A closer look at AI – and how to harness it in the markets
Regular readers know that I've long been bullish on the power of AI...
You don't need me to tell you that it's changing our world. It's transforming entire industries. And it will keep on evolving as the technology gets more powerful.
However, that doesn't mean it'll be a good investment for those who buy AI stocks... or for the companies that are sinking billions of dollars – in some cases, hundreds of billions over time – into the technology.
So I'm mostly watching from the sidelines – but I'm watching closely, given the importance of the sector.
And as I'll detail below, I've found a way to harness AI to make money in the markets...
1) The last time I wrote about AI was on July 3, when I covered the debate between ChatGPT owner OpenAI, led by Sam Altman, and software giant Microsoft (MSFT).
Altman believes we're on the verge of artificial general intelligence ("AGI"), which would achieve (or even surpass) humanlike intelligence. But Microsoft CEO Satya Nadella believes we're still years away.
I concluded that I tend to agree with Altman, whose views also reflect the industry consensus. But one of my readers, Brian M., e-mailed me to disagree:
I am not an AI developer, but I have worked in technology, operating large customer-facing technology platforms for 30+ years, and in my current role, I am working on developing AI operations tools to enable better service performance.
In the case of your comments on AGI, I think that by siding with Sam Altman vs. Satya Nadella, you are buying in to Sam "talking his book." It is in his interest to assert that they have achieved or are about to achieve AGI, because his company needs to secure continual funding, as they are burning billions of dollars every year.
I do not believe that we are anywhere near AGI. In fact I would go so far as to say that the path to AGI does not go through current foundation models. I'd say it's no better than a coin flip as to whether we will achieve AGI in my lifetime.
To make his point, Brian used an analogy:
Let's use driving as an example in examining where we are currently and why it is so difficult to create AGI. We have self-driving cars, which have sensors and cameras and radar that mimic the human ability to observe and respond to sensory input. All of this happens in a relatively contained, rules-based setting.
Now imagine we try to train an AI to drive an F1 car. We could train it on the tracks that are used in F1 races, perfect lines, all the data on acceleration, braking, maneuvering of the vehicle, but put it in the entropy of an actual race and it would fail. Give it a new track to run on and it would not succeed in learning it without additional training. A professional F1 driver can feel the grip of the tires change lap by lap, understand how the car performs with each liter of fuel consumed, know where the rubber pebbles accumulate and how it affects the optimal line, and know when a tiny opening occurs to attempt a pass. Even if you could put enough compute on board the car to run an extremely complex racing model, it still wouldn't beat a human with 3 pounds of meat in their skull.
He concluded:
The reality is that AI using foundation models has achieved extremely high levels of competence, but only in very specifically trained domains. These can in certain cases be better than [humans] and can appear to be approaching AGI levels of intelligence, but they are not generalizable. You cannot take an AI trained to generate pictures and put it in a race car and expect it to operate. Current AI is best described as extremely accurate stochastic parroting.
If OpenAI was to charge its user base the true token cost for the use of the service, no one would replace an employee with AI because it would still cost them more than a human. This is not to say that I don't believe that AI will be something. Just that I think we are going to see it follow the dot-com playbook of massive amounts of mal-invested capital, a bust that washes out a lot of that capital and then a strong economic return for those that survived.
Thank you, Brian! I always enjoy hearing from a different perspective.
2) My friend Harris Kupperman of Praetorian Capital makes a similar argument about "mal-invested capital" in a recent essay:
I've watched as AI went from an interesting parlor trick for making memes, to something that's increasingly integrated into my daily workflow. I use it a lot and get huge value from it. I am not here to belittle AI, it's the future, and I recognize that we're just scratching the surface in terms of what it can do. I recognize all of this. I also recognize massive capital misallocation when I see it. I recognize an insanity bubble, and I recognize hubris.
