Growing up, a friend’s dad would often make statements on sports figures as if they were novel observations. Commenting on Michael Jordan (who at the time was probably in his sixth year in the league and had already won the NBA scoring title a handful of times), I remember him saying, “You know, I think this Michael Jordan is going to be a pretty good ball player.” Today, that would be like talking about Artificial Intelligence’s potential and saying, “You know, I think this AI stuff is going to be a pretty big deal.” While factually correct, neither would qualify as bold statements at the time they were made. Perhaps the more interesting question when it comes to Artificial Intelligence would be, where AI is going at this point, and what’s next in its evolution.
To answer that question, let’s first start with the money that is being spent to build the underlying infrastructure to train, develop and deploy AI models. It is estimated that the aggregate capital expenditures over the next twelve months from just five large publicly traded hyper-scalers, Microsoft, Alphabet, Amazon, Meta and Oracle, will top $375 billion. That means that roughly 40% of the estimated annual $1.0 trillion of cap-ex in the S&P 500 is currently going towards AI infrastructure. The logical next questions might be how we are going to use all this AI capacity and whether the hyper-scalers will ever earn attractive returns on their investments.
For that, let’s look at JP Morgan, which is reported to spend roughly $18 billion per year on information technology, of which $2.0 billion is earmarked for AI use case development within the bank’s many departments and processes. It is estimated that AI, across 450+ internally-developed use cases, is now saving JP Morgan employees 15 million hours annually, has resulted in nearly $2.0 billion in productivity gains and has saved something approaching $1.5 billion in preemptive fraud detection losses annually. How ever you measure it, that sounds like a pretty good return on investment. Now, who gets to keep all those gains? Is it JP Morgan and its shareholders, is it customers, is it the hyper-scalers who built the infrastructure? The answer is probably all of the above. But it is real, and it is happening now.
I think we are all waking up to the potential of AI—both personally and in the businesses, organizations and enterprises we are involved in. We are just starting to scratch the surface. We believe we are entering a period where attention will begin to shift from the hyper-scalers building massive AI infrastructure to the use cases that will now permeate corporate earnings reports and success stories throughout the rest of the economy.
And as far as jobs go, we believe Nvidia CEO Jensen Huang summed it up well when he said, “You won’t lose your job to AI, but you may lose it to somebody who is using AI.” So, the time is here to lean in, experiment, try to automate something that is a nuisance to you, or if you have never used it before, try asking a question in the prompt window of any of the popular AI chatbots. You might be pleasantly surprised. Because, as my good high school friend’s dad would say, “This AI thing looks like it’s going to be a pretty big deal.”