Global trade in food and luxuries has been going on for centuries, but the next big commodity looks set to be invisible, yet essential. As we change how we react with each other and the world around us, artificial intelligence is destined to become far more than a future-gazing distraction for geeks and scientists.
"AI is going to be a commodity, available on mobiles and computer terminals," says technology analyst Peter Cochrane, "It may well be the most vital of all commodities, surpassing water, food, heat and light. Without it, we will certainly not survive as a species."
One of our problems is data - masses of it. A few hundred years of scientific inquiry and the invention of the data-generating and sharing mechanism that is the internet has left reams of crucial information unused and unanalysed.
Medical teams need to process and cross-reference both personal data and that from previous patients and studies, while financiers find it ever harder to wade through enough data to make coherent predictions of movements on the stock exchanges.
What these industries need - and will get - is "machine learning" from cloud-based computers capable not just of spotting trends in data but making predictions and learning from conclusions. AI is not about sentient robots, but machines that mimic our organic intelligence by adapting to, as well as recognising, patterns in data. AI is about making machines understand.
Watson is such a supercomputer. Comprised of at least 90 servers and 16,000 gigabytes of RAM, Watson - developed by IBM - can process the equivalent of a million books per second. And using revolutionary smart learning software called DeepQA, Watson can interpret the meaning of questions. So what makes Watson different to other supercomputers? "Hardware-wise, there's not much," says Cochrane. "It's the more than 300 algorithms it uses which makes the difference."
There are hundreds of other supercomputers, but none are as good as Watson - not by a long way.