During the internet boom, Cisco reached a stage where good quarters became a liability. Investors stopped asking whether the company performed well and started asking whether it performed well enough to justify the future they had already priced into the stock. Once expectations detach from reality, beating estimates becomes less important than beating imagination.
That was my first thought after reading the latest results from Broadcom Inc (NASDAQ: AVGO).
Broadcom reported second-quarter earnings with a revenue of $22.19 billion, up 48% year over year, while non-GAAP earnings per share climbed 54% to $2.44. The company also generated $15.19 billion in adjusted EBITDA and $10.26 billion in free cash flow during the quarter.
Then the stock fell as investors demanded perfection from the company.
Investor Expectations Got Ahead Of Them
Revenue is rarely the most interesting figure in an earnings report, especially when you’re dealing with a company generating more than $22 billion per quarter. What interested me was how many parts of the business accelerated simultaneously.
AI revenue reached $10.8 billion during the quarter, up from $8.4 billion in the previous quarter and more than double the level reported a year ago. Semiconductor revenue climbed from $12.5 billion to $15 billion quarter over quarter, while infrastructure software revenue increased from roughly $6.7 billion to $7.2 billion.
Management then guided for approximately $16 billion in AI revenue next quarter, which means Broadcom expects to generate more from AI alone in a single quarter than many semiconductor companies produce in total annual sales. That is an extraordinary one by any reasonable standard.
Yet the post-earnings conversation shifted almost immediately toward what Broadcom didn’t do. Some investors wanted a larger guidance raise, others wanted management to expand its long-term AI opportunity framework again, and others focused on VMware pricing dynamics and whether hyperscaler spending can sustain this trajectory indefinitely. Nobody was seriously debating whether Broadcom had a great quarter. They were debating whether it had a legendary one, and that is a fundamentally different standard to apply to any business operating in the real world.
One reason I think Wall Street is fighting the wrong battle is because the company’s actions tell a different story than the conversation surrounding the stock.
While investors spent earnings night arguing about VMware pricing, customer churn, and software licensing, Broadcom spent the last year embedding itself deeper into the infrastructure powering the AI buildout.
The company expanded its partnership with Meta Platforms (NASDAQ: META) to deploy networking technologies across next-generation AI systems. It expanded collaboration with Alphabet’s (NASDAQ: GOOGL) Google Cloud to strengthen cloud networking infrastructure. It shipped its 3.5D Face-to-Face Compute System-on-Chip, designed to support increasingly complex AI workloads. It also launched the industry’s first 400G-per-lane optical DSP platform built for next-generation AI networking…
…all within the same period the market was fixating on whether guidance was wide enough. Those announcements reveal where management believes the durable value in this business actually sits, and it is not in defending a mature software acquisition. It is becoming structurally indispensable to the largest infrastructure buildout of this generation, which is a different and more powerful position to occupy.
The Chart Suggests Profit Taking, Not A Broken Story
The price action surrounding this earnings report deserves careful reading rather than a reflexive interpretation, because the context changes what the volume and the selloff actually mean.
Broadcom entered earnings after a sustained rally that carried shares from roughly $290 in April to nearly $518 before the report – a move driven by institutional capital allocating aggressively into a theme it believes can keep compounding, not retail speculation chasing momentum.
The post-earnings decline arrived on volume exceeding 58 million shares, one of the heaviest trading sessions visible on the chart, which under normal circumstances would warrant serious concern about distribution.
But even after that decline, Broadcom remains well above its rising 50-day moving average and significantly above its 200-day moving average, with both continuing to point higher. And more tellingly, buyers stepped in around the same price levels where the stock previously consolidated during May, which tells you institutions were willing to actively defend positions despite the disappointment around guidance and the reset in near-term expectations. Momentum cooled.
The trend didn’t break. That distinction is the entire ballgame when you are trying to separate a temporary reset from a genuine deterioration in the thesis.
I’d Buy This Dip
I understand why traders sold Broadcom after earnings. I don’t agree with them.
When I look at this quarter, I see revenue growing 48%, earnings growing 54%, AI revenue reaching $10.8 billion, free cash flow exceeding $10 billion, and management guiding for roughly $16 billion in AI revenue next quarter. Not to mention the deepening relationships with Meta and Google while launching products designed for larger and more demanding AI clusters.
Most importantly, I see a stock being punished for failing to exceed expectations that had become detached from reality. When a company executing at this level gets sold because investors wanted something even better, I have historically seen opportunity on the other side of that reaction, not a warning to step away from it.
