Bubbles with a Boy from Oklahoma
- sam00070
- Sep 10
- 10 min read
Observing the Herd
As mentioned in our most recent fund update, I’m outlining what I see as an important time in history, which I’d like to make my opinions known for the record as financial headlines continue buzzing with the term “bubble.”
Bubbles with a Boy from Oklahoma
Global tensions were high, the Asian currency crisis was unfolding, and the Fed cut rates that September, which it would eventually cut two more times to ease the markets by the end of the year, following the collapse of Long-Term Capital Management and Russia experiencing a financial crisis. Large-cap tech had continued to carry the markets for years and a new technology that would forever change the world, the internet, was gaining attraction by investors clamoring to invest in any related IPO. Five separate IPOs that year gained over 200% in their first day of trading with the average first day gain of 20% when looking at roughly 380 IPOs that occurred that year. By the end of the year, the Nasdaq had gained over 39%. The year is 1998. And no, this was not the peak in technology companies. The Nasdaq would go on to gain 85% in the following year, 1999.
The dot-com bubble had years of foundations being laid, continuously being called “frothy” from at least 1995 onward with warnings from some of the greatest value investors like Seth Klarman alarmed by speculative excesses in growth stocks that year, which as a reminder the Nasdaq index closed up 39.9% in 1995, clearly several years before the peak with benefit of hindsight. The “excesses” would go on for five more years with the Nasdaq’s worst year in that period being 1997 when it gained a “modest” 21.6%. Buzz had been brewing around the potential of the internet, exciting investors to continue investing in the space. While it was the exuberance of humans that bid the value up to mind blowing levels, you could point to the internet, the most revolutionary technology the world had ever seen as the main catalyst.
Taking a step back to look at long-term market trends, you’ve probably come across the popular “PE10” chart, which compares the S&P 500’s price to its average inflation-adjusted earnings over the past decade. Most permabears like to utilize it to preach fear. But, since 1995, this metric has signaled “danger” with a value above 20, dipping below only during the 2008 financial crisis. Then there’s the Buffett Indicator, which measures market value against U.S. GDP—useful in the 20th century, maybe, but less relevant now in a global economy. If you would, think of how much of Apple’s foreign revenues contribute US GDP? None. Yet, it’s undoubtedly accounted for in its market capitalization that is embedded in the indicator’s numerator. And sure, these metrics look elevated compared to historical norms, but most of these extremes have logical explanations. So why should investors stay perpetually worried, year after year, just because the numbers look high on a chart?
The thesis of the dot-com bubble proved right. The internet changed the world unlike anything else had in the past, and many of those companies’ stocks are now magnitudes higher than their dot-com peaks. But, it would take investors years to recover and warm up to the idea of technology stocks again, previously being trained to “buy the dip,” but as the year 2000 unfolded, the Nasdaq would eventually decline 80% from its peak to bottom, although never breaching its 1995 high, an important note to reflect on. Looking at the PE10 chart, the markets hit “extremes” in 1995, yet price levels never returned. It would take over a decade and a half though for the index to reach new all-time highs again.
Enter: Artificial Intelligence
Artificial Intelligence (AI), the most revolutionary technology the world has ever seen, in my opinion, at least in terms of scale and scope. As a child, I was captivated by technology and concepts like quantum computing and AI, which seemed straight out of sci-fi. Now, decades later, it’s incredible to see these innovations coming to life. My car, for example, can drive itself and even engage in meaningful conversations about recent events with me thanks to the recent Grok update. When ChatGPT was released at the end of 2022, only a short three years ago, I was speechless. Yet, few people I spoke with had even heard of it. Those with no exposure to AI who I got to try it were unimpressed, unable to grasp how far the technology had come. On the creative side, what started as basic image generation of blobs has evolved into stunning visuals, produced at a fraction of their previous cost, rivaling even the best studios.
