Original URL: http://www.theregister.co.uk/2013/11/05/apple_sapphire_factory_in_arizona/

Secrets of Apple's mysterious Arizona sapphire factory: Our expert whispers all

How Tim Cook aims to shuffle his mountains of money around

By Tim Worstall

Posted in Hardware, 5th November 2013 18:19 GMT

As already noted here at El Reg, Apple is spending some of its vast cash pile on setting up a sapphire production plant in Arizona.

It has not been silly enough to try and manage the plant itself – that will be the responsibility of partner mineral crystal firm GT Advanced Technologies – but it is providing a vast pre-order for the plant's production. That order then finances the plant itself and thus the production run.

As I've covered here before, there is indeed an opportunity to make sapphire significantly cheaper simply by going through multiple iterations of the manufacturing process.

In effect, the idea is very similar to the way in which silicon ingot has fallen in price from $450/kg to some $25 or so now. Make more ingots at wider girths more often and the production cost of each will come down. It's that combination of increasing expertise and mass manufacturing that does it. We've seen it in silicon and the industry predicts that much the same will happen with sapphire.

So far, so newsy: but of course much the most interesting thing about any piece of news is what it implies for the future. And it's possible to make the speculation that the way Apple is going to use its cash pile is in this manner. It's not quite vertical integration – the actual owning and operation of the plants that supply it – but financing advances in production techniques in those upstream suppliers are certainly something closely akin to it.

We've seen some vertical integration, for sure. Apple now designs its own chips based on the ARM core rather than contracting out that design to Samsung. But it is still contracting out the fabbing of those chips, currently to Samsung and (as rumour has it) to TSMC in the future. True vertical integration would mean that Apple would then build its own fab – but I doubt it will.

We could, however, imagine the iPhone giant co-financing a fab on the next generation of production technology instead. And through its financing, via that purchase order, it takes part in the building of the next generation of sapphire production.

It's not unusual in business to find yourself with a chicken-and-egg situation. You know that if you make the effort you can reduce production costs per unit substantially by upping the size of production. But you're really very unsure as to whether you can actually sell all of the increased production even at that newly lower price that is possible (in technical terms you're not sure what is the elasticity of demand for the product is with respect to price). And it is also possible that there are people out there who would just love to gobble up all you can produce at that beautifully lower price but they don't know that you can in fact do this. We do, after all, live in a world of imperfect information.

Then there are those interesting times when the two can meet: the bloke who knows how to make lots more very cheaply and the bloke who will use lots more if it can be made more cheaply. This is really what the Apple and sapphire story is about. Apple has the financial resources to be able to support GT Advanced Technologies through the construction stage and also the appetite for the material once it is being produced.

Be your own market

As I say, this chicken-and-egg situation isn't unusual: my tiny scandium-based corner of the metals world has been facing it for over a decade. We all know that there are interesting resources out there that can be cheaply produced. But to be cheap, any contender would need to produce several times current annual global consumption. So, why bother going large if it just means going bust? The industry is therefore in a genteel pavane, looking for that company that might have that demand if only the price were lower. And yes, I do know who it is in the case of scandium and no, given that my pension is reliant on me and no one else knowing that, I'm not going to tell you that here.

I can give you an example of how Apple could make its sapphire another order of magnitude cheaper though, by, again, throwing its financial weight around and guaranteeing a market for the new production.

Sapphire is made from aluminium oxide (alumina, or Al2O3) and alumina is a pretty cheap substance - a few hundred dollars a tonne sort of cheap. However, the specific alumina used is high purity alumina (HPA) and this ranges from $140/kg to $260/kg. The purity difference isn't all that great.

Metallurgical alumina, the hundreds a tonne stuff, is 99.0%, the other two prices refer to 99.99% and 99.995% material. And you can see that the price difference is vast, well over a hundredfold, just for a bit of purification. Well, OK, just a bit of purification is making it all sound too simple, but there's no technological reason why the prices should be so different.

Scaling the sapphires

It requires, as with the production of sapphire itself, someone to scale up the processes currently used to produce it (HPA comprises perhaps 7,000 tonnes a year of the tens of millions of tonnes per year of alumina). There's no real reason at all why a scaled-up HPA process shouldn't be able to produce at $10/kg or lower.

But you'd want, of course, to make sure that someone was there to purchase that now cheaper and hugely more available material. Which is where someone like Apple could come in. It certainly has the financial resources to be able to make the required purchase commitments, and possibly the hunger for cheaper sapphire too.

Of course, I mention this because I've looked at the process and indeed, have shown that you could start with fly ash from coal burning to do it (with a nice side serving of gallium and germanium to use as dopants on the silicon chips as well). That Apple aren't going to run with this process is obvious, it never responds to anyone at El Reg.

But the basic idea still stands. Apple doesn't do manufacturing, it really does design and retail. Which is great, that's its comparative and absolute advantage. But it also has that vast cash pile and it can use this to improve the productivity of the people above it in the supply chain. This can get them over that chicken-and-egg situation of not knowing whether it's worth investing heavily in more efficient and larger production runs.

This sapphire announcement is only one example: given that Tim Cook is and always has been a supply chain guy, I can imagine the boffins are running their slide rules over where they can indeed use their financial weight to reduce the cost of components to them.

Instead of vertical integration – going back and making their own screens, fabbing their own chips, processing sapphire – it can instead use its ability to pre-pay for large orders to get the experts in those fields to make the advances for its own benefit.

And that sounds like a bloody good use of that cash. ®