Tuesday, August 28th 2018
AMD Chip Manufacturing to Lay Solely With TSMC On, After 7 nm - And Why It's not a Decision, but a Necessity
It's been a tumultuous few days for AMD, as the company has seen Jim Anderson, Computing and Graphics Group leader after the departure of Raja Koduri, leave the company, at a time of soaring share value for the company (hitting $25.26 and leaving short positions well, short, by $2.67 billion.) However, there's one particular piece of news that is most relevant for the company: Globalfoundries' announcement to stop all ongoing development on the 7 nm node.
This is particularly important for a variety of reasons. The most important one is this: Globalfoundries' inability to execute on the 7 nm node leaves AMD fully free to procure chips and technology from competing foundries. If you remember, AMD's spin-off of GlobalFoundries left the former with the short end of the stick, having to cater to GlobalFoundries' special pricing, and paying for the privilege of accessing other foundries' inventories. Of course, the Wafer Supply Agreement (WSA) that is in place will have to be amended - again - but the fact is this: AMD wants 7 nm products, and GlobalFoundries can't provide.To the forumites: this piece is marked as an editorial
AMD's CTO Mark Papermaster wrote thus in a blog post:
Should anything befall TSMC, should the silicon giant trip, AMD will have all of its product portfolio endangered. Consumer and professional GPU and CPU products will all be manufactured under TSMC's 7 nm process, as AMD has stated time and again - and in quite an aggressive manner. Here's hoping other players step up to the 7 nm manufacturing task, or, that being as hard as it is, that nothing affects TSMC's ability to deliver, lest one giant brings down others with it.
Sources:
AMD Blogs, via TechSpot
This is particularly important for a variety of reasons. The most important one is this: Globalfoundries' inability to execute on the 7 nm node leaves AMD fully free to procure chips and technology from competing foundries. If you remember, AMD's spin-off of GlobalFoundries left the former with the short end of the stick, having to cater to GlobalFoundries' special pricing, and paying for the privilege of accessing other foundries' inventories. Of course, the Wafer Supply Agreement (WSA) that is in place will have to be amended - again - but the fact is this: AMD wants 7 nm products, and GlobalFoundries can't provide.To the forumites: this piece is marked as an editorial
AMD's CTO Mark Papermaster wrote thus in a blog post:
AMD's next major milestone is the introduction of our upcoming 7nm product portfolio, including the initial products with our second generation "Zen 2" CPU core and our new "Navi" GPU architecture. We have already taped out multiple 7nm products at TSMC, including our first 7nm GPU planned to launch later this year and our first 7nm server CPU that we plan to launch in 2019. Our work with TSMC on their 7nm node has gone very well and we have seen excellent results from early silicon. To streamline our development and align our investments closely with each of our foundry partner's investments, today we are announcing we intend to focus the breadth of our 7nm product portfolio on TSMC's industry-leading 7nm process.The thing is, AMD going solely with TSMC for 7 nm isn't a decision: it's a necessity. It's a necessity of increasing need for AMD to bank on its powerful forward momentum against Intel on the CPU wars. It's essential for AMD's continued push in the professional, server field. And it's of utmost importance for AMD's relevance in graphics technologies against NVIDIA's frankly dominating position (despite AMD controlling all high-performance games consoles, a strategy that will only increase its fruits for AMD, should they be able to maintain this exclusivity - and all points towards that). Zen 2, EPYC 2, Vega 20, Navi - those are not only sizeable pieces of AMD's product portfolio, these are its bread and butter.As great as that part of the deal is for AMD, there's an obvious drawback of yet another foundry being left in the dust of new node developments: the weight of the world's semiconductor manufacturing capability on the 7 nm node - and AMD's efforts in it - stand solely on the shoulders of one player. And that's not even looking into actual output for TSMC's 7 nm node, how many customers will want to manufacture their chips on it, and whether or not TSMC has the ability to satisfy demand from all players.