For comparison to the 1990s, Mosaic, the first widely popular graphical web browser was released in 1993, followed by Netscape in 1994, which enabled the first secure online transaction and is often compared to ChatGPT due to its mass user adoption. Netscape would reach 12-14 million users by 1995 and 80-90% market share. ChatGPT was able to breach 100 million users by January 2023, two months after launching.
Today, AI is almost ubiquitous, with it being hard to find anyone not using the technology in some way. The adoption rate has been staggering, unprecedented, with AI now deeply integrated into everyday life. It felt like only a year ago, still few I spoke with actually used AI in any meaningful way. But now, even my wife, who still maintains a paper calendar, uses AI daily. And despite its rapid growth, AI’s mainstream adoption owes little to tech giant Apple, whose clumsy rollouts have somewhat dampened enthusiasm. Still, AI’s capabilities are now widely recognized. Many companies tout their ability to reduce procedural timelines by over 90%, with some tech firms even claiming AI writes an increasing share of their code base, quickly becoming the majority.
Unprecedent Adoption, Bested by Speed of Innovation
The only thing even more impressive than the rate of adoption has been the pace of improvement. While I tend to avoid dystopian fears about job loss, I’m optimistic about AI’s potential to improve the world beyond the next decade, especially watching the field of robotics accelerating alongside it. However, I find it hard to not foresee challenges for today’s educated youth trying to enter the workforce over that same time frame, especially those unaware of how drastically the landscape has or is changing, mostly educated by systems of the past. Imagine a law student entering a university today. By the time they graduate, the world will be vastly different, with AI already handling most entry-level tasks. By the time they seek employment, AI will most likely be analyzing cases better than any attorney on earth could dream of. Time spent sifting through discovery documents will take minutes rather than weeks and with better outcomes.
Consider that ChatGPT itself didn’t exist just three years ago and think about how far things have come in that little amount of time. I expect we’ll see similar trajectories in the years ahead. And on the flip side, technology companies themselves aren’t safe either, as many speculate that operating systems and software in general could be rendered a novelty of the past at some point as generative AI will create what you specifically need when you need it, essentially your own, truly personal operating system. I think future generations will be amazed that we use to use keyboards and mice, manually keying strokes or clicking through processes, similar to how we view punch cards for computing of the past.
To me, AI has been a technology deserving of hysteria and deserves the largest bubble the world has ever seen. But I continue to be shocked that it still hasn’t seemed to massively outpace reality, at least yet. Nvidia, the “king” of AI, still trades under a 30 forward earnings multiple. Palantir, which I previously wrote about in February as something that can’t be understood on a traditional value basis, has one of the highest growth rates accompanied by the highest margins within the S&P 500, so it seems justified to receive a premium. It’s a company capable of collectively saving the Fortune 500 tens if not hundreds of billions in expenses annually.
Eye Popping Numbers
Meta’s most recent earnings release made me step back as the numbers were staggering with their CEO stating total spend could reach $118 billion in 2025. The company is now all-in on AI, which I agree with that decision. It is the most revolutionary technology and is worth pursuing at all costs for their business. The dollars spent by these behemoths are real, trickling into many corners of the economy, the construction and engineering firms, electronic component companies, semiconductors, servers, cables, concrete, steel, copper, and, most sought after, power, among many other goods and services. This has been driving up multiples in many of these industries since late 2022, and rightfully so. It’s real. It’s rational. AI is actively and will far into the future change the world forever.
The Fed is gearing up to cut rates on September 17th, just as it did in September of 1998, leaving feelings of déjà vu for many. The markets are primed for excesses. Crypto clearly an easy place to observe this. But in terms of AI, the general public is now beginning to witness its spectacles firsthand, which will only continue to escalate. So, can you imagine watching the Nasdaq soar +85% next year, just like it did in 1999? Do you believe the critics that called AI or the markets in general a bubble for the last several years will be listened to? How many investors that don’t have exposure will finally relent and demand to buy in? Many have pointed to elevated valuations since 2019, using charts like PE10 above, relenting and saying that the market had finally peaked in 2021 as it suffered a sustained drawdown in 2022. However, here we sit again at all-time highs. Most notably, in 2021, consumer accessible AI was not even in existence yet.
