Question AMD 1Q21 Earnings Results

[Hitman928]

EPS beat ($0.45 vs $0.38 expected)
Revenue beat ($3.45B vs $3.18 expected)

Full year revenue growth projection of +50%.

More to come.

 

[B-Riz]

Direct link to the actual Bloomberg report, well worth the read:

Why Shortages of a $1 Chip Sparked Crisis in Global Economy

To understand why the $450 billion semiconductor industry has lurched into crisis, a helpful place to start is a one-dollar part called a display driver.

www.bloomberg.com

Good read, thank your for the link!

Sooo, we going to crowd fund a US based display driver company with chip fab?

 

[Mopetar]

A shortage in a $1 chip is pretty easy to fix for anyone who has spare wafers. There's obviously a lag between when you start manufacturing and when they come off the line, but this is not a hard problem. Are your $100 chips being held up due to a lack of $1 chips. Stop making so many of the former and more of the latter. So what if if costs $10 now because stockpiling more $100 chips doesn't earn a solitary cent without any $1 chips to go along with it.

 

[Doug S]

A shortage in a $1 chip is pretty easy to fix for anyone who has spare wafers. There's obviously a lag between when you start manufacturing and when they come off the line, but this is not a hard problem. Are your $100 chips being held up due to a lack of $1 chips. Stop making so many of the former and more of the latter. So what if if costs $10 now because stockpiling more $100 chips doesn't earn a solitary cent without any $1 chips to go along with it.

That's not how it works. If for example Apple needs more of those $1 display drivers they can't say "hey TSMC use some of our N5 or N7 wafers to make them".

Those display drivers are made on an older process that Apple probably doesn't buy any wafers for anymore, so they don't have the right wafers to trade off. Even if they did, they can't tell TSMC to make some other company's IP with their wafers. Maybe there's a deal to be made, but what possibly makes sense for TSMC to be a willing middleman for for their biggest customer and what makes sense for TSMC to be a willing middleman for all their customers isn't the same thing.

The companies that have a claim on whatever process the display drivers that Apple and others need may not themselves have any use for display drivers. So they will want to use all that capacity for their own needs and their own products, though I suppose a big enough check from Apple could change their plans. They may have shortages of other parts, on a different process, and the companies that have claims on wafers for those processes will also probably need all that capacity for their own needs. Any sort of "trade" would quickly become very very complex with a lot of moving parts. It is like one of those TV/movie tropes where you need 'A' and someone has 'B' which you can trade to someone else for 'C' which you can trade to someone else for 'D' etc. etc. to finally get yourself to 'A'.

Apple bought the old 8" wafer Maxim fab in California in 2015. I don't know what process(es) it is capable of, what its wafer output is, or if it actually manufactures anything Apple uses in its products or (more likely IMHO) is just for experimenting with devices (like the Apple designed PMICs) before having someone else mass produce them. It doesn't make sense for Apple to own leading edge fabs, but maybe it makes sense for Apple go one step further than designing their own PMICs to owning the production capability for them, as well as other little low value discretes like display drivers (which they can either design or license)

 

[Mopetar]

That's not how it works. If for example Apple needs more of those $1 display drivers they can't say "hey TSMC use some of our N5 or N7 wafers to make them". . .

Someone does have those wafers and they're probably worth a hell of a lot more to the all of the people supply constrained on a $1 part than whatever else those wafers are being used for. If Apple can't give up any of their own wafers I'm pretty sure they can give up some of their money to secure as many as are necessary for their own supply. It would hardly be the first time Apple invested money into one of their suppliers for some type of component.

It's rather doubtful that everyone else using those wafers has a business far more valuable than what the rest of the industry would be willing to pay for those chips. The fact that they used to be $1 seems to suggest that they weren't terribly expensive to manufacture.

 

[moinmoin]

The fact that they used to be $1 seems to suggest that they weren't terribly expensive to manufacture.

I suggest you to read the Bloomberg report:

The bottleneck is that these mature chip-making lines are running flat out. Wu says the pandemic drove such strong demand that manufacturing partners can’t make enough display drivers for all the panels that go into computers, televisions and game consoles -- plus all the new products that companies are putting screens into, like refrigerators, smart thermometers and car-entertainment systems.

