ATI announced after market close that their qtr revenue would be down by over 100 million $ to 520 million. They blame Intel yanking their chipset business faster than anticipated (what were they expecting - Intel to continue to put cash in the pocket of their competitor) and a "supply chain adjustment" by a major handheld customer. That's some phenomenal spin - a big cancelled order is now a "supply chain adjustment".
But I'm not here to pound ATI. Naturally their stock price is down post market close. However, AMD is down 3% post market having dropped 6.45% during market. Now it was a bad day for tech but this is going to raise questions on Wall Street and with shareholders on the financial wisdom of the deal. There is a chance that from a revenue & profit standpoint ATI may not be bringing the same level of upside in 2007/2008 as AMD projected. In addition AMD is locked in at 22$ price per share for ATI having used the low price of 18.26$ per share for their own conversion. The possible implication of this may be to put more pressure on their credit rating and the terms of the 2.5 bln $ debt they are taking on.
If you look at my earlier post on what uncertainties could hinder AMD on this acquisition, these are the kind of unforseen issues I was referring to. I'm not claiming I knew this specific hitch was coming but you understand now what I was talking about.
Frankly I think this deal is a mixed bag for AMD:
1. They will benefit from having more marketing and sales bundling bullets in their arsenal.
2. They will also be able to make some headway on platforms but mostly on the enterprise side. Consumer and SME is doubtful with their limited ability to build a large brand for these segments. You can see my comments on this here by scrolling down the comments.
3. Their ability to integrate high quality graphics (well better than Intel's sad IGP) and CPU on the die will help penetrate emerging markets as it should enable lower cost. Though I have to believe there are substantial design and manufacturing challenges to this I have a higher confidence in AMD than Intel here. But this will probably not come to light before 2008.
4. The acquisition will divert resources that were focussed on sustaining momentum against Intel. People in many places inside AMD will be forced to stop what they were doing and work through this merger. Heck, Hector and team will be forced to do that.
5. Managing their cash and debt position will be very difficult as long as this price war continues. A couple of bad quarters will make things sticky. Which doesn't mean Intel will have good quarters but remember - a good Q2 for AMD was net income 90 million $ and a bad Q2 for Intel was 800 million $. Going from 90 to 0 is a lot easier than going from 800 to 0.
I don't think AMD had a choice if they wanted to stay competitive in the longer term. Timing of the deal is not perfect but I think they did the right thing biting the bullet now rather than later. Hector and team have shown a high ability to manage the business but the next year is going to be tough. Any mis-step can cost them dearly. Which further tightens their ability to resource the incremental things they need to grow. And I think the first likely serious set back will be market and then revenue share reduction due to the price war + Conroe launches leading to a quarterly GAAP loss in the next 2-3 quarters.
Subscribe to:
Post Comments (Atom)
13 comments:
Azary - very good point. I can only guess but here's what I think it is.
Between AMD and Intel by end 2007 they will probably have more capacity than the global PC market needs assuming the market grows at a normal pace and every single product sold isn't quad core -:).
1. Which is why they are both investing heavily to explode emerging markets and create lower cost PCs for the next billion consumers. PIC, community PC, etc.
2. It's why Intel is investing in platforms and why AMD has acquired ATI. To find new usage models to stimulate the next wave of consumption.
3. It's why Intel is throwing big $s at initiatives like Wimax, ultra mobile PCs, etc and created a Digital Health group.
Both Intel & AMD have realised the PC market can no longer provide the kind of growth they need. Hector has less pressure because he can focus on incremental market share which gives him some more time. But as he grows bigger his rate of growth diminishes.
Otellini on the other hand realises to compete means lower margins for Intel the fat cat. My guess is that's why he's cutting costs now. Because the days of 62% margins are over if he must find the next billion customers with lower cost PCs. Frankly, there are two businesses I can think of where avoiding being commodotised is not an up-hill fight. Oil and telecom/content service providers. Oil this is not - so I'm surprised Intel has not looked at revenue models where you establish an ongoing source of revenue. Perhaps that's the game w/ Wimax. They want to sell the hardware but also want some kind of leg in to the recurring revenue of service. But I don't think any of their investments give them a position to benefit from service revenue - that have been traditional investments Intel Captial makes. But that would be interesting if they decided to go down that route - build the hardware, sell it at lower margins but make money from the service that connects your hardware.
If you think about it, this is exactly the lesson Microsoft is learning from Google right now. Make the application free and earn recurring revenue through advertising.
