This is really bad.
The fabs go dark
Array ( [content] => [params] => Array (  => /forum/index.php?threads/samsungs-chip-fabs-in-texas-ordered-to-shut-down-due-to-power-shortage.13756/ ) [addOns] => Array ( [DL6/MLTP] => 13 [Hampel/JobRunner] => 2000070 [SV/ChangePostDate] => 2010200 [SemiWiki/Newsletter] => 1000010 [SemiWiki/WPMenu] => 1000010 [SemiWiki/XPressExtend] => 1000010 [ThemeHouse/XLink] => 1000670 [ThemeHouse/XPress] => 1010394 [XF] => 2020371 [XFI] => 1040070 ) [wordpress] => /var/www/html )
From Samsung official data, it fabs 65nm-11nm logic.
More than 20 years ago the Philips Semiconductor fabs in Nijmegen lost power, it took a bit more than a day to get things running again, but the most advanced fab was .18 micron. One fab in the US actually caught fire due to a power outage, that took a lot more time to recover from. Now they knew up front, I assume that makes an enormous difference. At the time they could etch away the last process step and redo it.
For the Y2K precautions there were a lot of diesel generators for the Nijmegen fabs, and the fabs were only on standby, not working. So I am not sure if emergency generators are feasible for sudden power loss. It is really expensive for something that happens less than once a year, and unless capacity is used 100% like now (and then...) it is probably not worth it, as long as only the process step is lost and not all lots being worked on.
Important is to tell your customers though, they should hear from their supplier that the fab is down, not read it on the internet! Daily updates on status is strongly recommended if you are not 100% sure you can deliver as promised. (the Philips CEO got summoned and had to apologize, fortunately our department head had done so and was used as good example, some much higher level managers for other products had not and they were in for some shit).