Net Loss

The exact net loss calculations depend on which parts of AMD's business were hit the most (for example, did server marketshare decline? Was GPU revenue stable, or did it decline slightly too? Were handhelds weaker than expected? etc.) but basic and mostly reliable calculations are fairly easy to do, and the numbers they turn out are ugly to say the least.

First, let's just look at operating income excluding one-time items, which thus corresponds to gross profit minus operating expenses. As can be seen in the tables on Page 1, non-GAAP margins were 40% in Q406. Given that CPUs had higher margins than GPUs, and that GPUs will now represent a higher proportion of revenue while CPU ASPs also went down, it seems likely that non-GAAP gross margins of roughly 34% are an optimistic estimate at this point.

Based on that, we can estimate gross profits of about $415M. The problem obivously is that their operating expenses are in the $650M+ range, which implies an operating loss of more than $230M. In addition to that, interest income and interest expenses should be considered; the Q406 numbers would imply that's another $50M down the gutter. That is some quite problematic cash flow for a company with about $1.5B in cash and cash equivalents. As for the reported non-GAAP net loss, it will obviously be higher, possibly more than $400M, since you need to add amortization and one-time expenses there.

Given the likely further erosion of margins and ASPs in Q2, gross profit could go down as low as $250M in a worst-case scenario. Best case, you might imagine $350M, perhaps. At the current operating expense levels, that would indicate an operating loss between $325M and $400M. Of course, AMD will try to minimize operating expenses going forward through a variety of means, so you could hope for an operating loss that's about $40M to $60M lower than that. Also, it all depends a bit on the former-ATI operations' performance, but obviously, if the preivous numbers are correct, the situation is not one that AMD can sustain. There's no reason Q3 will be any better, either.

In addition to all of this, it should be noted that even after revising capital expenditures (CapEx) downwards by $500M+, AMD is still going to use up to $1.5B there this year, apparently. CapEx is defined as expenditures used by a company to acquire or upgrade physical assets - in AMD's case, that's mostly for fabs. That number is rather huge to say the least, considering that our estimates indicate AMD could be running out of cash within 15 months, based on operational expenses alone! Of course, this is assuming that AMD's new CPU products (that should begin ramping in Q4) don't help, which they obviously will - nobody knows how much, however.

Conclusion

The second quarter will see further ASP and margin declines, as AMD's April 9th price cuts take effect. However, there is no reason to believe that market share will go down further, at least in desktops, as AMD is pricing some of their chips very aggressively, even after taking into consideration Intel's presumed April 22th cuts.

In early June, Intel will introduce the Pentium E21xx and the Celeron 4xx product families, based on the Conroe-L core, thus getting rid of the last part of their line-up that wasn't competitive with AMD's: the entry-level Netburst-based processors. As we've said in the past, we believe Conroe-L to be a dual-core chip with one core disabled for the Celeron 4xx family. It should thus measure around 85mm², which is significantly smaller than AMD's 65nm dual-core offerings, but bigger than their single-core offerings (although with possibly higher yields due to redundancy and a more mature process...)

Conroe-L, according to early benchmarks, won't really be substantially faster than AMD's offerings at a given price point. But unlike previous Intel offerings in that market segment, they will be fully competitive. This, along with Intel's launch of the Santa Rosa platform, could hurt AMD's market share in Q3. So between today and the end of Q3, both AMD's market share and ASPs are likely to go down further, we believe. Simply put, the worst might be yet to come for AMD's CPU business. Even the higher ratio of 65nm parts won't offset much of anything, sadly.

Given the net losses that we estimate AMD will need to operate under in Q1, Q2 and Q3, we wish them to best of luck in getting new products out of the door and in finding sources of financing in order to weather the storm in what will be one of the company's toughest years, ever. Either way, we look forward to AMD's Q107 conference call on April 19th to get some more official information on AMD's financial position, and to confirm or deny our analysis.

---

Comments, opinions, feedback? Please post any of your reactions to the article in this thread.