Broadcom Stock Crashes 15% After Earnings: What Just Happened to AI Chip Stocks

By Mike Chen · June 5, 2026

Broadcom stock plunged roughly 15% after its Q2 FY2026 earnings report on June 3 -- and it dragged the entire AI chip sector down with it. The numbers were objectively excellent: ~$22 billion in revenue (up 47% year-over-year), ~$10.7 billion in AI revenue (up 140%). But the market wanted a guidance raise, and Broadcom only reiterated. That was enough to trigger one of the most violent semiconductor selloffs of the year.

What Went Wrong When Everything Looked Right?

I've been following Broadcom closely for over two years now, and I want to be blunt: there was nothing wrong with this earnings report. The revenue beat. The AI segment grew at a rate most companies would kill for. The VMware integration is progressing ahead of schedule. By any normal standard, this was a fantastic quarter.

But we're not in a normal market. We're in a market where AI chip stocks have been priced for perfection -- and perfection means you don't just meet expectations, you blow past them and raise guidance while you're at it. Broadcom did the first part but not the second. Management reiterated its full-year outlook instead of revising it upward, and that single word -- "reiterated" -- cost shareholders roughly $90 billion in market cap in a matter of hours.

I watched the after-hours action in real time, and the speed of the decline was genuinely startling. It wasn't a slow bleed. It was an instant repricing -- the kind of move that tells you a lot of fast money was positioned for a guidance raise and hit the eject button the moment it didn't materialize.

Why Did the Selloff Spread to Nvidia, Micron, and Others?

This is the part that frustrated me the most. Broadcom's specific guidance decision had essentially nothing to do with Nvidia's data center GPU demand or Micron's HBM memory production. These are different companies serving different customers with different product architectures. But in a market running on narrative momentum, one crack in the AI chip story was enough to spook everyone.

Nvidia saw immediate selling pressure despite having no new information. Micron pulled back. The Philadelphia Semiconductor Index dropped broadly. It's the kind of correlated selloff that tells you the trade was crowded -- way too many portfolios had the same "long AI chips" bet on, and Broadcom's report gave them a reason to de-risk simultaneously.

Broadcom vs. Nvidia: Understanding the Real Difference

One thing most coverage gets wrong is treating Broadcom and Nvidia as interchangeable "AI chip stocks." They're fundamentally different businesses with very different risk profiles, and understanding this distinction matters if you're trying to figure out what the Broadcom selloff actually means.

Nvidia sells general-purpose GPUs. Their chips work for any customer, any workload. That's their moat -- they built a universal platform (CUDA) and an ecosystem so deep that switching costs are enormous. Broadcom, on the other hand, designs custom AI accelerators -- XPUs built specifically for individual hyperscalers like Google's TPU program and Meta's custom silicon efforts. This gives Broadcom incredibly deep client relationships but also introduces concentration risk. Lose one hyperscaler contract and a meaningful chunk of revenue disappears.

Then there's the networking angle. Roughly 40% of Broadcom's AI revenue comes from networking chips -- the silicon that connects thousands of GPUs inside data centers. This is actually the part of the business I find most compelling long-term. As AI clusters scale from thousands to hundreds of thousands of GPUs, networking becomes the critical bottleneck, and Broadcom has a dominant position there. If you're interested in how other chipmakers have navigated this AI spending wave, the Marvell Technology surge story is worth reading for context on how custom silicon plays have performed.

Is This a Buying Opportunity or a Warning?

Here's my honest take after processing the earnings call, the price action, and the broader sector dynamics: the fundamentals haven't changed. Let me say that again because it matters -- nothing about Broadcom's business deteriorated. Revenue grew 47%. AI revenue grew 140%. Those are exceptional numbers by any standard in any market cycle.

What changed was sentiment. The market had priced in a guidance raise and didn't get one. That's a positioning problem, not a fundamental problem. And in my experience, post-earnings dips driven by elevated expectations rather than deteriorating business quality have historically been some of the best entry points in high-growth semiconductor stocks.

That said, I'm not rushing to buy tomorrow morning. Sentiment damage takes time to heal, and there's a real risk that the "peak AI spending" narrative gains traction in the short term. I want to see stabilization in the stock price and some confirmation from the next round of hyperscaler commentary before adding to my position. The broader chip rally that recently pushed the Dow Jones to record highs showed how quickly sentiment can shift in both directions.

What This Means for the AI Chip Boom Going Forward

If I zoom out from the noise, Broadcom's report actually tells us something important about where we are in the AI infrastructure cycle. We're past the "everything goes up on any AI headline" phase. We're entering the "show me the guidance raise" phase. That's a more mature, more demanding market -- and it's going to separate the companies with real competitive advantages from the ones riding the tide.

The AI buildout isn't slowing down. Hyperscalers are still committing hundreds of billions to data center construction. The demand for custom silicon, networking infrastructure, and high-bandwidth memory is still accelerating. But Wall Street's expectations have caught up to reality, which means the easy money in AI chip stocks is behind us. From here, stock picking matters more than sector allocation.

My bottom line: Broadcom's 15% crash is the market recalibrating expectations, not the AI chip boom ending. The companies doing the actual building -- the hyperscalers ordering these chips -- haven't paused a single purchase order. If anything, this selloff is the market's way of reminding us that even the best growth stories need periodic resets to stay healthy. I'd rather see this correction now than watch valuations stretch to truly unsustainable levels.