"AMD Moves Away From PCs Amid Steep Losses"

[blastingcap]

I suspect the tough questions do not generally occur on conference calls, but in private conversations before or after... so don't read too much into things

 

[mrmt]

I suspect the tough questions do not generally occur on conference calls, but in private conversations before or after... so don't read too much into things

In private conversations you have a bigger latitude in your answer and questions, but public statements that makes the executives liable for that are a lot more important than private talks that nobody listened or recorded, and that will be simply "forgotten" in case the predictions listed there were wrong.

 

[mrmt]

There are a lot more rules and regulations around giving guidance as a company officer. So I wouldn't chalk up their carefully chosen wording to being "incompetents".

Could you please show us what rules or regulations prevent you to commit yourself to a four quarters revenue/margin forecast?

 

[Vesku]

Could you please show us what rules or regulations prevent you to commit yourself to a four quarters revenue/margin forecast?

"In addition, penalties for fraudulent financial activity are much more severe. Also, SOX increased the independence of the outside auditors who review the accuracy of corporate financial statements, and increased the oversight role of boards of directors.[1]"


http://en.wikipedia.org/wiki/Sarbanes–Oxley_Act

"Proponents of the measure say that SOX has been a "godsend" for improving the confidence of fund managers and other investors with regard to the veracity of corporate financial statements.[4]"

Less but more reliable information.

 

[mrmt]

"In addition, penalties for fraudulent financial activity are much more severe. Also, SOX increased the independence of the outside auditors who review the accuracy of corporate financial statements, and increased the oversight role of boards of directors.[1]"

So what? Among other things SOX says you cannot take your numbers out of thin air and throw them in a financial statement without disclaiming them for what they are, and that you also cannot make forecasts, evaluations or disclosures based on data or facts *known* to be patently false.

This has nothing to do with giving your best estimate that came from your corporate financial model using your validated data available, and this is what AMD managers refused to do.

Less but more reliable information.

SOX means *less* but more reliable information? Have you been living under a rock in the last ten years? SOX mean A LOT more of information in the disclosures. And I mean A LOT more. Today we see miles-long disclaimers about the financial data presented and 100 billion companies disclosing 25 million liabilities in minute details. Banks fill 300 pages long quarterly financial statements today. Every company has dedicated SOX committees and they all spend a lot of time and money in compliance structure to attend SOX demands of disclosure or control.

Want a practical example? SOX would prevent exactly the kind of statement you made, "less but more reliable information", without an extensive disclaimer that this is *your* opinion based on your common-sense view coupled with a diagonal reading of an article at Wikipedia, and that you formed this opinion without reading the entire text of the SOX act. But if you put an disclaimer like this, there wouldn't be any problems in publishing it in your financial statements or declaring it in the EC.

 

[Idontcare]

He did disclose margins, he just wasn't explicit, but he gave everything you need to calculate margins. He didn't *commit* himself to that margin, and this is where the problem lies, as the target margins is a very low target, even lower than their current margins *without* the inventory markdown. That he would not commit himself to those "easy" numbers makes everyone wonders that AMD knows things have a good change of get a lot worse next year.

This is the problem for AMD in a nutshell.

Wallstreet, and analysts, don't wait for AMD execs to officially confirm what the analysts can compute for themselves, that is why they are called CC analysts versus mere CC listeners.

The problem for AMD is that the CEO and the CFO painted a rather dire and bleak future for AMD, one that the analysts would have little choice but to point out in their subsequent analyses of AMD as communicated to their own readers...unless the CEO and CFO took the opportunity that the analysts were attempting to provide them to further clarify if the future really was anticipated to be as dire and bleak as the CEO and CFO seem to be assuming it will be.

The CEO and CFO set expectations, which at first pass appear rather bad. The analysts wanted to have those expectations refuted, otherwise they'd have no choice but to communicate to their readers to put AMD on a sell list. The CEO and CFO did not refute the concerns they generated when they set expectations as low as they did.

And that is the issue for AMD stock and anyone who is long or looking to be long on the stock. Why set expectations so low if you do not expect them to turn out that way? And why back away from talking about the expectations you have set once it is obvious how the analysts are interpreting the ramifications of the expectations you have communicated as CEO and CFO?

That is when you cross the transition as an analyst of not only having an eroding confidence in the underlying financials of the company but also you begin to have an erosion in confidence of the people running the company.

 

[Vesku]

So what? Among other things SOX says you cannot take your numbers out of thin air and throw them in a financial statement without disclaiming them for what they are, and that you also cannot make forecasts, evaluations or disclosures based on data or facts *known* to be patently false.

