The 25-Year Round Trip: Cisco’s Lessons About Client AI Concentration

John O’ConnellAdvisor Perspectives welcomes guest contributions. The views presented here do not necessarily represent those of Advisor Perspectives.

Sometimes, it takes time for good companies to live up to expectations. On December 10, 2025, Cisco Systems closed at a new all-time high. The last time Cisco closed higher was March 27, 2000.¹ That 25-year gap was not a rough patch. It happened to a company that never stopped being profitable, never stopped growing revenue, and never came close to failing.

Think about that distinction during your next client conversation about AI exposure. Cisco was not a bad company that deserved to lose money for its shareholders. It was a good company with a price tag that assumed flawless and indefinite execution indefinitely. Unfortunately, its price took a quarter century to align with its fundamentals. Your clients' growth-heavy portfolios are making a version of that same bet today, and most of them do not know it.

The Mechanism Nobody Names: Cap Weighting Concentrates Itself

Market-cap-weighted funds allocate more of your client's money to a stock as its price rises, automatically, with no rebalancing decision required. A stock that doubles does not just contribute more return. It commands a larger share of every future dollar that flows into the fund, which pulls in more of the next dollar, too.

This is not a flaw in fund construction. It is how cap weighting is designed to work, and for most of market history, it has been a reasonable way to track the economy's actual capital allocation. The problem is that the mechanism does not distinguish between a company earning a larger share of the index because it is genuinely capturing more of the economy, and a company earning a larger share because a narrative has pushed its multiple ahead of its fundamentals. The index cannot tell those two apart. It only sees price.

On March 27, 2000, Cisco’s stock peaked at about $80 a share, giving it a market value of $555 billion. For a short time, it was the most profitable company in the world. Its peak weight in the S&P was about 4%. But its success was short-lived. Two and a half years later, in October 2002, Cisco fell about 89% to $8.60, taking about 3% to 3.5% off the index, representing about 7% of its total decline.

Though Cisco has remained in the S&P the entire time, the stock continued its decline as big technology grew much larger in the 2020s, with its weight in the index falling to about 0.4% in the 2020s.

Cisco serves as a warning sign for today’s ongoing concentration risk, as AI-linked stocks now constitute roughly half of the S&P. AI leaders have appreciated so dramatically that growth investing and concentrated AI investing have become close to the same trade, whether or not the client — or the fund provider — describes it that way.