The site didn't put my comment below its proper parent, so I just want to make it clear that I was replying to the idea that oil companies have actively held off on building refinery capacity because of the peak.  In addition, isn't it possible that they just underestimated the capacity necessary?  After all, I have heard here and elsewhere that these things take up to a decade to come on-stream, and the meteoric rise of the Chinese and Indian energy appetites has been well within that margin, hasn't it?
Economic growth in China and India is no surprise to anyone who follows such things. Our own history - inefficient use of energy early on in major economic expansion - tells us what to expect from developing economies.

Raw crude demand therefore was always a given. Oil sands never would have been developed in Alberta, for example, if there were not folks looking out decades. When they were first being touted as investments, many short term thinkers dismissed it as a crazy science experiment.

That didn't stop the original investors from pushing forward.

As for refineries... never underestimate the profit motive. If regulations makes gas more expensive, then it will sell at a profitable price point. Capacity constraints raise prices but someone has to be a visionary and step forward to build the damn things.

The only reason you wouldn't build it is if you don't think you need it.

That might be an oversimplification... clearly there are some that have long wanted to build nuke plants but have been prevented from doing so.

Not for much longer I suspect.

Expanding a plant is much easier than building a new one. I dont think there is a single community in the country that wants one of these messy, polluting things. Licensing a new plant in 1975 was much easier than today.  
Apart from the EPA and the NIMBYs, expanding old plants is also faster and cheaper than building new ones.  It also allows you to scale capacity gradually up in response to market demands, ensuring maximum return on your investment.  

And there's always uncertainty...  You'd hate to spend 10 years and several billion dollars on a new mega-facility, only to have it come on-stream right when the economy goes into recession and demand is slacking.