Harris gives some rough numbers for revenue and margins of AI companies:
The AI datacenters to be built in 2025 will suffer $40 billion of annual depreciation, while generating somewhere between $15 and $20 billion of revenue. The depreciation is literally twice what the revenue is.
Now, here is where it gets complicated as there is no gross margin in the AI game. They're literally giving away the technology and occasionally getting a nickel back for every dollar they give away. Calculated as a gross margin, it would be -1900%. This is the nature of trying to drive adoption and get customers attached to a product. [Venture capital] has a long history of funding this sort of thing, as long as the [return on invested capital] eventually flips positive. With nothing to go on, I'm going to take an optimistic guess here, and say that ultimately, the margins get to positive, and then gradually creep up towards 25%.
He says you would need $480 billion of revenue to hit your target return:
Now, I think AI grows. I think the use-cases grow. I think the revenue grows. I think they eventually charge more for products that I didn't even know could exist. However, $480 billion is a LOT of revenue for guys like me who don't even pay a monthly fee today for the product.
Harris concludes that there just isn't enough revenue to cover the current spend, and we're going to hit a wall with AI soon:
At the end of the day, this AI cycle feels less like a revolution and more like a rerun. I've seen this story before – fiber in 2000, shale in 2014, cannabis in 2019. Each time, the technology or product was real, even transformative. But the capital cycle was brutal, the math unforgiving, and the equity holders were ultimately incinerated. AI will be no different. The datacenters will be built, the chips will hum, and some of the capacity will eventually prove mind-blowingly useful. But the investors footing the bill today will regret ever making the investment. That's how bubbles end – not with a bang of innovation, but with the slow, grinding realization of negative returns, for years into the future. When shareholders finally wake up to the fact that AI isn't generating cash flow, only burning it, the guillotine will fall – on management, on the stocks, and on the broader market that bet its future on a fantasy.
I sent Harris' essay to a friend who's very knowledgeable about Meta Platforms (META). He replied:
Harris is dead wrong because he fails to understand what the real revenue number is. Take Meta – the revenue attached to its AI spend is over $100 billion a year (out of its $179 billion in total revenue), not the $15-20 billion he attaches to the entire industry. As he admits, he doesn't understand technology or the companies that deliver it...
My view is somewhere in between these two...
Across the entire sector, I do think AI is starting to feel like a bubble.
But I also continue to like the shares of tech giants that are harnessing AI, like Meta – and my other two favorites, Alphabet (GOOGL) and Amazon (AMZN). These companies are profitable and dominant, yet their shares are still trading at modest earnings multiples.
3) For those interested in reading more about AI from various viewpoints, here are the best articles I've read in the past few weeks:
- Today's New York Times: A.I. Could Make the Smartphone Passé. What Comes Next?
- Wall Street Journal: There Is Now Clearer Evidence AI Is Wrecking Young Americans' Job Prospects
- Virtualization and Cloud Review: MIT Report Finds Most AI Business Investments Fail, Reveals 'GenAI Divide'
- Post by Liz Ann Sonders on social platform X: When surveyed by the Conference Board, 93% of CEOs said they plan to increase productivity by leveraging tech in order to manage costs
4) In the financial markets, it's easier than ever to harness the power of AI...
The same technology that allows a chatbot like ChatGPT to hold a natural conversation can also sift through millions of data points... spot patterns... and deliver insights that would take humans much longer to uncover. (And that's if humans even find them at all!)
This technology is shaking up our financial markets. And it's creating a whole new way of making money.
Here at Stansberry Research, we've been working on exactly this kind of project. I'm talking about developing a system that makes AI work for you.
It took months and months of research and development. And it involved dozens of coders, developers, and research staff.
When we rigorously back tested this system, it beat stocks, bonds, gold, and even the legendary Warren Buffett himself.
Now, we're finally ready to unveil it in a special presentation... And we're giving you the chance to try this breakthrough out for yourself.
Best regards,
Whitney
P.S. I welcome your feedback – send me an e-mail by clicking here.