Azary - to add to my comments. If the PC market slows down or neither AMD nor Intel are able to find the "next big thing", they will both descend into a price war again. Blood will flow into the streets and at that point I will have to say AMD will be in a much worse position than Intel being one process generation behind, in serious debt from the ATI acquisition and continuing to invest ahead of the revenue curve in R&D & mfg (they've made a few rapid announcements on new lab in Shanghai, assembly & test in Singapore, etc). But that's a year off so let's wait and see.
Much can happen between now and then and either one of them could stumble.
cheers...
I think AMD is ready to take a short-term hit due to loss revenues from the ATI Conroe platform. But the main goal for AMD is to sell more cpus down the road via platforms, and bundling ATI chipsets will certainly does that. Another goal for AMD is gain mindshare of 3rd party vendors once a platform has established.
Motherboard makers will fight for contracts from HP, Dell, etc to supply them with AMD-ATI MBs.
Also, Broadcom and Atheros and other wireless suppliers are also in the WiMax consortium with Intel if I'm not mistaken, so AMD should have ample WiMax supports.
AMD's geode brand is also gaining supports in devices like UMPCs, thin clients and other electronics gadgets too and ATI will add more advanced features to that platform also.
The funny thing is I don't understand where AMD is getting all their money from.
1. They are buying AMD.
2. They are building a completely new NY Fab.
3. Converting Fab 30 to Fab 38.
4. Transitioning Fab 36 to 65nm by mid 2007.
5. Transitioning Fab 36 to 45nm in 2008.
6. Paying Chartered for chips.
7. Introducing new K8L architecture.
8. Introducing new "dedicated" mobile architecture.
9. Continuing changes to platforms.
etc.
You point out that they only make $90million, so I really don't see where they are getting all this money. Acquiring ATI will also add of their problems to the list which is a major reduction in chipset sales by the elimination of Intel chipsets, and also a downturn in graphics sales. The fact is that Conroe is a great gaming chip and will be for the next year until K8L can get on it's feet. That means that for the next year, Intel is going to be the dominant gaming platform and with people squemish of buying ATI it isn't the best position to be in. I guess it'll depend on how good R600 is against G80, but even then the G80 has the advantage of launching first and gaining traction.
Here's what I think is going on:
1. Spinning off Spansion was a big deal. They lost 311 million on their memory biz in 2005. Even though this takes revenue down by 2 billion, it increases profitability.
2. At the time of the ATI deal they said they would part fund with about 3 billion cash in the bank and 2.5 billion debt/stock.
3. They have various outstanding notes and debt. For example, Fab 36 is a partnership with a silent group who lent them around 150 million and the govt of Saxony with over 200 million invested.
4. They have been diluting equity by converting debt to stock in some places. I.e. they had 398 mln outstanding shares end of 2004 and 442 million at the end of 2005. This has no increased to 489 million...undoubtedly as their options get unloaded due to stock price jumps.
This is why I made the comment on managing cash and debt. Any mis-step means their credit rating declines and the cost of borrowing increases. In addition, their existing creditors have conditions in place that permit conversion of debt to stock, limit further borrowing, etc.
If they get hurt in the price war and it limits their profitability...or even removes it, they will have serious trouble. Their ability to invest in all the programs and infrastructure they need to grow will be hit badly. They will bleed. I think we'll need to watch Q3...but more likely Q4 to see the impact of the price war.
n4cr - how about end user system price, thermals, etc on Torrenza? Any idea whether it will make mainstream pricing or will it end up niche at the high end?
It's on thing to have a performance lead across the board like AMD did over Netburst. But it's another thing to have performance lead only in the very high end. That won't help win significant market share.
There is a chance that from a revenue & profit standpoint ATI may not be bringing the same level of upside in 2007/2008 as AMD projected.
Actually, it is more likely that ATI will sell a lot of chpsets in 2007. You don't seem to understand that AMD needs about 20 Million more chipsets this year than last and another 20 Million more next year. This is definitely what AMD needed.
The possible implication of this may be to put more pressure on their credit rating and the terms of the 2.5 bln $ debt they are taking on.
It shouldn't. AMD can pay this off in about 1 1/2 years.
2. They will also be able to make some headway on platforms but mostly on the enterprise side. Consumer and SME is doubtful with their limited ability to build a large brand for these segments.
Building Brand is the job of the OEM, not the supplier. You have this backwards.
4. The acquisition will divert resources that were focussed on sustaining momentum against Intel. People in many places inside AMD will be forced to stop what they were doing and work through this merger. Heck, Hector and team will be forced to do that.
