Wednesday, November 19, 2025

Party like it's 1999

 Prince recorded the song "1999" in 1982 as the title track of an album of the same name. The song was inspired by a TV show which speculated on what could happen after the turn of the century. Prince thought that the planet as a whole would experience a rough time at the turn of the millennia but wanted the song to give people hope. The song, when initially released reached #44 on the Billboard Hot 100 chart and a later release moved up to #27. The song was to become the last top 40 hit for Prince before his death in 2016.

     Many pundits have made a case comparing todays market situation to the year 1999. For anyone who can't remember, the dot-com craze of 1999 resulted in a lot of newly formed companies crashing to earth under the weight of heavy debt and a stock based on little more than dreams. There was clearly a bubble created in the stock market where any company with a dot com in their name was bid up to irrational valuations. Most of these companies had no earnings or sales and could not really be valued by investors. They were initially financed by wealthy investors and off loaded to a naïve investing public at ridiculous prices. People bought the stocks mainly because they feared missing out of the next big thing which was the internet.

     Well, the internet WAS the next big thing but people still lost their money in companies that had a half-baked business plan if there was one at all. So what is the difference this time with the advent of AI and what could go wrong? For one thing, up until recently, most of the companies that are at the forefront of AI are profitable and have been for a long time. Microsoft, Amazon, Meta, Tesla, Alphabet, Nividia, and Broadcom all make money but they are spending massive amounts to build-out data centers. What disturbs me is that recently, many of these "hyperscalers" have committed most of their cash flows on the AI build. But now they are willing to take on debt to continue to out invest each other in what could be a race to the bottom. I appreciate the potential for AI to transform our economy but worry that the spend may not justify the payoff. Even more concerning is the circular financing that is occurring in the industry. It seems that the center of this financial circle is a company called Open AI, founded by Sam Altman. They have great potential but no earnings. Not to be left out of the party, Oracle has bet the farm on AI and raised lots of debt. Any failure in this chain could bring a crash.

     Currently, most of the spend on AI lands on the income statement of Nividia. They have the best GPU and everybody wants the latest iteration. However, AMD is hot on their heels to sell an alternative for the value minded hyperscaler. Other members of the Mag seven are trying to develop their own GPU's for future use. At this time Nividia still has command of the GPU market and will for the foreseeable future. The term "bubble" has been used recently by financial pundits which has caused weakness in the AI stocks but my take is that there is a long way to go in this space. At a PE of 52, I am not adding to my NVDA position but not selling either. In my opinion, they will be the last one to fall.

     The major constraint in the AI build-out is access to reliable electric energy to power these data centers. The electric grid is simply too outdated to meet the new demands of AI. Electric generation also has lagged because of clean energy initiatives. Companies like Corning Glass has stepped-up to build low cost wafers that go into solar panels, Cummins has diesel generators to provide back-up power on site, electric generation investment may best be done by buying an ETF like the XLU which has a basket of different utilities.

     Even though the hyperscalers's stocks have run into a soft patch lately, I believe the party has a long way to go before any major corrections. It is my belief that new money invested in the AI space should be directed to the companies that will be suppliers to support the industry. Anyone who wants to benefit financially should take the time to research just what goes into building an AI data center and who will benefit from the effort. In the words of Prince: "I was dreamin' when I wrote this so sue me if I go too fast but life is just a party and parties weren't meant to last,"