Peak oil is not about "running out of oil," it’s about the peak rate of oil production. It’s not the size of the tank which matters, but the size of the tap.

When the production rate of oil reaches its geological limit and begins to decline, the world’s economies will be forced to live within a shrinking, not expanding, energy budget. The economic impact of peaking oil production is what concerns us, not the amount of oil yet to produce, because all economies depend on continuous growth. We won’t “run out of oil” for another 100 years or more, but it will be produced at ever-declining rates.

This is an essential concept. Talking only about the number of barrels of oil that might exist somewhere, without also talking about the rate at which that oil can be produced, and when, entirely misses the target.

Oil production rates generally follow an irregular bell-curve shape. It is simply the nature of petroleum extraction that it gradually ramps up, reaches a peak or short plateau (sometimes with a secondary peak) when roughly half of the recoverable oil has been produced, and then declines. This observation has been made in thousands of oil fields (and oil producing nations) worldwide, and is named "Hubbert’s Peak" in honor of the geologist who first described it, Dr. M. King Hubbert.

For the world, ASPO-Ireland’s working model of past and future oil production looks like this:

This model is based upon a detailed study of all the world’s major oil fields, with all forms of petroleum taken into account.
According to the June 2008 revision of this model, the peak of all petroleum liquids—including heavy oil from Venezuela, deepwater oil from the Gulf of Mexico, oil from the Arctic and Alaska, and natural gas liquids—is this year, 2008. But the exact date of the peak is almost irrelevant when considering the implications of peak oil.