There’s been a particular squeeze in driver ICs for automotive systems because they’re usually made on 8-inch silicon wafers, rather than more advanced 12-inch wafers. Sumco Corp., one of the leading wafer manufacturers, reported production capacity for 8-inch equipment lines was about 5,000 wafers a month in 2020 -- less than it was in 2017.

No one is building more mature-node manufacturing lines because it doesn’t make economic sense. The existing lines are fully depreciated and fine-tuned for almost perfect yields, meaning basic display drivers can be made for less than a dollar and more advanced versions for not much more. Buying new equipment and starting off at lower yields would mean much higher expenses.

“Building new capacity is too expensive,” Wu says. Peers like Novatek Microelectronics Corp., also based in Taiwan, have the same constraints.

That shortfall is showing up in a spike in LCD prices. A 50-inch LCD panel for televisions doubled in price between January 2020 and this March. Bloomberg Intelligence’s Matthew Kanterman projects that LCD prices will keep rising at least until the third quarter. There is a “a dire shortage” of display driver chips, he said.

 

[Mopetar]

That report doesn't make sense for what it's reporting on. Shortages in a $1 component shouldn't double the cost of a 50" LCD panel. The rest of it is far more expensive. Instead the cheapest displays or the displays being used for the cheapest products don't get chips because the people selling $1,000 or even $100 products can more easily pass off added cost of paying more to the end consumer.

Additionally the only way in which what's being reported makes sense is if the $1 chip were already consuming all available wafers at a particular fab and couldn't be ported anywhere else. For that to be true, the fab had to have been previously been running close to 100% capacity with no one realizing that regular year of year growth in the market would eventually lead to this very situation. Even if the pandemic created a spike in demand, you're to have me believe that no one noticed the looming production cliff?

Of course now that I think on it a bit more, Bloomberg also had a huge (and completely unsubstantiated report that was denied by everyone) report that China had compromised and infiltrated Apple and several other large American companies. I'm wondering if the quality of their reporting here is similarly crap because it doesn't make any sense. Fortunately it doesn't really matter, because even if they were right, it suddenly makes a lot of economic sense to increase production, port the design to a different process, or any other solution that will alleviate the problem.

 

[moinmoin]

That report doesn't make sense for what it's reporting on. Shortages in a $1 component shouldn't double the cost of a 50" LCD panel. The rest of it is far more expensive.

Think it through before writing. It's not the cost of the missing piece, it's the cost the piece missing is causing. With just-in-time manufacturing if the product can't be completed the whole production comes to a standstill while the running costs continue. With all the cost optimizations applied but no longer being effective the house of cards is crashing together. This is what the auto manufacturers currently suffer on of their own making*, but the excessive demand at the moment affects other industries as well, and the display driver is just one cheap piece that's cost optimized to a point the easiest way to resolve its scarcity is it not being a cheap piece anymore. And there are a lot of cheap pieces that are innocuous as long as not scarce (the other widely reported one last year already was Ajinomoto Build-up Film/ABF).

* On the topic of auto manufacturers another insightful report by Bloomberg (and use your brain to filter the obvious bs in any media, not to filter all kinds of information):

A Year of Poor Planning Led to Carmakers’ Massive Chip Shortage

Near-sighted planning, supply-chain complexities and a tradition of keeping inventories low caused the semiconductor shortage that is now forcing carmakers to idle production lines and straining their relationship with chip manufacturers.

www.bloomberg.com

 

[B-Riz]

Think it through before writing. It's not the cost of the missing piece, it's the cost the piece missing is causing. With just-in-time manufacturing if the product can't be completed the whole production comes to a standstill while the running costs continue. With all the cost optimizations applied but no longer being effective the house of cards is crashing together. This is what the auto manufacturers currently suffer on of their own making*, but the excessive demand at the moment affects other industries as well, and the display driver is just one cheap piece that's cost optimized to a point the easiest way to resolve its scarcity is it not being a cheap piece anymore. And there are a lot of cheap pieces that are innocuous as long as not scarce (the other widely reported one last year already was Ajinomoto Build-up Film/ABF).

* On the topic of auto manufacturers another insightful report by Bloomberg (and use your brain to filter the obvious bs in any media, not to filter all kinds of information):

A Year of Poor Planning Led to Carmakers’ Massive Chip Shortage

Near-sighted planning, supply-chain complexities and a tradition of keeping inventories low caused the semiconductor shortage that is now forcing carmakers to idle production lines and straining their relationship with chip manufacturers.

www.bloomberg.com

The pandemic showed how fragile JIT is, and I have read randomly that a lot of companies are second guessing it now, wanting to setup a more reliable supply chain, even if it costs a little more now.

