How to become chip builders

The idea that Bangladesh should become a producer of computer chips is an interesting one. We know that advanced and rich economies produce computer chips. Therefore, if we produce computer chips, we'll be an advanced and rich economy, right? 

Well, not so much really. It doesn't quite work that way. It's true that if Bangladesh were able to produce computer chips in volume then that would be evidence that Bangladesh was and is an advanced economy. But it's evidence of this fact, not something that causes this to happen.

The issue is capital. Chip production is hugely and massively capital intensive -- it costs $10 billion to set up afabrication plant (“fab” in the jargon). To set up one that's at anything like the leading edge of the global technology that is. Which links into the second problem -- it's only the fabs that really are at that cutting edge really making a profit. 

If you build a fab which will make older-style chips, then sure, you'll make a reasonable enough operating profit. The whole point is, after all, to turn what is essentially sand into a very valuable chip. But the plants that were cutting edge three years ago (the cycle is about as fast as that) are still operating, still producing the level of chip they were designed for. They made huge profits in those first two and three years of production. That's when there was a global scarcity of being able to make that level of chip -- margins were insanely high. 

More such fabs come online, making to this standard becomes a commodity business, commodity profits are made. But here's that little problem -- commodity profits don’t pay back that $10bn. That capital cost has to be recouped in those first couple of years, from those scarcity driven profits. 

The logic of this is that to enter the chip business it's necessary to enter at that leading edge. Enter at any of those older, commodity, levels and you're competing against plants that have already recouped their capital costs. That's not something that's going to work.

It's also true that the capital is by far, far, the greatest cost in the business. TSMC, the world's largest independent chip maker, invests tens of billions of dollars a year just to stay in business. Wages are such a small part of the overall cost that being in a low wage country really doesn't confer all that much advantage. 

So, how would Bangladesh enter the chip industry? It would be necessary to enter at that leading edge -- old tech just doesn't work here. That means many billions in capital must be invested in even just the one fab.

This then runs up against the obvious point -- Bangladesh doesn't have that much domestic capital. Further, what there is can be put to vastly better uses. In a country short of capital, it is far better to use what there is in a few millions here, a few tens of millions there to improve matters, rather than billions upon billions on just the one factory. 

Now it may well come to pass that Bangladesh will become a computer chip producer. But it won’t be part of the path to becoming rich, it will be as a result of having become rich. When there's enough domestic capital to build a fab then Bangladesh will indeed be rich -- because a good proof of being rich is that there's a spare ten billion of capital that can be used to build a fab.

The trick is not to try to work through how to build a chip industry in Bangladesh. Rather, it's how to make Bangladesh rich enough that it can finance a chip industry. More difficult, true, takes longer, also true, but that is the right way around to be thinking about it.

Tim Worstall is a senior fellow at the Adam Smith Institute in London.

Source: Dhaka Tribune.