While market observers note a widening of breadth in the market this year, the disparities are striking. The NASDAQ is up over 11%, almost entirely due to the performance of Nvidia# and two other members of the Magnificent Seven, Meta Platforms# and Alphabet#. The Dow is up less than 3%. Within industries, we see similar disparities. Nvidia is up over 120% while Intel is down 39%. Eaton is up 38% while Rockwell Automation is down 17%. JPMorgan Chase# is up 19% while PNC Financial is up 1%. Exxon Mobil is up 17% while Conoco Philips is up less than 1%. I could keep going but you get my point. About the only industry with monolithic consistent performance is the electric utility industry, benefiting from a slight moderation in interest rates recently, and an expectation that AI and EVs will stimulate greater long-term demand.
While overall earnings have edged up, once again the performance is bifurcated. According to data from FactSet, earnings for the S&P 500 companies excluding the Magnificent Seven were down 6%, similar to the fourth quarter of 2023. First quarter GDP growth was adjusted down to only a bit over 1% while the core PCE reading for inflation, announced Friday, was 2.7%. That isn’t quite stagflation, but it seems headed in that direction.
We see the diversity of performance everywhere. High sticker prices are impacting consumers. It isn’t about inflation last month. It’s about the cumulative impact of inflation over the past three years. Traffic isn’t down at McDonald’s because Americans have lost their taste for Quarter Pounders. It’s all about sticker shock. To be sure, there are chains like Chipotle that continue to surge ahead but there are always exceptions. Look at retail where Abercrombie and Gap are making nice comebacks while Lululemon and Nike appear to have lost some of their mojo. In simple terms, running a business in a period of very slow growth is tough. Some succeed; a lot don’t.
With that said, I want to switch gears and talk about AI. There are striking similarities, at least in my head, between what is happening today and what transpired in the early days of the Internet boom in the mid-1990s. We all remember that famous line in Field of Dreams, “Build it and they will come” referring to constructing a baseball diamond in the middle of an Iowa cornfield. And indeed, they came! By the late 1990s, the most valuable company in the world was Cisco#. Essentially, what Cisco made was the plumbing that made the Internet work. Its routers and switches were the hardware necessary for everyone to communicate. Intel made the chips, Microsoft made the operating systems and browsers, and Dell, along with others, made the gateways to the Internet. Back then, there were no smartphones. Apps wasn’t a spoken word.
So why did the bubble burst? Not because the Internet was destined to change our lives like nothing else since the creation of the PC. But because the infrastructure was far ahead of the applications that eventually would change our lives. We saw hints of what was to come in the late 90s with applications like AOL mail, MySpace and Yahoo. But while each was directionally right, each was flawed in its own way.
This is nothing new. It took years for digital cameras to achieve better performance and economic value than traditional film cameras. Early laptop computers were slow, had dim screens you couldn’t see in daylight and were expensive.
So, let’s leap forward to generative AI. ChatGPT and other software programs got everyone excited, so excited that almost $50 billion has been spent or allocated to build GPT, Gemini, and CoPilot, all of which are good starts. But they are probably as close to the end product today as AOL and MySpace were close to 30 years ago. In the mid-90s, the thought of Waze, OpenTable, or TikTok weren’t in anyone’s mind. Great technology breeds new applications but those applications take time to emerge. Facebook and Google didn’t blossom within the Internet’s first decade. Generative AI as we now know it is less than two years old.
In the business world, capital spending is all about capital efficiency. Productivity. You invest to grow, to become more efficient. No one wants to be left behind. Any large business that isn’t at least exploring the advantages that AI offers for better productivity and customer service is probably a business without a bright future. With so much computing capacity moving to the cloud, the cloud providers are in a race to offer clients everything they will need for a future world dominated by the benefits of artificial intelligence. Build it and they will come. And they will. But when?
Look at last week’s earnings report from Salesforce.com. Businesses are taking longer to make investment decisions. It isn’t that the new stuff isn’t working. It’s that the cost/benefit analysis isn’t always as compelling as the hype. Over time, AI features will become pervasive across everything we touch and do, from better responses to simple search queries to the development of future drugs that will save our lives.
Surveys show that a significant number of computer users have used applications like ChatGPT either for business or personal use. But using it doesn’t mean that it has become a core application. Rather, they are tipping their toes into the water. Good answers encourage further use. It’s a process.
In the real world, the hardware comes first, then the software/applications. Once the hardware is designed, building it in scale is relatively easy. Writing complex software takes more time. Early iterations are flawed. They aren’t robust. It took many iterations of Excel to make it work well. Amazon started as an online seller of books at a discount. Even within AI there are two phases. Systems must be trained before they can infer the correct response. Just as we humans go to school for 12 years before we are ready to apply what we have learned.
Disruptive technologies attract many participants. Most chasing a gold rush end up poorer than when they started. In the first two decades of the 20th Century there were close to 3000 businesses chasing the dream of making their version of the Model T. But only a handful got to profitability. The same held true for the PC revolution. Where are Commodore and Atari today?
A decade from now, history books will tell the tale. From 30,000 feet the story line will rhyme with what has happened in the past. The infrastructure will get built quickly, probably overbuilt. Some early pioneers will succeed. The vast majority won’t. Some existing IT giants like Microsoft and Alphabet will find their places in the sun just as IBM and Hewlett Packard did as the Internet and PC revolutions evolved. Many existing giants will disappear like Kodak or Digital Equipment. I think it is doubtful that the hardware winners will be dramatically different from what we see today simply because the capital cost of entry are so high. Look at what’s happening in the EV space to startups that try to enter the business. Tesla made it because it got there first. All the other newbies are likely to fail.
But the real winners in AI ultimately are going to be the ones that develop killer application families; not just one program, but an integrated series of applications, both horizontal and vertical in nature, that get widely adopted and become standards. Think Microsoft 365, for example. Or look at how Amazon has created the infrastructure for others to sell their goods.
Just as the Internet bubble burst around the year 2000 in the stock market, it is likely the current hype will have its comeuppance once reality sets in. As corporations become more deliberate about how they will invest in AI, we will see more sagas like the one Salesforce offered last week. Eventually, that will slow the pace of expansion in hardware.
I want to end by going back to my favorite word, pivot. In a world changing ever more rapidly, standing still is a death sentence. Facebook started as a program for the PC. The pivot to mobility and smartphones created the behemoth we see today. Microsoft pivoted from DOS to Windows, from the consumer to the enterprise, from the desktop to the server to the cloud. A decade from now, keyboards and passwords will be relics of the past just as texting has replaced talking and videos have replaced still pictures.
It’s going to be an exciting world. But there are absolutely going to be speedbumps. When the hype and reality fail to match, there will be disruption. In the stock market, that means a correction. When? I have no idea. But no revolution succeeds without some pain. The one certainty is that of the thousands of startups chasing the dream, most will fail. But there will be a few enormous successes. Most won’t be early pioneers. But they will be identifiable just as Google and Facebook were one or two decades ago. AI is going to change our world. Over the next decade we will be surprised more than once. Keep watching.
Today Rafael Nadal is 38. Billy Cunningham turns 81.
James M. Meyer, CFA 610-260-2220