Dense ARM servers, probably another console or two (Nintendo? SteamBox? Samsung? PSP?) and then who knows?
Meanwhile, their professional graphics market share is slowly but steadily growing, so there's that too.
The AMD Execution Thread [2007 - 2017]
- Thread starter overclocked_enthusiasm
- Start date
-
- Tags
- amd
- Status
Dense ARM servers, probably another console or two (Nintendo? SteamBox? Samsung? PSP?) and then who knows?
Meanwhile, their professional graphics market share is slowly but steadily growing, so there's that too.
overclocked_enthusiasm
Regular
You're right...AMD should just stay in their traditional market and be happy with 15% market share (mostly low end) against a competitor that outspends them 118 to 1 in CapEx ($13 Billion to $110 Million). They should stay in a market where they are consistently 1-1.5 dies shrinks behind their competitor and suffer the resultant deficiencies in cost, performance per watt, and profitability. They should stay in a market that is declining to the tune of 10% each year as customers transition to other non-PC platforms. They should be happy with a share price of $3.50 and market cap of $2.65 Billion....1/10th of their high water mark back in 2006. What's not to like?
"In almost all cases, Duopoly competition produces consistent profitability for the #2 company, except in heavily capital intensive businesses. When an industry requires very high capital expenses to remain competitive, the larger more efficient manufacturer can increase market share and profits, while the #2 company loses money. With the exception of a few product cycles, when AMD was first to market with an innovative design, Intel consistently used their manufacturing prowess to starve their significantly smaller competitor of profitability."
overclocked_enthusiasm
Regular
The game console business is poorly understood by many people. Here is what is generally accepted:
~ Typical 5 year lifespan peaking in year 3
~ Most predictions are 90 million units (Xbox & PS)...figure more for Steam and Wii
~ ASP of $60-$100 per APU (with VERY low CapEx due to royalty agreement)
~ Since these are "semi-custom" parts, the end buyer (Microsoft and Sony) pick up the lion's share of the R & D tab in the form of royalties that proceeds the actual revenue once the silicon ships
~ Gross margins of "mid teens" and "improving from there" per CFO on these semi-custom parts
UNITS
Q4 2013 - 5 million
2014 - 20 million
2015 - 25 million
2016 - 20 million
2017 - 20 million
REVENUE & EPS @ $80 ASP
2013 - $400 million @ 18% margin = $72 million or .10 eps
2014 - $1.6 billion @ 19% margin = $304 million or .42 eps
2015 - $2.0 billion @ 20% margin = $400 million or .55 eps
2016 - $1.6 billion @ 21% margin = $336 million or .47 eps
2017 - $1.6 billion @ 22% margin = $320 million or .49 eps
So AMD has something they haven't had in a looooong time. A steady revenue stream for the next 5 years, at profitable GM levels. This gives them the freedom to continue the transitioning of their business model from 90% traditional x86 to 50% non-traditional (ARM, APU, Semi-Custom, Embedded, etc.). On a $6 billion yearly revenue, the console business alone will be close to 20-25% of total revenues. With this financial breathing room they can move their new flagship APU (below) up the food chain and continue exiting the large core x86 bloodbath with Intel.
http://www.hardwarepal.com/amd-kaveri-apu-next-month-specs-pricing-revealed/
"It seems the A10-7850K will be AMD’s leading APU. AMD’s A10-7850K just like the rest of the “K” variants is an unlocked APU with 4 cores, a Base Clock of 3.7GHz and Turbo Core of 4.0GHz. While on the other hand the A10-7700K has a Base Clock of 3.5GHz and Turbo Core of 3.8GHz (also 4 cores)"
"Both the APU’s in the graphics department will have the R7 2XXD as a GPU unit and will feature the new GCN architecture bringing forth Mantle , DirectX 11.2, OpenGL 4.3 and AMD TrueAudio. While the A10-7850K will have 512 GPU cores and 8 GCN cores , the A10-7700K will have 384 GPU cores and 6 GCN cores. Both the GPU’s will be set at 720MHz which might not be enough for a dedicated GPU, but for an APU that might be more than enough."
