|
Mohamad Ali would be the first to tell you he was the greatest! No disputing that here. Michael Jordan, Jack Nicklaus, Tom Brady and athletes of all sports have earned the “GOAT” title, The Greatest of All Time. The stock market has produced two GOATs in recent bull market history: first, Cisco of the 90’s dot-com boom and now Nvidia in the current AI powered bull market. I want to offer a comparison between the GOAT stocks. Hopefully this is helpful perspective. The saying that “stocks take the escalator up and the elevator down” is wisdom. The explosive rise in Cisco stock ended with a tremendous fall. From its peak of about $80 in March of 2000, CSCO fell steeply and continuously to $8 by October of 2002. After nearly 10 years of climbing, it dropped 90% in about 18 months. The stock did not return to its 2000 price for 26 years until February of this year. Will Nvidia encounter a similar fate? It is likely a matter of when, not if. RCA was the ultimate “tech darling” of the Roaring Twenties, rising 200-fold in value on excitement over radio technology. After peaking in September 1929, the stock lost roughly 98% of its value by mid-1932 during the Great Depression.
Ali was the “undisputed heavyweight champion of the world” from 1974 to 1978. Cisco stock was the dominant outperformer of the still unprecedented internet driven bull market of the 90’s. From early 1991 to its peak in 2000, Cisco Systems stock soared from a split adjusted price per share of about 0.50 to $66. The stock split 9 times between March of 1991 and March of 2000. The meteoric move in Cisco’s stock price over that time was 15,900 percent. Nvidia, boosted by the AI bull market, has gained 2,200 percent since a low of about $10 in October of 2022.
It is helpful to identify parallels in the stock market. Mark Twain is famously credited with the saying, “History doesn’t repeat itself, but it often rhymes.” Valuable perspective can be gained from the study of stock market “rhymes”. History dating back to 1890 provides insightful stock and stock market behavior throughout market cycles, economic booms and busts, and revolutionary technological advances. Many Wall Street analysts are calling the AI bull market a “bubble” and point to the burst of the dot-com bubble burst as a warning. As my first mentor often said, “Trees don’t grow to the sky.” No trend, asset price, or economic growth can increase forever without a limit. So, too, will this AI bull market end. Like the extended dot-com bull market, history indicates these types of advances exceed expectations and, as Sir John Templeton said, “… die on euphoria.” The FOMO effect (Fear Of Missing Out) leads to that euphoria and ultimate market cycle top.
The structural parallels between Nvidia (NVDA) post-ChatGPT and Cisco Systems (CSCO) during the late-1990s Internet boom represent two of the most striking “picks and shovels” market rallies in financial history. Both companies emerged as the indisputable infrastructure backbone for a generational technology shift — Cisco building the routers and switches for the web, and Nvidia supplying the GPUs and CUDA ecosystem for generative AI. However, while the narrative framing is nearly identical, the underlying fundamental mechanics tell a nuanced story. One strong parallel between the two is expressed best by the “picks and shovels” analogy. These GOATs mastered the selling “shovels in a Gold Rush.” In March 2000, Cisco briefly became the most valuable company in the world (reaching a ~$500 billion market cap) because you could not build the internet without its hardware. Similarly, Nvidia’s post-2022 run to multi-trillion-dollar valuations was driven by compute dominance, with tech giants aggressively vying for its accelerator chips. Today Nvidia’s market cap exceeds $5 trillion! Also, both rallies were propelled by massive enterprise and hyperscaler infrastructure spending. Service providers built out fiber networks at any cost in 1999; hyperscalers built out clusters of data centers starting in late 2022. Furtherer, both management teams commanded peak gross margins exceeding 70% as demand severely outstripped supply, giving each company immense pricing power during the buildout phase.
These companies have some key differences. Who Cisco did business with and how they did it was less sustainable than the way Nvidia does business. Cisco sold heavily to speculative, debt-laden telecom startups (such as Winstar and Global Crossing) and routinely used vendor financing—loaning money to its own customers to buy its routers. When those startups collapsed, Cisco faced massive inventory write-downs. Nvidia’s core buyers are the best-capitalized, cash-rich megacaps on earth (Microsoft, Alphabet, Meta, Amazon), who finance AI buildouts using operational cash flow rather than high-yield debt. Also, the nature of Cisco’s hardware business and Nvidia’s business ecosystem vastly differ. Cisco sold commoditizable networking hardware running proprietary IOS, which competitors like Juniper eventually chipped away at. Nvidia sells compute hardware tightly integrated with CUDA—a software and compiler ecosystem refined over nearly two decades. Switching off Nvidia isn’t just buying another chip; it requires rewriting an entire enterprise’s software architecture.
Here’s a side-by-side comparison of the Cisco of the 90’s and Nvidia today:
|