This has nothing to do with giving your best estimate that came from your corporate financial model using your validated data available, and this is what AMD managers refused to do.



SOX means *less* but more reliable information? Have you been living under a rock in the last ten years? SOX mean A LOT more of information in the disclosures. And I mean A LOT more. Today we see miles-long disclaimers about the financial data presented and 100 billion companies disclosing 25 million liabilities in minute details. Banks fill 300 pages long quarterly financial statements today. Every company has dedicated SOX committees and they all spend a lot of time and money in compliance structure to attend SOX demands of disclosure or control.

Want a practical example? SOX would prevent exactly the kind of statement you made, "less but more reliable information", without an extensive disclaimer that this is *your* opinion based on your common-sense view coupled with a diagonal reading of an article at Wikipedia, and that you formed this opinion without reading the entire text of the SOX act. But if you put an disclaimer like this, there wouldn't be any problems in publishing it in your financial statements or declaring it in the EC.

Less but more reliable as in company executives aren't going to publicly say anything without the required research and compliance examination. I wouldn't have even made my original post considering the amount of compliance work that you are pointing out would be required.

In regards to the AMD announcement: They are embarking on a restructuring in the middle of a slowdown, as much as it irks analysts I don't see how they could make reliable statements on revenue/margins upon completion when they are at such a preliminary stage. It is about as bleak as the market is perceiving, all indications are that AMD is expecting further cuts as a distinct possibility.

 

[mrmt]

And that is the issue for AMD stock and anyone who is long or looking to be long on the stock. Why set expectations so low if you do not expect them to turn out that way? And why back away from talking about the expectations you have set once it is obvious how the analysts are interpreting the ramifications of the expectations you have communicated as CEO and CFO?

One of the rumors that floated around was that AMD did indeed consider deeper cuts on its workforce, ranging from 20% to a whopping 30%. As usual, you should take these with a mountain of salt, but it is always worth remember them.

When they refused to acknowledge breakeven at Q3 with 1.3 and settled for 15% of the workforce, I had a gut feeling that the rumor was right on the spot, that AMD did indeed consider more aggressive cuts and that breakeven at Q3 isn't something that they are confident they will reach. The 1.3 billion isn't the best estimate the executives have for revenues, but a compromise between cutting down costs to adapt to smaller revenues and keep engineering alive because of a BoD decision, and that the executive team does not think they can reach these goals.

The fact they had to repeat at the EC that they still are an engineering company smells to me that they were considering forfeiting one of their markets and/or somehow floated inside the company the possibility of becoming an IP company in a not-so distant future.

 

[mrmt]

Less but more reliable as in company executives aren't going to publicly say anything without the required research and compliance examination.

With SOX they must both disclose a lot of things *and* explain how they got the numbers. Plus an auditor must say the input data is ok. Look at an EC from multiple sectors, look at how much information is common. Sales, inventory, cash position, debt, expenses. All that those subjects will be treated at the EC but with different levels of details.

But SOX does not apply to this disccusion, because SOX is more focused on past and current facts, things you know now that impacted or will impact your future. Forecasts are obviously beyond the scope of regulation, except when they are made with current facts that are known to be false.

There is nothing on SOX saying that I cannot predict X% of growth increase or Y% in cost increase. What I can't do is to predict X% in growth increase knowing that my biggest client formally notified me last week that he will terminate the contract next month.

 

[Vesku]

And how exactly do I provide guidance if internally I know that there may be more layoffs without having to reveal to you those layoff plans and the internal forecasts that backs them? If I don't want to bring that up it's safest to avoid commenting in that information arena altogether.

I think letting analysts fill in the blanks is better than saying something along the lines of "oh and in our conservative projections of the next few quarters we'll be slashing another 10-15% of the workforce 3rd or 4th Quarter 2013."

 

[mrmt]

And how exactly do I provide guidance if internally I know that there may be more layoffs without having to reveal to you those layoff plans and the internal forecasts that backs them? If I don't want to bring that up it's safest to avoid commenting in that information arena altogether.

You don't give guidance at all, or at least you give a very conservative estimate, something you know you can deliver, the kind of thing that the markets will push you down today but will generate a lot of credibility in the future when you deliver the target.

 

[Idontcare]

And how exactly do I provide guidance if internally I know that there may be more layoffs without having to reveal to you those layoff plans and the internal forecasts that backs them? If I don't want to bring that up it's safest to avoid commenting in that information arena altogether.

I think letting analysts fill in the blanks is better than saying something along the lines of "oh and in our conservative projections of the next few quarters we'll be slashing another 10-15% of the workforce 3rd or 4th Quarter 2013."