"Both the APU’s will have 4MB L2 cache while all this will require a 95W TDP for the A10-7850K , the A10-7700K will have a configurable TDP of 65W depending on the performance of the whole APU. As far as prices go, both the APU’s are rumored to be tagged at around $150 something that Intel seems to fail to achieve with their higher priced products. All this will be confirmed on the 14th of January when both the APU’s are set to be launched. Further information will be available at CES 2014 and a full list of features will soon be unveiled"
For sub $200 parts, likely $150-$185, I see great opportunity for AMD here to grab share against other Intel integrated solutions and many discrete Intel/Nvidia solutions as well. I am betting AMD's margins on these parts are better than recent norms since they don't have a direct apples to apples competitor in Intel...because these parts are becoming more powerful...especially on the graphics front as they leverage their ip from the ATI aquisition.
~ Typical 5 year lifespan peaking in year 3
~ Most predictions are 90 million units (Xbox & PS)...figure more for Steam and Wii
~ ASP of $60-$100 per APU (with VERY low CapEx due to royalty agreement)
~ Since these are "semi-custom" parts, the end buyer (Microsoft and Sony) pick up the lion's share of the R & D tab in the form of royalties that proceeds the actual revenue once the silicon ships
~ Gross margins of "mid teens" and "improving from there" per CFO on these semi-custom parts
UNITS
Q4 2013 - 5 million
2014 - 20 million
2015 - 25 million
2016 - 20 million
2017 - 20 million
REVENUE & EPS @ $80 ASP
2013 - $400 million @ 18% margin = $72 million or .10 eps
2014 - $1.6 billion @ 19% margin = $304 million or .42 eps
2015 - $2.0 billion @ 20% margin = $400 million or .55 eps
2016 - $1.6 billion @ 21% margin = $336 million or .47 eps
2017 - $1.6 billion @ 22% margin = $320 million or .49 eps
So AMD has something they haven't had in a looooong time. A steady revenue stream for the next 5 years, at profitable GM levels. This gives them the freedom to continue the transitioning of their business model from 90% traditional x86 to 50% non-traditional (ARM, APU, Semi-Custom, Embedded, etc.). On a $6 billion yearly revenue, the console business alone will be close to 20-25% of total revenues. With this financial breathing room they can move their new flagship APU (below) up the food chain and continue exiting the large core x86 bloodbath with Intel.
http://www.hardwarepal.com/amd-kaveri-apu-next-month-specs-pricing-revealed/
"It seems the A10-7850K will be AMD’s leading APU. AMD’s A10-7850K just like the rest of the “K” variants is an unlocked APU with 4 cores, a Base Clock of 3.7GHz and Turbo Core of 4.0GHz. While on the other hand the A10-7700K has a Base Clock of 3.5GHz and Turbo Core of 3.8GHz (also 4 cores)"
"Both the APU’s in the graphics department will have the R7 2XXD as a GPU unit and will feature the new GCN architecture bringing forth Mantle , DirectX 11.2, OpenGL 4.3 and AMD TrueAudio. While the A10-7850K will have 512 GPU cores and 8 GCN cores , the A10-7700K will have 384 GPU cores and 6 GCN cores. Both the GPU’s will be set at 720MHz which might not be enough for a dedicated GPU, but for an APU that might be more than enough."
"Both the APU’s will have 4MB L2 cache while all this will require a 95W TDP for the A10-7850K , the A10-7700K will have a configurable TDP of 65W depending on the performance of the whole APU. As far as prices go, both the APU’s are rumored to be tagged at around $150 something that Intel seems to fail to achieve with their higher priced products. All this will be confirmed on the 14th of January when both the APU’s are set to be launched. Further information will be available at CES 2014 and a full list of features will soon be unveiled"
For sub $200 parts, likely $150-$185, I see great opportunity for AMD here to grab share against other Intel integrated solutions and many discrete Intel/Nvidia solutions as well. I am betting AMD's margins on these parts are better than recent norms since they don't have a direct apples to apples competitor in Intel...because these parts are becoming more powerful...especially on the graphics front as they leverage their ip from the ATI aquisition.
I never said they should not try, but I don't see any "good" place for AMD with their current IP portfolio. That's why I reacted. Out of Intel, it's not easier... at all...
I stopped here. Do you really think these crippled mid range PCs called next gen will have the same lifespan as the previous models ?
I don't think so and I'm not the only one:
http://www.expertreviews.co.uk/game...e-to-have-shorter-lifespans-claim-ea-and-sony
With Steam machines around the corner, mobile SoCs more and more powerful each generation, PS4/X1 will have hard time to last, so your revenue prediction is way off the chart...