So, does a bubble come or are we already in one? Honestly, I do not believe I know more at predicting that than anyone else. With that said, I do believe we are not in one yet. However, the foundation has been laid for one to occur. OpenAI CEO, Sam Altman, implied AI is in a bubble, but he was speaking to the immense size of real dollars being spent, not hypothetical valuations. The parallels to 1998 with multiple +100% IPOs already this year, the indices numbing investors with +20% annual returns as normal, the Fed cutting next month and anticipated to cut more by the year end, large-cap tech carrying the market for a decade, many decrying over extended valuations for years, and the development of the furthest reaching and most impactful technology to the world upon us, I think there’s a high probability of watching the markets evolve into hysteria. And personally, I believe AI deserves a historic bubble and that we may be finally walking out of those doors with the potential to witness numbers and valuations never imaginable, much in the way investors felt in 1999, but now enabled by that same technology that caused the dot-com bubble, the internet.
As history echoes through the markets, it’s hard not to look at today’s AI frenzy mirroring the dot-com era. Valuations are elevated but not unjustifiable, with investors beginning to price in transformative growth in some areas, which will surely lead to inflating the most speculative AI ventures far beyond their achievable fundamentals. This exuberance could eventually become untethered from reality, which then raises the risk of a true reckoning. Like the dot-com bubble, the excess could unravel, plunging the sector into a collapse that stifles investment and ushers in a new “lost decade,” which the markets are long overdue for. Should the Nasdaq surge to dizzying heights, nearly doubling in a single year like witnessed in 1999, it may signal the peak is perilously close.
Yet, for now, I stand in awe, witnessing a pantheon of innovative companies redefine the boundaries of possibility, reaping monumental rewards in the process. And, watching Oracle, what was a $675 billion market capitalization AI infrastructure business yesterday, move over over +40% (over $270 billion) today following its recent earnings release and making Larry Ellison the richest man in the world, makes me believe the market is still finding itself behind. For additional context looking back, Oracle’s split and dividend adjusted peak in 2000 would have been just under $36 per share before falling under $6 by 2002. It now trades well over $300 after this release. I show this as an emphasis that the dot-com era was rational, investors just got carried away in both directions. Much of Oracle’s stock move today can be attributed to the chart below, which allowed investors a glimpse into the company’s expected five-year future. The accounting is already being scrutinized, rightfully so, which Ellison ironically was accused of irregularities during the dot-com bubble in late 2000. But, there is no denying the acceleration of AI-focused investments.
In 2024, Nvidia’s earnings were jarring in terms of the size and scale of the numbers, increasingly difficult to comprehend. Just a few months ago, Meta shocked the world with their budgeted levels of spend. And last night, Oracle blew away investors with almost incomprehensible numbers, posting its best single day return since 1992. I personally believe no one can predict how wild things could get. But be aware, this AI-theme could skew markets, benefiting specific segments or companies far more than others, just like what was observed in the late 1990s, especially if the rest of the economy splits into a recession when adjusting out all AI-focused firms.
But fortunately to that end, AI’s reach is far greater, capable of accelerating and sprawling into nearly every industry. Outside of Nvidia or other tech behemoths, this has only really been observable since late 2022, where stocks from engineering to cooling to electrical component and power providers have all accelerated with only a temporary, though significant, decline seen earlier this year. Since then, almost all of them have hit new highs again. We’ve yet to see massive disruptions from AI in sectors like healthcare, which is undoubtedly around the corner. In some ways, it’s hard to not feel like we’re just getting started. So, I’m in the camp of expecting the AI trend to continue rather than “popping” or collapsing at current levels. But, what do I know? I’m just a boy from Oklahoma.
Comments