Your points have excellent merit, but I would not have realized it or recognized it as such if it weren't for my personal employment experiences in real-life mirroring what you are getting at. (in other words my brain took a while to process this POV and I probably am still missing a key point or two)


Your last sentence, the one with the paraphrasing quotation - it suddenly occurs to me that this is exactly what must happen IF AMD wants two things - (1) they want to finish tapeout of steamroller and possibly excavator, and (2) they have no strategic need to keep the steamroller/excavator design engineers on payroll once they get those designs into production because the future of AMD is not power-hungry fat cores targeting the desktop but rather the future is low-power jaguar derivatives (which is entirely handled by a completely separate design team).

This happened to me at Texas Instruments. I was formally assigned to the 65nm and 32nm teams while in parallel we had a 45nm team that operated entirely independent of the team I was on. The 45nm team was working on 45nm while the 65nm team was putting the final touches on 65nm for production ramp.

It turned out that TI had no intention whatsoever of fully developing its own 45nm, they had long since already decided 45nm would be entirely outsourced to the foundries. But how do you keep your 65nm team contently working on finishing up 65nm on time, something you are banking on happening at the corporate level, if that 65nm team should find out they have no future at TI for 32nm because 45nm itself was squashed and everyone let go?

The answer for TI management was to absorb the short-term cost hit of keeping the 45nm team employed, creating a facade of a future for CMOS development, just long enough such that the 65nm team finishes the job. Then you let everyone go all at the same time, both teams.

And that is what TI did in the spring of 2007. They retained a total of 6 process development engineers from thee 65/32nm team and redeployed them to work on 130nm analog (I was one of the lucky six).

So now you got me thinking that this is exactly the challenge Rory knows he is facing, he knows he needs to cut headcount by 20-30%, but 10-15% of the people he intends to let go can't be told now that they are planning to be laid off as soon as they finish steamroller or excavator tape-out. Because if those folks did find that out then they'd start hunting for jobs now and get no real work done in the meantime.

But in the meantime Rory and company have every expectation of coming back in two or three quarters and announcing a layoff of the folks that AMD no longer needs - the one's who work on projects that are on the left-hand side of the following slide:

I'll admit it is a rather cynical perspective that I am taking here, but I would not have considered it plausible were it not for the fact I've lived through something very analogous as well as the fact that the CEO and CFO can't be stupid, they must be intentionally playing a poker game here with their employees and that results in them having no choice but to play poker with the analysts at the same time.

 

[Vesku]

That's a plausible scenario Idontcare, keep an illusion that 50% of the company focus will continue to be performance desktop/server while actually planning to refocus completely after Steamroller is launched. Retention of key staff needed to complete current projects must be a concern.

Not something I was directly thinking, but it would be a reason to not want to go into more detail regarding projections.

Heck even if 2013 cuts are from a different section of the company, the worry of not knowing who was on the chopping block for next year would impact productivity.

 

[mrmt]

I'll admit it is a rather cynical perspective that I am taking here, but I would not have considered it plausible were it not for the fact I've lived through something very analogous as well as the fact that the CEO and CFO can't be stupid, they must be intentionally playing a poker game here with their employees and that results in them having no choice but to play poker with the analysts at the same time.

You may have a point here regardind the workforce of steamroller and excavator. If they shave another 15% of their workforce their OPEX goes to about 400 million per quarter, which at the expected gross margins of 35% gives us a break even point around 1.1 billion in sales per quarter, much more alligned with the Q4 revenue prediction than that 1.3 billion.

But costs are something you have far more control than revenues, and it is in the later that AMD problems are. They cannot spend more than one year bleeding over 100 million per quarter and they somehow must still gauge how big is the market for those new chips of them. Will those new revenue streams be enough to keep the company afloat and do they have the cash position strong enough to cover this transition? The lack of disclosure in revenues means that they are not confident enough for predictions, and this is what bites the stock. Much, much more than whether they will cut 15 or 30% of their workforce.

 

[podspi]

I really would hate AMD to pull out of the high-performance market. One thing that bothers me though, is that while they have a clear disadvantage due to their process lag, doesn't this hurt them MORE in the low-power low(er)-performance segment?

My understanding is that the power advantages of the newer nodes decrease as clock-speeds increase, and clock speeds are generally higher on high(er)-performance products.

Is the logic behind this that AMD is willing to accept lower margins than Intel, and so will be able to compete with them that way? This is not even mentioning ARM...

 

[Phynaz]

Wasn't spinning off the fabs supposed to bring AMD's break even point down to $1B? And now they are saying they need $1.3B?