Previous generation launched 8 years ago, he only listed 5 years there. Additionally the old consoles, despite their age and the things you talk about related happening, the sales of those units are still high.
overclocked_enthusiasm
Regular
Hi Dave, nice to see you! I have a question for you. Can you leverage the GPU power in an APU coupled with a discrete graphics card for a mini Crossfire effect? If so, then the APU has legs to move up into the mid-range and higher desktop as the CPU performance scales higher yes?
How goes the adoption of Mantle with the game devs for the consoles? Seems like there may be some performance advantages for AMD hardware once the games are fully ported to PC.
How goes the adoption of Mantle with the game devs for the consoles? Seems like there may be some performance advantages for AMD hardware once the games are fully ported to PC.
overclocked_enthusiasm
Regular
My estimates are conservative...they exclude Wii and Steam...both which are also likely to use hardware from AMD. My sales estimates come from industry sources who crunch the numbers of the existing user base, the refresh cycle, and other metrics. As Dave said, the last cycle is still going strong after 8 years...and sales for this new cycle are super strong.
I will concede one small point to you...I think the next generation consoles (2018 and beyond) will be more similar to sff PCs with easily up-gradable components. The hard core console gaming base will want to have better and better technologies (especially as 4K hdtv becomes affordable) AND the PC gaming base continues to migrate to consoles or sff PCs for the living room instead of the office. So imagine a sff PC/console for 2018 and beyond that will have upgrade options during its lifespan...RAM, CPU, GPU, APU, power supply, hard drive, etc.
So will there be some migration from consoles to your vision of sff PCs for every living room during THIS console cycle (2013-2018)? Sure...and Steam will be the testing ground for that. So I'll give you 5-10 million of my 90 million (Xbox/PS). But guess what? You have to give me those 5-10 million right back because Steam Machine will use AMD hardware also! You also owe me another 10 million or so for Wii. so we are up to 100 million units over 5 years...not 90 million.
So in essence you further prove my point. AMD has a sustainable and robust revenue stream for the next 5 years (at least) with multiple platforms (Xbox, PS, Steam, Wii). If you are assuming my comments favor or advocate consoles over sff PCs you miss the larger point. The point is AMD is knee deep in the middle of BOTH consoles and sff PCs for living room gaming. Capisce?
overclocked_enthusiasm
Regular
The success of their semi-custom APU wins in consoles clearly demonstrates their ability to leverage their IP (in this case from ATI) into real world products with good margins and substantial revenues. This IP will be further leveraged in tablets, laptops, low end desktop and even higher depending on the performance of the upcoming APU lines.
The other area they can leverage their IP is in the form of ARM servers...based on their Sea Micro IP. This was proven by the huge Verizon deal (and don't try the "but Intel shipped more CPUs than AMD did in that Verizon deal" line). Just two examples of their IP leading to wins...
If by "good place" you mean an unsaturated market where they can dominate with a unique product...I don't think those markets exist. BUT those players that can leverage their IP to distinguish themselves from their competition will thrive. NOBODY can outfab Intel...nobody...but you can invent a better mousetrap for different markets if you expand your competitive base...and the results could be surprisingly positive.
I can't speak for Dave but I can tell you that the answer to your first question is yes, it already exists and is called Hybrid Crossfire. However most reviews show that it only works really well in a small number of games. I suppose Kaveri might help with that since it raises APU graphics power closer to the level of low~mid-range discrete graphics.
As for Mantle, AMD has been pretty clear that it's a PC API.
"Hybrid Crossfire" is the technology description, at the platform level the implementation is called "Dual Graphics".
http://www.amd.com/US/PRODUCTS/TECHNOLOGIES/DUAL-GRAPHICS/Pages/dual-graphics.aspx#1
http://www.amd.com/US/PRODUCTS/TECHNOLOGIES/DUAL-GRAPHICS/Pages/dual-graphics.aspx#1
The main reason that older [PS3 and Xbox 360] consoles are still selling well is that the price keeps trending lower and lower over time. Nothing magical about that.
While the new console business will certainly be a fairly reliable stream of revenue for many years to come for AMD, the gross margins (and hence profitability) on these console APU's will be fairly low (note that the Xbox One APU alone has > 5 billion transistors!). Also note that a shift towards 50% "non-traditional" [non-x86] sources of revenue is only fruitful if revenue gains from non-traditional more than offsets the revenue loss from "traditional" [x86] business (and over the next one or two years in particular, this positive scenario is very unlikely to hold, and the more likely scenario is that revenue decrease from "traditional" business will outweigh revenue increase from "non-traditional" business). Finally, in moving to "non-traditional" sources of revenue, there will be significantly more competition to AMD, not less!
