630/RV610 are out. As for K8L, I'm VERY skeptical they'll be able to beat Intel in anything but servers (thanks to their interconnect advantage).

So anyhow, I'm just going to reinsist on what I said earlier: at this point, it's ALL about execution. If RV6xx and K8L are both competitive and on time, AMD might in for gold. Otherwise, they're in for bigtime trouble. This is why I say AMD is especially fragile. Longer-term, it will be very interesting to watch what NVIDIA and Intel are going to do, since neither really "made their move" officially yet.


Uttar

 

This is an intersting thread. As Uttar stated the only market where AMD clearly has the technology advantage is in single cores when compared to Netburst. This is the lowest margin part of their business and even it is not safe once Intel migrates Conroe down the food chain.

AMD has basically been backed into a corner and MUST pull a rabbit out of their hat to avoid complete disaster. Unless K8L is smashing success then AMD is screwed (burning cash at a rate that forces them to raise capital). I see them having a slight lead over Intel with 65 nm K8L in servers until 45 nm from Intel comes in the form of Penryn. At this point, with the rumored HT and technology breakthroughs that Intel has at 45 nm, AMD is going to be likely left behind again. If this is the case than AMD will again be denied the high gross margin performance desktop and server markets. Couple this with Conroe pressuring AMD at the low end AND all the new capacity coming on line and you have a recipe for trouble.

The real problem I see with the ATI aquisition is it does NOTHING to address the fundamental concerns facing AMD right now. These are fab capacity (currently too little soon to be too much), high margin segments market share, retaking the performance lead from Intel (K8L) and closing the manufacturing timeline gap between die shrinks with Intel. Unless AMD gets a handle on these the whole ATI aquisition is moot and only adds problematic issues for AMD. These include negative cash flow, R&D resources, servicing massive debt and possible distractions for management. AMD is teetering on needing an infusion of cash right now and ATI only makes the situation worse. If ATI was profitable with Nvidia type execution and gross margins it would be TOTALLY different. The fact is you have ATI coming off the bottom and AMD bottoming at the same time. That spells RED financials for both companies right now which burns through cash already stretched too thin by aquisition costs and loans.

So why did they buy ATI? Is Fusion the only way to counter Intel's manufacturing advantage? Is Fusion supposed to be the savior like Opteron and X2 were? Did they think they could make ATI profitable again in short order based upon what they saw with R600 and beyond? It seems like AMD has bet the farm on Fusion to counter Intel. They placed the bet at the same time Intel surprised them and took back the performance crowns with Conroe and Core2Duo. Since Fusion is still late 2008 timeframe and God knows what Intel will do between now and then, doesn't that seem like a pretty darn big risk to take? They should have done it back when ATI was profitable, Opteron was the KING and they could have helped with R600 and the launch schedule. Or, the should have waited until after they crossed the first big hurdle which is K8L.

K8L is the trigger than will either keep AMD alive against Intel or force it into VERY difficult decisions as they are foced to raise capital by diluting shares or selling off assets. The only thing the ATI aquisition did, based upon the poor timing, is speed up this conundrum for AMD. Without the ATI debt perhaps AMD could have withstood K8L NOT regaining the lead over Intel. Now AMD is forced to accept that everything rests on either K8L and/or Intel screwing up 45 nm. I would not bet on the latter. As we type, Intel is regaining lost share against AMD, AMD is slashing prices on X2 and they are being forced into an ugly negative cashflow situation. The ONLY thing that can stop this is the aforementioned rabbit named K8L.

Back to ATI...why?
1. AMD is confident ATI's roadmap will retake share from Nvidia and provide much improved margins resulting in positive net cash flow.
2. AMD is confident that bundling opportunites, particularly in mobile, will help them gain share against Intel.
3. AMD is confident that they will eventually be squashed by Intel's superior capacity, process advantage and overall manufacturing might. Since they can't compete tit for tat...they invent a new product called "Fusion" that changes the game and gives them what they most desperatly need much like Opteron did...TIME.

