The Channel logo


The Channel Columnists

Contact Mail Follow RSS feed
By | Tim Worstall 22nd October 2014 09:02

It's Big, it's Blue... it's simply FABLESS! IBM's chip-free future

Or why the reversal of globalisation ain't gonna 'appen

Worstall on Wednesday IBM has coughed up $1.5bn to get Global Foundries to take its chip manufacturing arm off its hands as reported here at El Reg.

There are a number of economic implications to all of this, perhaps the most important of which is why the eco-lunatic fantasy that we could reverse globalisation is simply nonsense (of course, whether it is desirable to do so is highly debatable).

The capital requirements in certain businesses are simply so large that only a global supplier can possibly meet them: and that even IBM wasn't in fact a company large enough to support being a player in one of those industries shows quite how large "big" needs to be these days.

What went wrong?

IBM had several basic business problems with its chip-making unit. It was losing money, which isn't quite what you might desire in a subsidiary. A normal reaction to that sort of situation is simply to close it down. Over here we tend not to do that because the redundo costs and so on associated with simply firing everybody can be huge: that's why BMW coughed £500m to rid itself of Rover – it was cheaper than simply paying everyone off.

That's not something that applies in the US environment: and anyway, there's not actually a legal requirement for a company to stand behind the debts of a subsidiary. You can, if you're brave enough, simply shrug and say, 'Well, it's bust'.


Power8 die shot

In this case, the now smaller Blue couldn't really do that: New York State would be apoplectic if it did. And there's a deal at the IBM research site there whereby the State owns the buildings and equipment and IBM employs the people. That would have died if IBM had just walked. It's nice of the local politicians to give them that free money, of course, but free doesn't always, in the long term, mean free.

Also, Big Blue will be wanting to maintain control of the design and intellectual property of the chips it uses, and reportedly plans to plough $3bn on R&D over the next five years.

There's also the slight problem that those plants are the ones that make the chips, the chips with a different architecture and design from everyone else, that Littler Blue puts into its own computers. So “bugger it, we're closing the thing” was never really an option.

And thus the need to cough up that $1.5bn to the people taking it off Big Blue's hands, plus the 10 years supply contract of those chips and so on. All of which is interesting as a series of business problems but their root problem also tells us a great deal about what's wrong with the delusions of the anti-globalisation crowd. For IBM's economic problem was that its own internal (and, of course, including the machines they shipped to clients) demand for chips simply wasn't large enough to cover the costs of the factories needed to make them.

Economies of scale and globalisation

It's a common enough thought that there are economies of scale: produce something in greater volume and the cost of each piece being produced falls. However, there's another bit that economists like to point to: there are also diseconomies of scale. Some time around when you get the sub-sub-committee of the standing committee on comestibles appropriations meeting to decide whether the sub-committee should be getting choccy bics or plain at the tea break, they tend to think that the bureaucracy is costing more than the greater volume of production saves.

However, when that switch takes place is going to depend upon the specifics of the industry: this is the area where Paul Krugman got his Nobel – which was to do with the effects of economies of scale and globalisation on where that point might be.

Krugman delved into things like network effects and showed that it's possible that the economically efficient structure is actually a global monopoly: or at least, it's theoretically feasible that this should be so.

Chip-making isn't quite like that (industry specificity again!) but given that it can cost $5bn and up to create a new fab, obviously there has to be a large volume flowing through it to justify that capital investment – a larger volume than IBM itself can provide and thus their problem. So while this isn't at the extreme of the Krugman case, it's still a useful example of his point.

It's entirely possible that the economically rational size of a company can be larger than one country or one continent can provide as a market – or, in this case, larger than one particular global variation of a technology can provide.

This is something of a problem for those who insist that the economy should retreat to one of living in hand-woven yurts made of locally organic lentil yoghurt.

There's simply areas of the economy that just cannot be efficiently run at that level: the economies of scale available to us are such that we'd be entirely mad to do everything at that local level. Sure, we could say that those that need to be done on a global scale should be done on a global scale while those that should be done locally should be done locally. Our problem then is going to be how to decide which is which. Haircuts are obviously local, computing chips global, great: but how do we decide about everything else?

For example, crocheting those little booties to welcome the new grandbabby into the world is a delightful expression of love, but as a whole we're probably better off with one village in China making 80 per cent of the socks for the world (which really is how it happens today). Gardening for the joy of gardening is similarly great: but the economies of scale of pineapple plantations mean that Yorkshire's never going to be self-sufficient in gammon accoutrements. The method of deciding what should be produced where is obviously going to be suck it and see: or as we might put it more formally, a market in what is being produced and by whom and where.

This really is something of a crimp in the style of the localists, that the method we need to use to decide what is produced locally is a global marketplace. This is quite a long way from why Not Quite So Large Blue had to cough up 1.5 billion spondoolies to get rid of a couple of factories. But it is the same underlying question: what is the correct economy of scale upon which to be producing whatever it is? And to really finish off the eat local movement, it's generally assumed by economists that agriculture has economies of scale right up there with computer chips, rather than down there with haircuts.

Another way to put this is to agree entirely that we should be using appropriate technology, another of those green-style arguments. It's just that the appropriate technology for a great number of things is on a far larger scale than those people seem happy about. ®

comment icon Read 71 comments on this article or post a comment alert Send corrections


Frank Jennings

What do you do? Use manual typwriters or live in a Scottish croft? Our man advises
A rusty petrol pump at an abandoned gas station. Pic by Silvia B. Jakiello via shutterstock

Trevor Pott

Among other things, Active Directory needs an overhaul
Baby looks taken aback/shocked/affronted. Photo by Shutterstock

Kat Hall

Plans for 2 million FTTP connections in next four years 'not enough'
Microsoft CEO Satya Nadella


League of gentlemen poster - Tubbs and Edward at the local shop. Copyright BBC
One reselling man tells his tale of woe