Broadcom (AVGO +0.39%) is worth about $1.7 trillion today. It earned every bit of that valuation through custom AI chips and the networking gear that ties massive data centers together.
I'm not here to argue that Broadcom's business is anything less than excellent. But five years is a long runway in semiconductors, long enough for smaller, faster-growing companies to close a gap that looks enormous today.
I think two of them manage it.
Image source: Getty Images.
AMD doesn't need a miracle, just more of what's already working
Advanced Micro Devices (AMD -2.03%) has just crossed the $1 trillion mark, which means catching Broadcom would require roughly 70% growth over five years. Framed that way, it stops sounding like a moonshot and starts sounding like a company simply continuing to compound at a reasonable pace for half a decade.
What makes me believe it actually happens is the deal AMD signed with OpenAI, covering 6 gigawatts of GPU capacity, with the first gigawatt of its new MI450 chips set to roll out in the second half of 2026. AMD's own finance chief has said the arrangement should generate tens of billions of dollars in revenue, according to Reuters reporting.
Nvidia (NVDA -0.52%) still runs away with the AI chip market, and nobody serious is arguing otherwise. But AMD was never trying to dethrone Nvidia outright. It just needed a real, growing second slice of a market expanding fast enough for both companies to come out ahead, and landing a customer the size of OpenAI tells me that slice is now locked in.
NASDAQ: AMD
Key Data Points
Palantir is the bolder bet, and I'm making it anyway
Palantir Technologies (PLTR +5.17%) is currently valued at around $455 billion, which means that passing Broadcom would require that it roughly quadruple from here. That's a genuine stretch, and I'd rather say so plainly than pretend it's a sure thing.
Still, look at what the company is actually putting up. Revenue grew 93% year over year last quarter, and the U.S. commercial side of the business, the part that proves Palantir isn't just riding government contracts, grew 149%. Management didn't just hold its full-year guidance steady; it raised it again, now calling for commercial revenue growth above 134%. The company also closed $2.1 billion in new U.S. commercial contract value in a single quarter, up 271% from a year earlier.
Growth rates like that rarely stay stapled to a $456 billion valuation for long. Either the growth cools off hard, or the market eventually rerates the stock to match it, and I'm betting on the latter. Palantir's software now lives inside government intelligence operations and, more and more, inside ordinary commercial supply chains and factory floors, which strikes me as a wider moat than any chipmaker gets to build, since it isn't something a competitor can simply out-manufacture. Palantir's disclosures show that its services help government and defense intelligence, while its Foundry and AIP focus on commercial customers, including manufacturers and enterprises.
NASDAQ: PLTR
Key Data Points
What could go wrong
So, what could go wrong? Plenty, if I'm being honest. AMD still has to execute close to flawlessly against a rival with a five-year head start and considerably deeper pockets, and if Nvidia's software lock-in holds as tight as it has, AMD's share gains could stall out well short of what I'm projecting here.
Palantir carries the bigger risk of the two by a wide margin. The stock trades at a valuation that already assumes years of uninterrupted hypergrowth, so one disappointing quarter or one slowdown in that commercial engine could cut the shares in half long before they ever get the chance to quadruple. I'm not waving that risk away. I just think the business underneath it is strong enough to outrun it over a five-year stretch.
