Oh boy, here we go again. Prepare for incoming higher DRAM prices, overpriced GPUs and SSDs. High tech BS, if you ask me.

I don't buy this at all. Pure BS. Any factory such as this will have redundant active backup generators that will instantly kick in. My bet this is part of their scheme to artificially limit the supply and surge the prices. Outage here, outage there, contamination here and there, some dude sneezing in a wrong direction and yada yada. I don't buy this. What's next - a roof leak stops manufacturing process for a week? Or a pigeon that magically flew in?

cryohellinc:

I don't buy this at all. Pure BS. Any factory such as this will have redundant active backup generators that will instantly kick in.

Wait, that's illegal. You're not supposed to know this. /hides.

anticupidon:

Wait, that's illegal. You're not supposed to know this. /hides.

[spoiler]https://cpb-us-e1.wpmucdn.com/wp.wwu.edu/dist/d/4219/files/2018/11/cropped-pika-2i0clzo.jpg[/spoiler]

Still remember when a flood at one WD factory was used as an excuse to raise HDD prices for like 3 years, including Seagate, and they both had record profits. The blatant collusion and gouging in tech gets really tiresome.

Oh dear, prices were low again.

As stated . We don't buy this

Despite being one of the most developed countries in the world, the South Korean power grid apparently isn't praiseworthy.

the cutover to a backup is not clean enough to guarantee that the wafers are cut or masked properly.

Oh christ here come the loony bin comments again.

cryohellinc:

I don't buy this at all. Pure BS. Any factory such as this will have redundant active backup generators that will instantly kick in. My bet this is part of their scheme to artificially limit the supply and surge the prices. Outage here, outage there, contamination here and there, some dude sneezing in a wrong direction and yada yada. I don't buy this. What's next - a roof leak stops manufacturing process for a week? Or a pigeon that magically flew in?

A serious roof leak could stop manufacturing for MONTHS easily. A significant leak could quickly destroy billions of dollars of custom built equipment. It's not quite as simple as getting one day day amazon shipping and plugging it back in.... As anyone who has actually used them can confirm backup generators are not guaranteed to work reliably enough or kick in fast enough for something like this. They are designed to help with long term outages. For short term you need a whole building UPS. Of course this is a moot point to begin with since we don't even know if they even had backup generators but of course that never stopped this community from beginning wild speculations about things that they have no information about or even any background in.

Mundosold:

Still remember when a flood at one WD factory was used as an excuse to raise HDD prices for like 3 years, including Seagate, and they both had record profits. The blatant collusion and gouging in tech gets really tiresome.

Are you claiming the weather is a conspiracy? Or that the flood didn't happen? Or it didn't damage the factory? I mean jesus hundreds of people died in that flood you know. These comments disgust me to read.

Astyanax:

the cutover to a backup is not clean enough to guarantee that the wafers are cut or masked properly.

Thank you! Anyone who has actually worked in the semiconductor industry would know this. Hell I work in IT and we have problems with generator cutovers with our clients all the time. But these armchair analysts have the nerve to sit here and pretend it doesn't happen just so they can continue to believe whatever they want.

NaturalViolence:

A serious roof leak could stop manufacturing for MONTHS easily. A significant leak could quickly destroy billions of dollars of custom built equipment. It's not quite as simple as getting one day day amazon shipping and plugging it back in.... As anyone who has actually used them can confirm backup generators are not guaranteed to work reliably enough or kick in fast enough for something like this. They are designed to help with long term outages. For short term you need a whole building UPS. Of course this is a moot point to begin with since we don't even know if they even had backup generators but of course that never stopped this community from beginning wild speculations about things that they have no information about or even any background in.

I couldn't care less. They make billions, and can't afford BOTH active generators and UPS? I don't buy it. All of those are lies aimed at further inflating the prices.

Please go take a job as a risk management director in the semiconductor industry so that you can realize that you have no idea what you're talking about. Nobody runs whole building UPS on this scale in developed countries. It's ridiculously expensive and the risk simply isn't great enough to merit it. You would need to have consistent outages every year for it to even begin to make financial sense. And the SK grid is just not unreliable enough to justify that.

NaturalViolence:

But these armchair analysts have the nerve to sit here and pretend it doesn't happen just so they can continue to believe whatever they want.

This is the natural consequence of Samsung price fixing and lying multiple times throughout the years.

I worked on a huge data center design 10 years ago, they had dual redundancy measures for both telecommunications and power from "separate" companies. They don't "kick in" like a UPS as they are always on and if any one point fails it just carries on working from the dual stream. Of course all the main hardware inside the date center was designed for dual power/comms too. Conspiracy theory or not, my assumption that the worlds largest multi trillion dollar chip plant is not capable of delivering sustained power for a 1 minute stretch is a little hard to believe. Especially as this has happened before and is not an unreasonable risk to mitigate.

I work in the semiconductor industry growing compound semiconductor crystals and we also have on site cutting and polishing of wafers, honestly even smallest blip in power can cause massive delays in production, whether it's losing the growth of a crystal or having melt back a partially grown crystal that now isn't within spec to start again, this will put you back either several hours or in the case of a complete loss of a crystal several days. Crystals aren't a simple case of load the material into the puller, push a button and off it goes, it can often take several attempts to grow a single crystal over several days and sometimes weeks. Cutting the crystal, making sure wafers are in spec and polishing them is also extremely delicate and time consuming work, any interruption in power can lose entire crystals during cutting, and several hundred wafers during polishing. Having UPS's on everything really isn't an option, they're good enough to keep the control consoles active, but you simply can't have enough of them to keep an entire site active, these machines pull serious power.

So, someone sneezed and it was acted upon like someone has died. We can't have anything cheap so it seems. And some wonder why I'm moving towards console only gaming. 🙄

Keitosha:

And some wonder why I'm moving towards console only gaming

I would say that AMD have played it smart in that regards. Long term vision and innovation on multiple fronts. Intel and NVidia really have milked us into a perilous future for themselves.

I have worked with automation systems for the semiconductor industry for over twenty years and we have several systems in that very Samsung plant. This kind of thing is not all that uncommon. A good amount of the equipment will have a UPS and a fairly seamless switchover when power is lost. However, quite a number of systems do not because they use such huge amounts of power. Things like diffusion furnaces and ion implantation machines use so much power the UPS for one of them would fill a room. One type of system I worked on did chemical processing. If wafers are in an acid tank and it over-etches OR under-etches by only a second or two the entire cassette of wafers can be trashed. That can be millions of dollars worth of silicon rendered worthless in a matter of seconds. We have a UPS on nearly all of our systems and they handle the switchover pretty well but they have to be rather big because robots draw a lot of power. These and other issues often take production down for hours or days at time. If you read the right publications you will see it happens at least once a week somewhere in the world. We had once case at a fab in San Jose, CA where someone inserted an infected flash drive into one of their computers and they were down for an entire week while they disinfected and reloaded backups for almost every computer in the building. That is, nearly all of them that ran windows. We escaped that one, thankfully, but it's just another example of what can happen and the effects it can have.

Windows? As in WindowsCC- Scada?

cryohellinc:

I don't buy this at all. Pure BS. Any factory such as this will have redundant active backup generators that will instantly kick in.

Power outages are a routine event in semiconductor industry, they were very lucky to stall the production for only three days, not months. The power outage problem has no easy solution. Diesel backup generators have a start-up time, so they're not uninterruptable, and battery-backed uninterruptable power supplies with megawatt ratings simply do not exist. There are a few industrial-grade power storage banks in the entire world but they are installed by electrical grid companies, and these banks would be too expensive to build and maintain by individual industries.