Banks bet on bigger, faster super computers
During the last financial crisis, JP Morgan realized it didn't have the computing power to keep up with the collapse. Bankers didn't know what their portfolios of bonds and complex derivatives were worth at any given moment or how they would change in value as all hell was breaking loose. The computers that were supposed to offer up those answers were choking on the complexity of the financial instruments JP Morgan and other masters of the universe had created.
In the fall of 2008 it took up to 8 hours for JP Morgan's fastest machines to do the math. So traders would know what their positions were worth the day before - but could not anticipate what was coming- or even know what they were worth at the time. And if anyone made a programming mistake there was no time to re-run the calculations before the next trading started for the day, which meant the bank was flying blind. One exec compared it to driving down a highway at 90 miles an hour by looking in the rear-view mirror.
Since then, JP Morgan has invested in a newer, faster machine. Bonds are just as complex, but now running the same calculations on its book of business takes less than 4 minutes and some calculations that used to take more than an hour finish in 12 seconds.
So, is it possible supercomputing might help us avoid a financial meltdown this time around? Sounds like wishful thinking, but maybe this time the bankers will see the cliff coming before we all careen off it into the abyss.