What could possibly go wrong...
Excellent, so we've got the ability to have a much wider supplier base as a result of globalisation and also the possibility of that supplier base clustering and actually making it more, not less, vulnerable to geographically specific shocks. How they play out against each other just sort of depends, which is what the answer is so often in economics, “it depends”.
The third thing we've got is economies of scale.
Now all you technical types are entirely familiar with this, although you tend to call it network effects. Windows sells trillions because everyone uses Windows. Facebook has youth by the the short and curlies because all youths are on Facebook. Paul Krugman got his Nobel largely for work in this very field: when might globalisation, the freeing of trade, lead not to more suppliers but to fewer? When might the economies of scale (the network effects if you like) just keep going so that we end up with not hundreds of producers, but only one, or perhaps just a handful?
Economists have always been aware of economies of scale, of course, as are most people. But economists, unlike most people, are also aware of diseconomies of scale. Sure, making a million of something might be cheaper per unit than making 100: but what about the costs of managing an organisation capable of making one million things? The accretion of layers of bureaucracy perhaps, or the slow down in technical innovation from being part of a multi-layered organisation (vide just about every tech company - even the likes of Google and Oracle - that buys in bright ideas that would never survive its own development bureaucracy). What Krugman mused upon was what happens if the economies of scale are so great as to continue to overwhelm these diseconomies even as a company moves to globally bestriding colossus?
OK, put like that it doesn't sound all that amazing, not Nobel level stuff, but you get the gongs for (a) being first to muse along these lines and (b) being able to prove what would happen. Which is what he did do and he was quite right too.
We can put this is Adam Smith terms too: the division and specialisation of labour (the dreaded pin factory for anyone who ever did any economics). If we split up production into multiple stages then each worker is going to get better at their specific task (both Smith and Marx warned they'd get bored stupid too but that's another argument) at hand. So, now we've got a team of specialists, each doing their part of the divided labour, we can get more production from our division and specialisation. What we've learned more recently is that this also applies to things that aren't just direct labour. Teams of hard drive designers seem to be better at designing hard drives than people who flit from graphics cards to hard drives to motherboards.
No, this isn't absolutely true, for there are still huge benefits to cross-fertilisation across disciplines, but it is more true than we had perhaps thought at one time. Which is where we get these ever increasing returns to scale from in some industries, where this is indeed true.
I dimly recall the time of the Kyoto earthquake in 1995 (forgive the dimly, I was more worried about being paid for a delivery made two days before the quake; we got the cash, just, a transfer on the last working day before it hit) and how the computer world was in uproar. Yes, death and destruction, appalling, but we found out that the only company in the world that made the glue (this is the dimly part, I think it was glue) to stick chips into their casings was at the bottom of the rubble.
Which is sort of but not quite where we are with the deadly Thai floods and the looming hard drive shortage.
Globalisation has made it possible to allocate work across the resources of the planet. We get our hard drives assembled by the Thai women making €200 a month in factories (as opposed to a possibly less pleasant career earning higher wages in the bars) but there's also these other effects as well. Economies of scale seem to be still large, which is why the number of drive manufacturers has gone, according to one estimate, from 20 to only 7 in the past few years. We've also got these clustering effects, where although those manufacturers might have multiple manufacturing locations, they tend to end up putting their factories close to those of their competitors.
How it all works out in any specific industry depends upon the specifics of that very industry: and that working out might well mean that globalisation does not lead to a larger number of suppliers nor even to greater reliability of supply. It, as so often, just depends. ®