AI Will Create A Futures Market For Compute

In my previous essay, I argued that datacenters resemble futuristic central banks in that cloud service providers can effectively “print” money in the form of cloud compute credits representing the ability to use a unit of future compute time. This means that cloud service providers, who manage the majority of global data centers, are effectively “central bankers” for the future, who can introduce financial derivatives representing future compute cycles that can be packaged into a wide range of financial derivatives including loans, bonds, futures and options.

In this article, I will talk a little bit about why skyrocketing demand for GPUs for future AI computation purposes will necessitate the formation of a national “Futures Market” for GPU access. By financializing GPU access into a tokenized “Real-world Asset” or “Synthetic” instrument, governments of the world are likely to be incentivized to improve the efficiency by which scarce GPU resources are accessed.

The best pricing known to economics? Market pricing. GPU cycles should be exposed on an open market place for buyers to place bids. Furthermore, GPU cycles should be tokenized and made available as Futures, enabling customers to pre-buy GPU compute at today’s prices, knowing they will need it in the future.

A GPU / compute futures market is a natural outgrowth of the skyrocketing demand for GPUs. Only a futures market is a logical solution for how to structure the market for compute for optimal pricing and access.

As AI continues to skyrocket, most of the trading for these futures will be performed by intelligent machines. It isn’t a matter of “if” there will be a worldwide futures market for compute, and derivatives, it is “when.”