Should anything befall TSMC, should the silicon giant trip, AMD will have all of its product portfolio endangered. Consumer and professional GPU and CPU products will all be manufactured under TSMC's 7 nm process, as AMD has stated time and again - and in quite an aggressive manner. Here's hoping other players step up to the 7 nm manufacturing task, or, that being as hard as it is, that nothing affects TSMC's ability to deliver, lest one giant brings down others with it.
87 Comments on AMD Chip Manufacturing to Lay Solely With TSMC On, After 7 nm - And Why It's not a Decision, but a Necessity
That said, I don't necessarily agree. I even got awesome tech support from an overseas Indian guy recently (you'd think that'd be annoying, but he went above and beyond).
On the graph you'e shown they're behind Nvidia and it's 2014. A lot has changed since then. After 7nm it's 5nm. then 3 nm (sometimes called 3.5 nm).
3nm test chips are already out.
Other than that... 5/7 nm mark the end of silicon in mainstream CPUs. We'll have to move to other materials.
For example: GaN not only makes 3nm doable, but also tolerates higher temperatures.
High quality Si processors stop working at 250*C or something like that. GaN is still fine at 500*C. We could see consumer CPUs hitting 7GHz etc.
Quantum computers are not coming to PCs. :)
Good day.
I'm mostly concerned about TSMC not feeling any need to push forward faster, as Intel, well, is exclusive to Intel anyhow. You are making that statement at the time when even Intel with it's market dominance and omg margins is struggling with new processes.
B2B chip business has many advantages. Uh, remind us, how many console chips, CPU chips are out there.
Going fanboibuthurt about such a meaningless metric is insane...
I still don't understand what will make AMD the leading customer. NVidia has not only grown since 2014 (in revenue, not in quantity), but they're also making a lot more things than just GPUs.
Don't underestimate the automotive business they have. Soon it will overtake pro graphics in revenue and may even approach what Nvidia makes from datacenters. NVidia vs AMD is not about getting "more business". It's a limited demand market. Either Greens or Reds sell a GPU. TSMC manufacturers for both, so what's stopping them from giving a priority to the company paying more? And NVidia has higher margins to share. :-)
BTW: I don't understand why Qualcomm and Broadcom appeared here. Care to explain? But wouldn't they pay more to basically control AMD's access to 7nm? :-)
People also need to understand that Samsung exists, and will make 7nm products eventually. Additionally GF is ramping up 12nm (non-Finfet) that has almost the same efficiency as TSMC's 7nm. That will be perfect for cheaper APU's and GPU's. How much more? Nvidia's gpu's are small fish compared to EPYC server contracts.
Even though this does impact the prognosis of volume shipments of 7nm parts, we still have to remember that in best case GF was still ~3Q behind TSMC, and would have produced chips with quad patterning on DUV. Maybe those extra wavers at GF wouldn't have been worth the cost for either party? "7nm" is not going to be a cakewalk for any of these foundries, and it would probably take significantly more investments to push it to EUV or other means which could eventually make volume production sustainable.
2) The reason people in low-skill low-pay jobs are paid poorly and treated poorly is, more than all other factors, because they are not as intelligent as those who demand better employment and thus seek it out. That does not mean there aren't highly-intelligent people doing those jobs. Of course there are. They are not, though, the dominant statistic. Odd circumstances can and do put talented people in menial jobs. We don't have much of a meritocracy, particularly when it comes to pay distribution. However, the idea that low-skill low-pay employment represents the best humanity can do, in terms of matching the human intelligence level with the work is utter nonsense. Just looking at what has been in fields like chip design and space observation blows your opining out of the water.