Sorry, but you are way off here. This actually effects very few people at each company. There should be perhaps five teams working now. One would be looking into merging order/inventory, one for technical projects, one for accounting, one for marketing, and one for human resources. The order/inventory merger team probably includes outside contractors. The accounting team probably includes an outside auditing firm. The actual effects of the merger don't really happen right away. For example, AMD may replace whatever accounting/order/inventory application ATI is using with one compatible with their own so this would require training before switchover which is what companies do even without mergers. Engineering projects are nothing unusual except that some engineers may have to travel or relocate. This is nothing unusual as this happens now. This takes some executive time but not as much as you think. This won't distract Hector or upper management.
5. Managing their cash and debt position will be very difficult as long as this price war continues. A couple of bad quarters will make things sticky.
Not really. AMD can fund the debt even at 2005 income levels.
Which doesn't mean Intel will have good quarters but remember - a good Q2 for AMD was net income 90 million $ and a bad Q2 for Intel was 800 million $. Going from 90 to 0 is a lot easier than going from 800 to 0.
You don't understand the money. In the 2nd quarter alone, Intel piled on $800 Million in extra inventory and saw non-computing losses increase by $200 Million with a drop in income of $1 Billion. That's a $2 Billion drop in one quarter. You'll know it is serious if Intel keeps stock buyback low for several quarters.
but the next year is going to be tough. Any mis-step can cost them dearly. Which further tightens their ability to resource the incremental things they need to grow.
I don't think so. I think that AMD had considered building a FAB in New York and then decided to delay that and spend money on ATI instead. This would be consistent with the bump and test expansion.
And I think the first likely serious set back will be market and then revenue share reduction due to the price war + Conroe launches leading to a quarterly GAAP loss in the next 2-3 quarters.
I doubt it. AMD should have definite revenue growth in the 3rd and 4th quarters over 2005.
Theres one thing that's been bothering me lately, AMD now has 65nm going plus they're on 300mm wafers now, right? Thats not all: rev G is going to be die shrink, so my question is why do they need so much capacity for? I mean for real, with all of the above, even with yields in 50's and all chips made dual core they still going to have enough to flood the market.
No. Your estimates are way off. AMD has had a problem in the 2nd quarter because FAB 30 is still the the same capacity as last year. So they have not been able to expand enough because of the limited capacity. There is additional pressure now because of increased demand for X2's. The FAB 36 ramp, 65nm production, and additional chips from Chartered will be needed to meet demand. Also, my estimate for yield was 80% for 90nm and just under 70% for 65n.
Between AMD and Intel by end 2007 they will probably have more capacity than the global PC market needs
Really? Well, I understand where AMD's increase comes from but where do you see Intel's increases coming from during 2007? Intel will be almost all 65nm by end of 2006 and no new FABs online until the very end of 2007.
1. Which is why they are both investing heavily to explode emerging markets and create lower cost PCs for the next billion consumers. PIC, community PC, etc.
This is a Sempron class market. AMD doesn't have to have this. They are putting much more effort into gaining server and desktop share.
2. It's why Intel is investing in platforms
These are not likely to increase sales much.
and why AMD has acquired ATI.
To deliver what OEM's wanted.
To find new usage models to stimulate the next wave of consumption.
This has nothing to do with it. Companies buy new computers. AMD just wants to gain in this market.
Both Intel & AMD have realised the PC market can no longer provide the kind of growth they need. Hector has less pressure because he can focus on incremental market share which gives him some more time. But as he grows bigger his rate of growth diminishes.
No. This isn't correct at all. Cpu sales are still AMD's most profitable area by far. This is what AMD is concentrating on. Nor has AMD reached a point where growth is slowing. They should continue to grow fairly rapidly until 2009.
Otellini on the other hand realises to compete means lower margins for Intel the fat cat.
No. Otellini is desperate to stop share loss.
My guess is that's why he's cutting costs now. Because the days of 62% margins are over
Intel would actually be at 60% margins if they divested all of the non-computing comapanies.
if he must find the next billion customers with lower cost PCs.
Really? Then why is he making the E6300 Conroe with 2MB's of L2 instead of 256K of L2 like AMD uses for Sempron?
BTW, your speculations about businesses for Intel are also incorrect. These too have lower profits than cpu sales.