Anyway, there are some headwinds that will limit the growth of new console business for this generation. One is the growth in ultra mobile computing. Two is the much lower consumer cost and stronger software ecosystem for previous generation consoles. Three is the commoditization of some new console features. Four is the growth in cloud computing and cloud gaming.
To be a serious player moving forward in what is increasingly becoming an ultra mobile world, one will need an ultra low power CPU, an ultra low power GPU, and an LTE modem that can be integrated on die with these other components. At the moment, AMD is well behind in all of these areas. So, all things considered and knowing how unpredictable the equities market is, you should really think about diversifying your portfolio, even as you take a long position on AMD.
Last edited by a moderator:
overclocked_enthusiasm
Regular
The console gm question was specifically addressed by the CFO at the CS conference last week. In 2Q AMD guided "low teens"...in Q3 they came in at "mid teens" and last week the CFO said "gms will improve further as yields improve. Your number is not out of the question" in response to a 20% gm number floated by the host. Since the end user (Microsoft/Sony) picks up the lion's share of the R & D costs associated with these semi-custom parts the impact on AMD's CapEx is very low with high flow through of these console gms to the bottom line. The "lower consumer cost" is irrelevant as long as you can make a good margin on enough units.
Gaming will continue moving to the cloud, will continue to migrate away from PC, will continue to migrate to anything mobile...I agree on all these things. But the installed user base in console gaming is so high, there is such loyalty in platforms (i.e. "all my friends and I play together on XXX"), that I see strong sales for at least this current new generation (2013-2018). These kids and other hard core gamers want to be on the couch or in their bedrooms with their headsets on talking with their friends while they pwn noobs together with a bag of Doritos and a Coke (or smoking bowls)...hardly doable on a mobile device. Everything in technology, features and/or hardware, turns into a "commodity" at some point. How you leverage your product in this commoditization process with your superior IP or marketing separates the wheat from the chaff. I still see a stable enough console revenue stream for AMD to have the luxury to hyper-focus on their APU and ARM strategies...both of which have the potential to pay off HUGE.
From an investment perspective, I believe diversification is for wealth preservation...not creation. I'm not your typical "7-10% per year is great...as long as I beat the S & P". I allocate a substantial chunk of my net worth and swing for the fences with an all in and heavily margined trading style where doubles, triples, or more are the goal. I swing trade...weeks, months, sometimes years. I rode ATYT (ATI) from $4 to $20 back before AMD bought them and rode AMD from $13 to $42 down to $23 (I was an idiot for not selling sooner)...so I have had some success in this space before. I am an unconventional investor, a contrarian, who sees opportunity where others see risk...buying more today at $3.65.
overclocked_enthusiasm
Regular
It should also be noted that the consoles are often sold at or below cost...even at the high water retail price point in their life cycle. The profit is derived through the games, peripherals (2nd controller etc.), online, and the halo effect the consoles have on the larger business. With that being said, do the hardware vendors share the pain in this price reduction over time? Likely, but by that time yields have improved to the point where there is some room to give a little back in the form of asp.
The PS4 and Xbox One were described as being a narrow loss in the worst case for the PS4 (depending on the exchange rate) and break-even or proft for the Xbox One right now.
The contracts for the APUs are most likely to have aggressive cost-reduction schedules. The per-chip price may be relatively high right now for a console component, but the idea is that they will reduce over time. What the console makers do with that reduction is their concern.
If circumstances arise that prevent the cost of manufacturing the chips from falling fast enough, that is most likely AMD's problem.
I'm not sure how the ordering process would work, but I am curious if the contract nature of the process will insulate AMD from its historically poor inventory and channel management.
That is actually not really true. Xbox 360 S [with 4GB HDD] launched slightly more than 3 years ago for $199 USD retail price (while the stripped down $199 USD Xbox 360 Arcade edition without HDD was discontinued). So the net result was more hardware for the same end user cost. Also note that street prices of Xbox 360 have clearly trended down over time, even within the last few years (at least here in the USA). In fact, this holiday season the Xbox 360 console had a street price as low as $149 USD at various retailers (for obvious reasons of course). There are other factors too that made the Xbox 360 platform a better deal over time, such as a stronger software ecosystem, bundled games, lower average street price of games, greater availability of rental games, and the introduction of Kinect.
- Status
