Not to play devil's advocate, but...
I recently got my hands on a McKinsey & Co. report about "Beyond the Supercycle: How Technology is Reshaping Resources".
A supercycle is a period of 6% or more of GDP going to a commodity. For this report, they looked at oil, iron ore, copper, nat gas and coal. There have been 2 supercycles: 1976 to 1983 and 2003 till 2014 (small dip for the global crash). Before the 1970's, there was about 2% of the GDP spent on those 5 items. Between the 2 supercycles, it rose to about 3%.
The first supercycle was OPEC and such. I'm sure everyone older than me remembers the oil crash after that.
The 2nd one was China becoming industrial instead of agarian. In the report (about 1/2 through it), unless you can find another China or group of countries to be LIKE China, there is a low chance of there being another supercycle with these commodities.
Their overall gist is that with adoption of technologies, there might not be a bit boon to oil/gas/coal/iron. Copper will remain strong as electricity isn't going anywhere. But for example: ride sharing.
By utilizing ride sharing, the car density can actually decrease. This will impact oil usage COUPLED with the fact that newer cars are more efficient. Ride sharing isn't rolling around in a 1973 Dodge Ram; they are going in a 201X "camry" (or similar), getting 3X MPG (not 12 MPG). The newer cars are also less reliant on iron (steel), so there is an overall decrease. Now plastics make up more of the weight, but the overall result is a lower demand for oil and iron (and nat gas or coal to process things).
Smaller changes like maximizing robotic movements (to reduce electrical consumption) or 'smart' metering of electricity reduce overall electrical demand (think NEST or occupancy sensors). Coupled with a big change in power production (moving from coal to nat gas / solar/wind) means less demand. This is all in step with India and others coming from 3rd/2nd world status to pseudo-1st world.
They have 2 scenarios: 'normal technology' adoption and 'full' adoption. Some of the full adoption ideas are autonomous trucking (they give examples of that in mines and getting 30% efficiency gains) on the road, self-driving cars, high electrical cars, 'cheap' energy storage, etc. In one of the "full" adoption scenarios with cars, electric/driverless cars would reduce the need for 4.5 MMBBL/day of oil. That's essentially half the gasoline we currently use. I would think if you did that to cars, trucks/heavier equipment would not be far behind.
Part of their reasoning for the lack of energy demand for any other 'large' country (or group of countries) is that energy is more efficient now. From 1980 to 2015, China's GDP increased 18-fold. But energy consumption was only 5-fold. Would India and other countries (Africa) use older technologies? Sure, but they won't be the old ones from 1950-1970's USA; they will be 1990-2010's USA (assuming the US keeps up the technological improvement pace).
They mention that "Peak Oil" might be around 2025. Not that we don't have anymore, just that we don't have the demand increase.
I think there will still be oil demand and obviously replacement drilling. But with fracking and the transition from carbon to electrical economy (battery storage, solar, etc.), the demand for HC fuels will take a beating. What's going to happen to those that are in the industry still (with 25+ years before retirement)?
That's what I wonder about.
~egon