Hacker News Reader

Comment by simonbarker87 | Hacker News Reader

I’ve been in and around manufacturing for the better part of a decade. I’ve run a small factory myself and have friends and family working at all levels of some of the largest factories in Europe.

At the scale these things work out there’s no way to really “have a backup” not in the true sense that backup means. If you build a factory you use a factory - otherwise you’re wasting masses of money.

What you should do is diversify and leave the option to quickly (relative terms here, still many months) ramp up production in other factories should one have an issue.

Which is what most companies do. Manufacture rewards scale though so the diversification comes with its own set of problems (usually in logistics and stock management).

So they do think about it for sure but due to the physical nature of things it’s not a case of “have a backup and decide to start going one day at a new location.”

View on HN

That is a thoughtful reply. Working in IT spoiled me, and I like having a backup or failover server or data centre but for Apple factory. But, it makes sense now, knowing the scale of this Apple iPhone factory, it is next to impossible to build the backup factory. Thanks!

There are aspects of this[1] book that I really enjoyed. I have no hardware/factory experience either but I had taken some operations research courses in college. In general, holding onto inventory more than you need is a waste of time, space, cost and introduces other complexities as well - like a defect identified even a little bit later in the process could lead to huge waste if you’re holding on to too much. But of course not having any sort of reserve also leads to issues.

During the pandemic one of the only car companies that didn’t have major disruption was Toyota and they were the pioneers of this model of operating but they also understood the need to keep reserves of critical components.

In the case of Apple, they do have a backup plan - the lower end models were already being manufactured outside China and they could scale that up while they scramble to fix the Pro lines.

[1] https://www.amazon.com/Goal-Process-Ongoing-Improvement/dp/0...

Consider that each new iPhone and iPhone size required retooling entire portions of the manufacturing line. And by retooling I mean physical tools - the shoots that deliver the right sized screw to the individuals assembling that part, soldering techniques changing and requiring different equipment, spot weld equipment etc.

Building a “backup factory” would also require that massive investment every year to revamp the assembly line, and never get used.

Then there’s the thousands of components that go into an iPhone, and their deliver schedules perfected down to 30 minute windows - you’d have to recreate that, and just ship, what, nothing? To the backup factory?

You don't build a backup factory to sit there unutilized.

You build four large factories instead of one mega factory and run all four. Similar number of machines, similar investment to update them - each is big enough for efficiencies of scale - but if one gets shut down by the local government or has a massive fire or there's a coup you've still got 3/4 of your production going.

>At the scale these things work out there’s no way to really “have a backup” not in the true sense that backup means. If you build a factory you use a factory

Not really true at all... I mean, if you want to scale, just build another line?

I worked on Capex at Tesla and there was explicit talks of backup factory's. Elon was always talking about WW3 during covid and how we need to massively scale up our operations, have backup factorys in every region or we will not make it.

When we had issues scaling our GA3 line in fremont, we cut the conveyors off the line, built a tent in the parking lot and build a new line in a month. (Granted, 14 hour days every single day) but it's generally easier to build a new line with the lessons learned than try and fix the old one.

The scale here is completely different, and I’d bet on Tim’s supply chain chops over Elon’s every single time.

Appreciate the input. I would say that another line at the same location isn’t a back up, that’s scaling and you can choose to run that line or not. It’s not uncommon to run a second or third line at lesser capacity than the primary line (the Nissan plant in Sunderland flex the secondary lines a lot for example).

Much of my manufacturing knowledge comes from automotive and it’s very very very rare to run a literal back up factory (as in an unused facility in a different location that’s good to go to replace another factory). They take so much man power to run, so much time to set up that it’s not economically viable to just have it on standby as a back up.

Even if you did there would be masses of business pressure to utilise the capacity.

That’s why I pointed to diversification instead, Nissan make the Qashqai in 3 locations, UK, Russia and Japan (from memory) which gives them a level of resilience but they aren’t back ups.

If Sunderland goes down then Russia and Japan plants can’t just step up and fill in.

Setting up a new factory for the DBX nearly killed Aston Martin (May still do arguably) and that was one being set up ti be used. Not sat idle waiting.

Re: Nissan

And, somewhat ironically (do not put production in dictatorship countries), Nissan lost one of these locations. They sold Russian plant for $1.

The factorys tesla is making can make all models in different regions, with nothing required to be shipped between the factorys.

I didn't mean back up as in they sit idle, but are able to produce every portion of the car. (Well.. technically fremont can't because it needs batterys from Reno, but thats besides the point)

The Berlin, Texas, Shanghai factorys are essentially 'back-up factorys'.

Reply on news.ycombinator.com