This discussion has become all the rage in the news media and it’s interesting to see where everyone is landing. The investing community is willing AI into functional and profitable existence, though it’s unclear whether this will ever be possible.
Then again, it took 20 years for Amazon to turn a profit.
The capital being poured into AI infrastructure is unlikely to get a return on investment anytime soon if at all. The amount of money being funneled into a narrow corridor of computing capacity is different from the gold rush of internet companies that were built to run on the capacity of the internet. Those companies went belly up but the internet remained and the real power players emerged over time. And the rest of the economy found ways to incorporate it into their processes and systems.
So, is this the dot-com bubble, or something more systemic like the housing crisis?
I think the answer to all of the AI questions is “yes, and.” The main AI companies such as OpenAI and Anthropic are so well capitalized by investors they don’t need to turn a profit. In this way, they’re a lot like Amazon was during the unprofitable decades. So they’re not going to collapse. Nvidia is playing a risky vendor financing game because it’s very likely the demand is there, but the ability to pay their exorbitant prices isn’t. So this will become a fractional game with certain giants pulling away from the pack and getting even bigger, and a constellation of companies who utilize AI to enhance their systems and tech stacks.
Then again, the valuations are already over-inflated, so when the correction comes it could go from isolated to systemic because our policy response might be insufficient to stop the vicious cycle of negative sentiment.
Naysayers who debate the hype around AI are probably right about whether we’re close to achieving artificial general intelligence (AGI). I’m beginning to come around to the realization that AI is the wrong term for what we have now. What we have now, which is incredible, is advanced automated workflows built on sophisticated probabilistic models. These workflows are making tasks easier and faster, and they’re being built into SaaS systems like Salesforce and other platforms to increase productivity and generate more consistent outcomes. These systems are already in use in the medical and diagnostic field, supply chain and logistics, myriad creative disciplines, call center and support services, and a thousand places I can’t even imagine.
Then again, the primary architects of the field such as Geoffrey Hinton—known as the Godfather of AI—and ethicists like Roman Yampolskiy—who coined the term “AI safety”—are convinced that AGI is coming soon and has the potential to eradicate most jobs and will likely start a world war on its own. Cool.
This level of tension between competing narratives screams out for government intervention. Oversight. A cooling down phase. Any attempt at regulation. Instead, the scions of Silicon Valley, all of whom have apparently gone to the dark side, are embedding in the White House and we are moving forward with reckless abandon. Dangerous times indeed.
In terms of the collapse or the bursting of the AI bubble, I have mixed feelings. I think the named players we know are the ones we’re going to know for quite some time. Outside of Nvidia, these others are pre-IPO so the investor risk is isolated to Wall Street and private capital. The so-called “Magnificent 7” might take a hit but as we’ve covered in recent months, the rest of the corporate landscape is humming right along.
The incestuous AI funding cycle reveals that investment risk is just as concentrated as the risk to the overall market. In other words, if the top handful of AI players lose significant value in a market shock, it shouldn’t bleed out to the rest of the market on pure fundamentals. But it would have a significant impact on confidence and probably lead to a badly needed correction.
But the real worrisome thing to me is this…
Companies are performing well in terms of earnings and cash on hand. Yes, they are overvalued, but the business cycle is robust. And yet, the labor market is shrinking by the day. Challenger, Gray & Christmas indicated mass layoffs in their projections. Corporate America is learning that it can do more (or the same) with less. So if this is the revolution before the real revolution, what does that mean? AI doesn’t have to realize its full potential because automation capability is growing exponentially as a result of the capital investments into sheer computing power. Food for thought.