I have personal experience in these matters. I have a high IQ and very high ambitions and yet I have worked at many menial jobs in my life because of unfortunate circumstances, like having been born into poverty. I have many files full of IP that will never make it to market, wasted contributions. It is extremely difficult to rise socially right now. We live in a country where it is openly mocked as fantasy, for instance, for people to be able to go to college without being plunged into massive debt — even though a college degree is not even a guarantee of suitable employment anymore for intelligent people. The idea of investing in our cultural intelligence (and thus competitiveness) is openly mocked here. In Austria, by contrast, college students think it's insane that we make college students go into heavy debt to get an education. This, though, is racism (because "those people" never deserve what "my people" deserve) and class warfare (class = race, basically — regardless of the optics). Barriers are put into place to prevent competition between the more elite families and the masses. Tokenism is used, like lotteries, to give people the illusion of an open fair playing field. It's not. Not even close. However, there still is the fact that the average IQ isn't so special and there are a lot of people with that and with IQs below it.
It's unfortunate that eugenics policy is nearly always advocated by people who are foolish. They don't advocate for its benefits and instead advocate for it based on its drawbacks, drawbacks that can be sidestepped a lot with attention. For instance, you have to start with the realization that IQ is not monolithic. So, you have to be able to value a lot of different aspects of human intelligence. That includes the arts. It includes the humanities. It includes empathy. It includes unpopular innovations, not just the consensuses of the status quo intelligentsia. Sociopathic narcissistic self-enrichment (the current model of the ideal person in a culture that is based on the concept of net worth, one where billionaire worship is right out in the open and people conflate organized charity with altruism) is definitely not the height of human intelligence. Yet, it is what is rewarded. The goal is to cheat the system as much as possible and get away with it as much as possible. Unfortunately, that kind of mentality ranks below the top tier(s) of human capability. Eugenics should not be about punishing people. It should not be about hate, bigotry, and other kinds of narcissistic insecure gloating. It should be about providing incentives to high IQ people to breed more. Some may argue that that's the same as providing disincentives for the rest but there are massive differences if this dichotomy is implemented humanely/efficiently. One common mistake, for instance, is to believe that efficiency is inhumane. The human definition of efficiency is humanness. It encompasses all aspects of humanity. So, it is not merely about enabling certain fortunate people (high IQ individuals) to exploit everyone else. On the contrary. The goal is for everyone to be able to contribute and to benefit. That's the basis of civilization itself: specialization and reciprocity. It's unfortunate that our current dominant civilization models are Machiavellian perversions.
Perhaps ironically, our best hope for rational policy is a takeover by AI.
No one seems to think about this, based on all the reading on this topic I've done, but a major concern seems to be that TSMC can now charge more.
People express concern about capacity but what about monopolist price increases. Less competition + increased demand = the ability to increase prices.
This could push more cost onto the consumer, contrary to the idea that GF's closure will save us money.
I think it might depend where you're at. Here, Southern Hospitality still exists.. maybe that has something to do with it? Just a wild guess. I know that in Asia, many take service seriously too. Sometimes to crazy levels (Japan).
This is how society has gotten used to cheap and/or slave labor again, for example. People talk about these things as if they were of the past, but they are happening under our noses. Outright human trafficking as well. And this is one just one aberration. Current society is full of things that it gives a slide too, which it used to forbid.
I don't think destroying everything is the solution though. lol
Server CPU market is worth around $18bln yearly. AMD expects to get maybe 10% of that in next few years, so around $2bln.
Nvidia's revenue for 2017 was $10bln. Short term - surely not. Seriously, the demand for GPUs is the same - no matter how many companies make them. If TSMC becomes a monopolist for GPU chips, they can choose who to work with. NVidia has higher margins, so it's likely they'd be paying more for chips. So in short-term going 100% NVidia would make sense.
Long-term is different, because of diversification. It's better to have more clients that make different products. This is a reason why staying with AMD makes sense... unless TSMC is already negotiating with Intel. :-)
That's $18 billion a quarter (probably last year) & not annually! The sever market could actually top $100 in a year or two easily, where do you think Chipzilla's $6 ~ 7 billion (annually) in operating income from DCG came from?