Heck, even the tariffs on foreign metal in the past few years show how overly dependent industries get on having a cheap commodity instead of a stable reliable source of a commodity can vastly impact the cost to consumers.

Toyota was the about the only car mfg to plan ahead properly, lol. I suppose you don't get to the #1 car maker in the world for no reason.

How Toyota Steered Clear of the Chip Shortage Mess

The auto industry could lose $60 billion in sales this year because of the scarcity of semiconductors. Toyota’s exhaustive monitoring of small suppliers led it to stockpile early.

www.bloomberg.com

But then, this also ties back to AMD, AMD is one of two USA based x86 IP owners and mfgs because IBM wanted a second source of x86 chips. I would bet dollars to donuts that really the federal government wanted a second source of x86 chips before giving IBM a huge contract.

 

[Doug S]

That report doesn't make sense for what it's reporting on. Shortages in a $1 component shouldn't double the cost of a 50" LCD panel. The rest of it is far more expensive. Instead the cheapest displays or the displays being used for the cheapest products don't get chips because the people selling $1,000 or even $100 products can more easily pass off added cost of paying more to the end consumer.

Then I guess you don't understand economics. If a shortage of display drivers means that fewer LCD panels are made, the ones that are made can be sold for more money. It is basic supply and demand.

There isn't some magic reallocation of $1 display driver chips across the entire industry so only the most expensive products get them. The ones who get them are selling their products, whether they are cheapo Chinese 50" LCD TVs. a 10" LCD in a "smart" fridge, the in-dash LCD for navigation on a new car, or whatever.

 

[DAPUNISHER]

Think it through before writing. It's not the cost of the missing piece, it's the cost the piece missing is causing. With just-in-time manufacturing if the product can't be completed the whole production comes to a standstill while the running costs continue. With all the cost optimizations applied but no longer being effective the house of cards is crashing together. This is what the auto manufacturers currently suffer on of their own making*, but the excessive demand at the moment affects other industries as well, and the display driver is just one cheap piece that's cost optimized to a point the easiest way to resolve its scarcity is it not being a cheap piece anymore. And there are a lot of cheap pieces that are innocuous as long as not scarce (the other widely reported one last year already was Ajinomoto Build-up Film/ABF).

* On the topic of auto manufacturers another insightful report by Bloomberg (and use your brain to filter the obvious bs in any media, not to filter all kinds of information):

A Year of Poor Planning Led to Carmakers’ Massive Chip Shortage

Near-sighted planning, supply-chain complexities and a tradition of keeping inventories low caused the semiconductor shortage that is now forcing carmakers to idle production lines and straining their relationship with chip manufacturers.

www.bloomberg.com

Good stuff, and thanks for sticking to the facts on the ground. Instead of conjecture, and fallacies like shoot the messenger.

 

[zinfamous]

Think it through before writing. It's not the cost of the missing piece, it's the cost the piece missing is causing. With just-in-time manufacturing if the product can't be completed the whole production comes to a standstill while the running costs continue. With all the cost optimizations applied but no longer being effective the house of cards is crashing together. This is what the auto manufacturers currently suffer on of their own making*, but the excessive demand at the moment affects other industries as well, and the display driver is just one cheap piece that's cost optimized to a point the easiest way to resolve its scarcity is it not being a cheap piece anymore. And there are a lot of cheap pieces that are innocuous as long as not scarce (the other widely reported one last year already was Ajinomoto Build-up Film/ABF).

* On the topic of auto manufacturers another insightful report by Bloomberg (and use your brain to filter the obvious bs in any media, not to filter all kinds of information):

A Year of Poor Planning Led to Carmakers’ Massive Chip Shortage

Near-sighted planning, supply-chain complexities and a tradition of keeping inventories low caused the semiconductor shortage that is now forcing carmakers to idle production lines and straining their relationship with chip manufacturers.

www.bloomberg.com

I think this happened to Toyota some years ago as well, right? The whole JiT chain was on the verge of collapse because of some hit to one or two parts that were manufactured at a few plants in Japan, and they actually shifted production by ordering an all-hands-on-deck throughout their suppliers to convert their factories to make those parts. I think they were something like 2 weeks ahead of entire production halting for a month or so, and they beat it by about a day or so?