Azary - to add to my comments. If the PC market slows down or neither AMD nor Intel are able to find the "next big thing", they will both descend into a price war again. Blood will flow into the streets and at that point I will have to say AMD will be in a much worse position than Intel
This is incorrect. AMD can stop getting chips from Chartered without effecting their income.
being one process generation behind,
AMD is currently making 65nm chips and will be at about the ratio as Intel is with Conroe at the end of the year. In fact, the indications have been that Intel is having some difficulty ramping Conroe (which is far below Intel's actual 65nm capacity) whereas AMD is ahead of schedule. There should be 65nm AMD chips available in another month. Intel counts release as start of production; AMD counts release as after 2 months of production. If we use the Intel Calendar, then these chips were available last month. This would put AMD 65nm just 2 months behind the release of Woodcrest.
in serious debt from the ATI acquisition and continuing to invest ahead of the revenue curve in R&D & mfg (they've made a few rapid announcements on new lab in Shanghai, assembly & test in Singapore, etc). But that's a year off so let's wait and see.
The debt for the ATI purchase is fine even with last year's revenues. The cap ex items you've mentioned are quite small compared to the costs of FAB upgrades which are taking place from now through 2008. The FAB upgrades can run $1 Billion.
The funny thing is I don't understand where AMD is getting all their money from.
AMD actually made $2 Billion in 2004 and $2.4 Billion in 2005. Income is the amount left over after all of your expenses. Capital expenditures are an expense and so is R&D. So, in 2004 AMD spent $900 Million on R&D and $900 Million on Cap Ex and had $222 Million as income. In 2005, AMD spent $1.1 Billion on R&D and $1.1 Billion on Cap Ex. and had $232 Million left over as income. The Cap Ex. budgets for 2004 and 2005 were $2 Billion and $1.5 Billion of that was the money spent to build FAB 36. Another $500,000 is tooling.
1. Spinning off Spansion was a big deal. Even though this takes revenue down by 2 billion, it increases profitability.
True. They are much better off without Flash memory; so would Intel be.
lent them around 150 million and the govt of Saxony with over 200 million invested.
You missed the fact that their long term debt declined by $600 Million from Q4 05 to Q2 06.
4. They have been diluting equity by converting debt to stock in some places. I.e. they had 398 mln outstanding shares end of 2004 and 442 million at the end of 2005. This has no increased to 489 million...undoubtedly as their options get unloaded due to stock price jumps.
This is in line with actual assets. They divested FAB 25 with Spansion but replaced it with the much more valuable FAB 36.
I think we'll need to watch Q3...but more likely Q4 to see the impact of the price war.
You need to include the fact that costs drop for AMD toward the end of the year and all through 2007. Costs don't actually drop for Intel except for a small amount as they ramp what is left of 65nm and improve yield.
Scientia:
Capacity - note, I said AMD and Intel will have more capacity than the global market. I didn't say Intel will have more capacity. Having said that, Conroe die sizes are smaller than Netburst and this will boost their output.
Intel's success in platforms - we've had this debate already. You're just stating your opinion which doesn't prove anything. In fact, Centrino is a big fat "you're wrong". However, as I said in another post, Intel is not driving their new platforms as effectively or hard as they did Centrino so far.
60% margin for Intel w/o non-computing businesses - prove it. Show us the analysis.
AMD 65nm ratio similar to Conroe by end of year - first, AMD has just retracted their claim that 65nm will be out in Oct. Second, the observation is Intel has a process generation lead. Had they not had this lead they would be in a worse position than they are today. Now with Conroe, a product that's better than AMD and smaller die size, that lead will become a real advantage.
AMD can pay off the 2.5 bln debt for ATI purchase in 1.5 years - prove it. Show us the analysis.
Building brands is the OEM's job - ha ha...which is why Intel is the 5th most valuable brand in the world and 7x AMD's revenue and commands a price premium and is able to survive even with a crap product like Netburst. And btw - Microsoft is #2. Building your own brand is every corporations job. How and to whom is the decision. Brands allow you to not become commoditized and be driven only by price all other things being equal. Should I recommend a book for you to read?
Managing the acquisition - dude, you are way off. You make it sound like 20 people in AMD will manage the ATI acquisition. When you acquire a company like ATI that fundamentally changes your business model, you have to shift the mindset of 10k existing AMD employees. You need to re-set the org focus, establish new teams, build the teams, clarify roles, build new relationships, change business processes, etc. Man...in a merger like this almost everyone at AMD and ATI will be touched.
AMD can fund debt at 2005 revenues - first, assumption is they will sustain those revenues/profits. I say not. Second, prove it - show us the analysis.
Intel costs don't drop till 2007 - wrong. Costs drop as they toggle to Conroe which is a smaller die size than Netburst.
Post a Comment