Wall Street demands absolute perfection right now. If you want proof, look at what just happened to Broadcom. The company dropped its second-quarter earnings report, beat expectations on both profit and total revenue, and still watched its stock tumble over 13% in after-hours trading.
Investors aren't reacting to a failing business. They're throwing a tantrum because Broadcom didn't raise its long-term artificial intelligence projections.
When you're trading at a massive valuation premium, merely meeting expectations is a cardinal sin. Tech investors have grown drunk on Nvidia-style beats, where companies don't just clear the hurdle—they fly over it and rewrite the outlook for the next three years. Broadcom CEO Hock Tan played it safe, stuck to his guns, and didn't change his 2027 AI chip forecast. The market punished him for it.
Here is the real breakdown of what happened, what the numbers actually show, and why the panic over Broadcom stock plunges on weak software sales, unchanged AI chip forecast for the year is missing the bigger picture.
The Revenue Miss and the AI Speed Bump
Let's look at the headline numbers. Broadcom brought in $22.19 billion in revenue for the second fiscal quarter of 2026. While that represents a massive 48% jump year-over-year, it technically fell a hair short of the $22.27 billion consensus estimate from Wall Street.
When a stock has surged roughly 40% since the start of the year and added hundreds of billions in market value over just a few days leading up to the announcement, even a tiny miss feels like a car crash.
The real friction lies in the upcoming projections. Broadcom expects its third-quarter AI chip revenue to land at $16 billion. While that is a staggering 200% increase compared to the same period last year, Visible Alpha data shows analysts were quietly modeling $16.36 billion.
- Q2 Revenue: $22.19 billion actual vs. $22.27 billion expected
- Q2 Adjusted EPS: $2.44 actual vs. $2.39 expected
- Q3 Projected AI Revenue: $16.0 billion vs. $16.36 billion expected
- Long-Term 2027 AI Sales Forecast: Unchanged at $100 billion
The market wanted a blowout. Instead, it got a highly predictable, incredibly steady performance from a management team that refuses to over-promise.
Why the Unchanged 2027 Forecast Scared the Market
The biggest blow to investor sentiment wasn't the current quarter; it was the long-term outlook. Hock Tan reiterated Broadcom’s long-range forecast of $100 billion in AI chip sales by 2027. He did note that the company expects to ship slightly more advanced processors—moving past 10 gigawatts' worth of AI chip capacity—but the revenue target stayed exactly where it was.
As tech analyst Ben Bajarin from Creative Strategies pointed out, nothing is slowing down. The company didn't lower its guidance; they just didn't raise it. But in a hyper-growth market, an unchanged forecast feels like stagnation to momentum traders.
You have to look at how Broadcom builds its business to understand why this happens. Unlike Nvidia, which sells off-the-shelf graphics processing units to anyone with a checkbook, Broadcom focuses heavily on custom application-specific integrated circuits (ASICs) and AI networking hardware. They build custom accelerators for giant hyperscalers like Alphabet, Meta, and Anthropic.
These massive, multi-year custom projects take time to develop, test, and manufacture. The revenue visibility is clear, but it doesn't scale in a chaotic, exponential spike overnight. It moves in massive, steady blocks.
The Growing Competitive Threat
Another reason investors are suddenly getting skittish is that Broadcom isn't the only player in the custom chip space anymore. The boom in inference—where trained AI models actually process user queries in real-time—has made custom silicon a hot commodity. Everyone wants cheaper, more energy-efficient alternatives to Nvidia’s expensive hardware.
That demand is drawing serious competition. Marvell Technology recently turned heads by announcing its custom chip business would cross $10 billion in revenue by 2029, beating short-term Wall Street estimates. When rivals start making visible inroads with the same hyperscale cloud clients, investors start questioning if Broadcom can maintain its iron grip on the custom accelerator market.
Supply chain constraints have also been a lingering headache. Advanced chip packaging requires specialized components that are tough to source. While Broadcom management explicitly stated during the earnings call that they are very comfortable with their secured supply for both 2026 and 2027, the mere mention of supply chain friction makes a nervous market even more twitchy.
The Enterprise Software Reality Check
Beyond the hardware drama, Broadcom is managing a massive corporate transformation on the software side. After buying virtualization giant VMware, Hock Tan did what he always does: he slashed operating costs, overhauled the product line, and forced enterprise customers into subscription models.
That strategy is brilliant for creating predictable free cash flow, but it has a shelf life for top-line growth. The initial massive revenue surge from forcing legacy clients into new contracts has largely played out. Infrastructure software revenue grew just 1% in the first quarter of the year, showing a sharp deceleration from previous quarters.
The enterprise software side is doing its job by keeping margins incredibly high. Adjusted EBITDA margins are running at a clean 68%, and the company generated billions in free cash flow, allowing them to buy back $7.8 billion in stock recently. But software isn't providing the growth spark that investors are looking for right now. They want AI growth, and they want it to look infinite.
How to Handle This Selloff
If you're holding Broadcom or looking for an entry point, you shouldn't view this drop as a sign of fundamental rot. The business is fundamentally a money-printing machine. The core issue is a mismatch between unrealistic Wall Street expectations and the reality of complex hardware manufacturing timelines.
Don't panic sell based on a tiny revenue miss or a flat long-term forecast. Look at the massive infrastructure buildout happening globally. Cloud providers aren't stopping their capital expenditure. They need Broadcom's Ethernet switches and custom chips to make their data centers work.
The smart move here is to watch the valuation settle over the next few days. Broadcom was trading at an incredibly high multiple relative to its historical median. This pullback clears out the impatient momentum traders who were hoping for an easy post-earnings pop. Use the weakness to assess your portfolio allocation, look for a stabilized price floor, and remember that a steady $100 billion long-term forecast is still an insanely profitable target.
