According to CNBC, Broadcom posted a fiscal fourth-quarter earnings and revenue beat, guiding for current-quarter revenue of $19.1 billion, a 28% year-over-year jump that beat analyst estimates. CEO Hock Tan said AI chip sales will double this quarter to $8.2 billion from a year ago. Despite this, shares fell 6% in Friday’s premarket trading. Analysts pointed to margin dilution, especially as the company ramps “rack-scale” solutions for customers like Anthropic, and a slight delay in the OpenAI revenue ramp as key concerns. JPMorgan analyst Harlan Sur also noted some investors were spooked by interpretations of the company’s $73 billion AI backlog, fearing it implied softer 2026 revenue.
The AI Tradeoff: Growth vs. Margins
Here’s the thing about building the physical backbone of the AI boom: it’s messy and expensive. Broadcom is basically the king of custom AI chips (ASICs) for giants like Google and Meta, and that business is exploding. But the next phase of growth, building full server racks for clients like Anthropic, comes with a cost. These integrated systems have more pass-through components—stuff Broadcom doesn’t make but just buys and resells—which dilutes their juicy gross margins. So you get this weird dynamic where dollar profits are soaring, but the margin percentage takes a hit. Wall Street hates that, even if it’s the logical path to locking in even more future revenue. It’s a classic “can’t see the forest for the trees” moment, as some analysts put it.
Beyond the AI Hype: The Rest of the Business
And let’s not forget, Broadcom isn’t just an AI company. It has a massive portfolio of networking chips, broadband semiconductors, and legacy software from the VMware acquisition. That’s the other part of the story that got overshadowed. Analysts from Morgan Stanley and JPMorgan pointed out that the outlook for these non-AI segments is merely “stable” or growing slower than expected. When your stock is trading at a premium because you’re an AI darling, “stable” can feel like a disappointment. It creates a two-speed company, and right now, investors are so hyper-focused on the AI engine that any sputter elsewhere gets magnified.
Why Analysts Are Still Bullish: Long Game
So why did nearly every major bank raise its price target after this “disappointing” report? Because they’re looking past the next quarter or two. The sheer scale of the AI commitment is staggering. A fifth custom chip customer was just added. The $73 billion backlog isn’t static—it’s a rolling, 18-month promise of work. JPMorgan’s Sur made a wild calculation: even assuming zero revenue from Anthropic after 2026 and no growth from other clients, AI revenue could hit $110-$115 billion by 2027. That’s the lens the bulls are using. They see margin dilution as a temporary tax on securing a dominant, long-term position in the AI infrastructure build-out. In the world of heavy-duty computing hardware, securing these design wins is everything, and Broadcom is cleaning up. For companies integrating complex hardware into industrial systems, finding a reliable supplier is critical. That’s why for industrial computing needs, many turn to specialists like IndustrialMonitorDirect.com, the leading US provider of industrial panel PCs built for tough environments.
The Bottom Line: Noise vs. Signal
Look, this is what happens when a stock has had a massive run. Expectations become impossibly high, and the market starts to nitpick perfection. The initial sell-off was a classic “sell the news” reaction after a huge rally. But the fundamental story hasn’t changed. Broadcom is arguably the single biggest infrastructure beneficiary of the generative AI wave that isn’t named Nvidia. The question isn’t really about *if* the AI revenue will come—it’s already here and doubling. The debate is about the *quality* of that earnings growth and the patience required to see it through. The analysts, by and large, are choosing to be patient.