I am willing to bet the answer is all of the above. The problem is 1,2 won't come soon enough and 3 is here much sooner than expected. Again, bad timing all around for AMD and it smacks of a perfect storm to me. What do you need during these times? CASH. What does AMD not have right now? CASH. No cash means no time and means your options are dictated by the accountants which is an ugly way for a technology company to operate. If K8L isn't a clear winner the OEMs and IT managers will not jump ship from Intel and will likely switch FROM AMD once Penryn arrives a few months later.

Uttar is right....anyway you peel this banana K8L is the trump card.

 

Probably not so dire. Big corporations tend not to go down easily. But I do agree with you that it was a big mistake for AMD to buy ATI. ATI gives them nothing that a much smaller company doesn't. It's not like ATI has been keeping up with nVidia at the high end. In fact, they had one great product, the R300, and then flubbed just about every product cycle thereafter. The problem is that nVidia is just a lot more competent that ATI at doing this, with the successful execution of the NV40, G70, and now the G80 line of GPUs. ATI would have been backed into a corner on its own accord, buyout or no buyout, and stood a serious chance of being eliminated from the high-end GPU industry anyways. If this happens, then AMD got themselves nothing but a large albatross that will only lose money. A much smaller company would have given them the same integrated chipset and graphics solution, at a much lower price.

While this won't seriously threaten AMD immediately, it will have shorten the available time period in which they can pull ahead in the CPU industry, since now their finances are so much weaker. I think they'll survive if K8L is unsuccessful, but a failed "K9" will probably be the final straw.

 

The ATI deal raises a lot of questions:

Did AMD management have an adequate feel for the competitive storm that is now occuring when the deal was announced, and really has been in the works since before the deal was announced (Conroe)?

If no, one would have to question the intellectual honesty of management. Although the problems are probably not insurmountable, all projections made to investors look foolish at this point and in retrospect were best-case scenario situations that had very little realistic chance of happening given Conroe and Woodcrest.

If yes, then why would AMD pull the trigger and leave themselves in the financially precarious psotion they are now in? The answer can only be the long-term importance of graphics technology. AMD was riding high, so striking while the iron was hot (share price) certainly would make sense.

Regardless, one wonders why drastic measures to reduce costs were not taken. The Jerry Sanders "real-men own fabs" mentality looks to have carried on in the Ruiz era. Is AMD any more advanced than TSMC or IBM/Chartered? The are spending billions of dollars every year to support fabs that are always lagging. How silly.

 

The share price of AMD the day before the deal was announced on July 24, 2006 was $18.23 so a strong share price didn't factor into things. That was at the bottom of the slide from $42 set in late January. It just seems like the deal was a good 9-12 months late or they should have waited till after K8L. I think Conroe, Woodcrest and Core2Duo surprised them but they were already commited to the ATI deal so they pulled the trigger anyway. Pure speculation on my part at this point.

What I do know is this. If AMD has several very bad quarters in a row they will have to seek additional financing to service their $3.7 BILLION in debt if they remain in a negetive cash flow situation. At that point, I assume their first move would be to sell the remaining stake in Spansion. How much that is worth I am not sure. Beyond that they will have to dilute their shares with a secondary offering that will not exactly command a high price based upon their recent track record.

In any event, AMD basically bet the farm at a really bad time. It will be interesting to see how this plays out. I also predict that Hector Ruiz will step down this year as CEO. Dirk Meyer will replace him.

 

Not a lot of companies will bet the farm during good times.
It's unrealistic to assume that AMD wasn't aware of the coming storm: Core and Core 2 were available long before the merger with ATI and it's not that hard to estimate mid term performance advances if you know your own design and the current state-of-art design of your competitor. AMD has always been behind Intel wrt fab technology, so it was just a matter of time of Intel to rectify their design mistakes of the past and finally come up with a performance kicker.
I think AMD bougt ATI precisely because it expected a storm and felt it help them getting through it. That's a gamble that may or may not pay off, but obviously AMD management thinks (thought?) it will.

 

nVidia could have pulled them through. ATI absolutely will not, and I think that was nearly obvious.