Inside the Nvidia Broadcom Palantir Investment Return Boom Wall Street Can’t Stop Talking About

In markets, five years is not a long time, but it is sufficient for a narrative to take shape. You’ll hear the same three tickers repeated like a chant if you walk around practically any trading floor or browse any Reddit investing thread. Palantir, Broadcom, and Nvidia. The names remain the same, but the combinations and rationales shift. Everybody knows someone who invested $10,000 in one of them in 2020 or 2021 and now discusses a down payment, a kitchen makeover, or a child’s tuition. It’s the kind of informal wealth talk that was once exclusive to housing bubbles.

The clarity of the math is almost offensive. Five years ago, a $10,000 investment in Nvidia was worth close to $126,000. The same sum was converted by Broadcom into almost $69,000. Palantir, which was by far the least successful of the three, increased that $10,000 to about $65,000. When all three are combined, a $30,000 wager is now worth over a quarter of a million dollars. Even at uncomfortable prices, mutual fund managers—the ones who are actually responsible for defending benchmarks—can’t seem to stop adding these names for a reason.

The cleanest story is that of Nvidia. Everyone involved in a data center build-out has been waving money at Jensen Huang since his chips became the picks and shovels of the AI gold rush. The company continues to print record numbers, including $62.3 billion in data center revenue alone last quarter, with a 75% gross margin that would shame most luxury brands, despite revenue growth slowing to an even more astounding 73% in its most recent quarter. It’s not even very costly at a forward P/E of about 22. The odd thing is that. The world’s most valuable company trades like a well-established industrial.

The more subdued and businesslike cousin is Broadcom. Although Hock Tan is not as well-known as Huang, Broadcom’s custom-silicon work with hyperscalers has grown into a sort of empire. AI revenue increased by 106% year over year to $8.4 billion in a single quarter, and management has been talking about $100 billion in revenue from AI chips by 2027 in a tone that seems to be genuine. Broadcom has an antiquated feel to it. The dividend, the consistent cash flow, and the strategic purchases. Until you see the chart, darling, you might almost forget it’s an AI.

The wild card is and most likely always will be Palantir. The stock trades at multiples that would make a value investor pass out, and Alex Karp conducts interviews that sound more like philosophy seminars than earnings calls. However, last quarter’s growth in U.S. commercial business was 137% year over year. Once thought to be the slow part, government revenue increased by 66%. The company’s $7.18 billion revenue forecast for 2026 indicates that it is now a category rather than a niche investment. The valuation cracks could still occur. It might also develop into it.

As this develops, it’s difficult to avoid thinking about Cisco in 1999 or Tesla in 2018—times when investors were advised to lower their expectations but failed to do so, sometimes with good reason. This trio seems to have already made their easy money. The returns from this will likely be lumpier, slower, and more reliant on whether or not AI capital expenditures truly compound as promised by the slide decks. Nevertheless, these three names consistently appear when a fund manager in Chicago rebalances a quarterly book or a Faisalabad investor pulls up a chart on a phone at 11 p.m. The trade in AI has developed. Somehow, the conviction hasn’t.