 

[moinmoin]

I think this happened to Toyota some years ago as well, right?

Yeah, @B-Riz linked an article about its case above. In short Toyota's JIT production was heavily affected by 2011's Tohoku earthquake and tsunami. The lessions it took from that now allows it to navigate the current crisis mostly unscanted.

 

[Mopetar]

Think it through before writing. It's not the cost of the missing piece, it's the cost the piece missing is causing.

That's why it's ultimately self-correcting in terms of overall resource utilization. That means people using that previously $1 part will value it a lot higher, which means the people producing it can sell it for more, which allows them to purchase more production capacity for it. From the perspective of people who need to buy it, the ones that can pass off the added cost most easily can afford to pay more and will get allocated more of the limited supply of those parts. If you're making something that you can only sell retail for $20, then those will stop getting manufactured when compared to something that can be sold for $2,000 because the relative price increase or ability to reduce margins is felt less in the product that's priced higher. Even if you're the company making the $20 gadget and manage to have a lot of spare chips that are now in high demand, you might make a lot more money selling those at a markup to other companies than you would making your own $20 gadget with them.

There isn't some magic reallocation of $1 display driver chips across the entire industry so only the most expensive products get them. The ones who get them are selling their products, whether they are cheapo Chinese 50" LCD TVs. a 10" LCD in a "smart" fridge, the in-dash LCD for navigation on a new car, or whatever.

There isn't any kind of magic reallocation, but they will tend to go towards those who are willing to pay the most just like any other time where there's a shortage. We're seeing that happen now where there aren't enough GPUs to go supply the total demand so more and more of them are going towards miners who will pay higher prices. The same will happen here as well over time as prices shift in response to this and we figure out which products that need these chips the end customers are actually willing to pay more for due to the constraints. Maybe those aren't the products that someone suspects necessarily, but in the end the market will adjust an let producers know based on purchasing habits.

 

[Doug S]

That's why it's ultimately self-correcting in terms of overall resource utilization. That means people using that previously $1 part will value it a lot higher, which means the people producing it can sell it for more, which allows them to purchase more production capacity for it. From the perspective of people who need to buy it, the ones that can pass off the added cost most easily can afford to pay more and will get allocated more of the limited supply of those parts. If you're making something that you can only sell retail for $20, then those will stop getting manufactured when compared to something that can be sold for $2,000 because the relative price increase or ability to reduce margins is felt less in the product that's priced higher. Even if you're the company making the $20 gadget and manage to have a lot of spare chips that are now in high demand, you might make a lot more money selling those at a markup to other companies than you would making your own $20 gadget with them.

There isn't any kind of magic reallocation, but they will tend to go towards those who are willing to pay the most just like any other time where there's a shortage. We're seeing that happen now where there aren't enough GPUs to go supply the total demand so more and more of them are going towards miners who will pay higher prices. The same will happen here as well over time as prices shift in response to this and we figure out which products that need these chips the end customers are actually willing to pay more for due to the constraints. Maybe those aren't the products that someone suspects necessarily, but in the end the market will adjust an let producers know based on purchasing habits.

There isn't a spot market for everything like there is for oil. Companies that need those parts as part of their supply chain will have contracts in place that specify price. If they need more well then too bad if their supplier isn't able to get more.

If you're a company with a contract for delivery of 100,000 display driver chips you put into cheap products, sure they would be worth more than $1/ea if you sold them to someone else. But where is the marketplace where you're going to be listing them - and I mean somewhere trusted that a major company would be willing to buy from for critical products (i.e. not Alibaba) But it will be the customers with the contracts for delivery who would have to figure this out, the supplier can't decide to break existing contracts because someone calls them up willing to pay $5/ea or whatever.

I consulted on a startup about 10 years ago (partially compensated in worthless stock options, luckily I insisted on most of my compensation in cash) that was trying to create something like what would be required to make these sorts of deals. Sort of a "B2B eBay", but with reputational scores based on "who you know / who vouches for you" so also sort of like B2B social media. Well, at least in one iteration of the plan. The main reason it failed from my perspective were that the founders couldn't agree on its focus so it tried to be too many things. In the end the VC decided to cut them off before they could produce anything viable.