Broadcom has spent the past few years doing something Wall Street finds irresistible: producing numbers so large that perfectly rational investors begin behaving as though disappointment has been permanently removed from capitalism. Revenue is surging. Artificial intelligence semiconductor sales are accelerating. VMware is producing mountains of recurring software revenue. Free cash flow is pouring in. Management is forecasting growth rates that sound less like mature-company guidance and more like the early stages of discovering electricity. Naturally, the stock market has responded by attaching a heroic valuation to the business and quietly assuming that nothing inconvenient will ever happen again. That leaves me with a deceptively simple question: Can Broadcom continue outperforming the largest technology companies, or have investors already paid for several years of perfection in advance? I can build a powerful argument on either side. The bullish case says Broadcom occupies one of...
There are expensive stocks, and then there are stocks that make me stare at the valuation and wonder whether investors have begun pricing earnings from a quarter that has not yet been invented. Arista Networks belongs firmly in the second category. I understand why the market loves Arista. This is not some barely profitable technology company sprinkling “AI” across its investor presentation like parmesan cheese. Arista is growing rapidly, producing extraordinary margins, generating substantial cash, and selling critical networking equipment to some of the largest technology companies in the world. It is one of the highest-quality businesses associated with the artificial-intelligence infrastructure boom. The problem is that Wall Street knows this. At a closing price of $198.82 on August 14, 2026, Arista Networks—ticker symbol ANET—had a market capitalization of roughly $254 billion. Its trailing price-to-earnings ratio was approximately 62. That valuation does not merely expect Arista ...