ASP's must be far below earlier projections.

 

[Idontcare]

Is the logic behind this that AMD is willing to accept lower margins than Intel, and so will be able to compete with them that way? This is not even mentioning ARM...

That is the logic that has enabled AMD to exist as a competitor to Intel for 20 some years. As a business they are non-profit over the sum total of their existence. Every dollar of profit they have ever made has been wiped out by losses at one point or another.

So how do you operate at a deficit of 1/6 the resources that your competitor has? You hope your competitor is distracted with something other than your demise...and for AMD they have had the good fortune that Intel's decision makers are perpetually distracted with a goal of hitting or exceeding 60% gross margins.

This is what left AMD the opportunity to spend less on R&D, sell their products at what is essentially break-even for 20yrs, with much lower gross margins than Intel.

And they can continue to do that for the foreseeable future provided the desktop segment doesn't evaporate. And while it won't evaporate, desktop is not the growth story for the first half of the 21st century, so AMD management has no choice but to chase after the growth story if they have any hope of generating uplift in the stock's P/E ratio and increase that stock price while continuing to operate with low margins.

But access to leading-edge process nodes is not the end-all be-all story when it comes to margins. Look at Qualcomm and Apple. They do just fine without access to Intel's fab technology to give them a competitive edge over their competition.

In fact, if you think about it, only Intel appears to need the crutch of process node advantages in order to hit their gross margin targets. Management at pretty much every other semico company achieves the same without the expense and crutch of the technological bleeding edge.

 

[Phynaz]

I really would hate AMD to pull out of the high-performance market.

They already have. It's been at least two years that they have been saying they will not be designing to the high end.

 

[Idontcare]

ASP's must be far below earlier projections.

Definitely. You don't find yourself knocking $100m off the valuation of existing CPU inventory unless the reality is that you have no hope of selling that inventory at the ASP you were targeting when you assigned value to the inventory.

Price pressures must be ridiculous right now with the slowing global picture. Where do you stuff millions of chips if nobody wants them at any price?

 

[Genx87]

I wonder if Intel seeing the rise of ARM thinks it is time to end AMD in preperation for the fight of its own life? Consolidate the x86 market and try to fight ARM in the mobile market. Anti-Trust case non-existent due to change in market conditions.

 

[mrmt]

Wasn't spinning off the fabs supposed to bring AMD's break even point down to $1B? And now they are saying they need $1.3B?

No. The spin off aimed at reduce their cash needs, not change break even point.

 

[moonbogg]

If AMD stops making desktop CPUs at some point, do you think they would be replaced by something like ARM to compete with intel on the desktop? I know nothing about arm so i'm just wondering.

 

[mrmt]

Definitely. You don't find yourself knocking $100m off the valuation of existing CPU inventory unless the reality is that you have no hope of selling that inventory at the ASP you were targeting when you assigned value to the inventory.

Actually the situation is a bit worse than that.

AMD doesn't assign the value to its inventory based on valuation only. The inventory value is always the smallest value between manufacturing costs and market value (the inventory valuation), so charges like the one we saw in AMD results happen only when the market value of the inventory drops below manufacturing costs.

This means that for the next quarter AMD will be selling Llano below manufacturing costs, at margins close zero, and those sales will compete with Trinity sales. We are also not seeing too many single module parts in the market, and because of that I wouldn't be surprised if a small but significant part of the inventory markdown was comprised of single module Trinity chips.

 

[Genx87]

If AMD stops making desktop CPUs at some point, do you think they would be replaced by something like ARM to compete with intel on the desktop? I know nothing about arm so i'm just wondering.

Nvidia has project denver that is supposed to be a higher performance part. I think ARM can make it into low end laptops. At this point it cant compete with intel beyond that. But one has to ask if the market is moving towards that low end segment.

 

[KompuKare]

Have to say this whole $100 million Llano stock thing is a bit strange and implies complete incompetence.
If they knew they were stock pilling why not reduce the price earlier?
Or if they were just continuing to produce Llano because some OEM had an excess of FM1 boards, why not both reduce the CPU and the board prices?

Of course DIY'er and enthusiasts cannot make up a huge lack of demand from the box shifter channel but Llano's are not such poor performers that they are unsellable. Ok, not comparable to high-end SB or IB but better than a lot of Celerons and Pentiums for people who will never buy a dGPU. What's the one saying: "no such thing as a bad product, just a badly priced one"? Well aside from the media PC niche, Llano and Trinity are largely overpriced (which is what you get when half your die is used for graphics), so a large % discount a few months ago should have shifted a lot more stock and not left them with a $100 million write-off they cannot afford.