Broadcom’s latest earnings report gave investors two very different stories to weigh: rapid growth in AI semiconductors and a softer showing in infrastructure software.
That split helps explain why the stock came under pressure even though the company posted strong year-over-year revenue growth and guided to another large revenue increase for the current quarter. For buyers comparing Broadcom with other AI infrastructure stocks, the question is no longer whether AI demand is helping the business. It is whether expectations have already moved faster than the reported numbers.
The Numbers Buyers Should Compare First
Broadcom reported fiscal second-quarter revenue of about $22.19 billion, up 48% from the same quarter a year earlier. Adjusted earnings per share came in at $2.44. The company also reported net income of $9.31 billion, or $1.91 per share, compared with $4.97 billion, or $1.03 per share, a year earlier.
The headline revenue number was close to Wall Street expectations, but not cleanly above every consensus figure cited around the report. That matters because Broadcom had already been priced like one of the major beneficiaries of AI infrastructure spending.
| Metric | Reported result | Buyer takeaway |
|---|---|---|
| Total revenue | About $22.19 billion | Strong annual growth, but close enough to expectations to leave little room for disappointment |
| Adjusted EPS | $2.44 | Modestly above the estimate cited in the source material |
| AI semiconductor revenue | $10.8 billion | The strongest part of the report and the main reason investors still focus on Broadcom |
| Infrastructure software revenue | $7.18 billion | A weaker point versus the estimate cited in the source material |
| Current-quarter revenue outlook | About $29.4 billion | Above the estimate cited in the source material, but not enough to offset concern about AI expectations |
AI Chips Remain the Main Bull Case
Broadcom’s AI semiconductor revenue more than doubled from the year-earlier period to $10.8 billion. Management attributed that growth to custom AI accelerators and networking components used to connect large AI systems.
For investors, this is the part of Broadcom that competes for attention with Nvidia, AMD, Marvell and other AI infrastructure names. Broadcom is not selling the same kind of general-purpose GPU story as Nvidia. Its appeal is tied more to custom silicon, networking, and deep relationships with large cloud and AI customers that want purpose-built chips.
Chip War by Chris Miller
For readers trying to understand why custom silicon and AI networking matter, this book gives useful context on how semiconductors became a strategic technology market.
As an Amazon Associate I earn from qualifying purchases.
The source material says CEO Hock Tan discussed six core custom chip customers, with names including Anthropic, Google, Meta and OpenAI. Because customer-specific order details can be difficult to verify independently, buyers should treat that list as management commentary rather than a complete map of future revenue.
Why the Stock Still Sold Off
The disappointment centered on expectations. Broadcom had already risen sharply before the report, and investors appeared to want a larger raise to the company’s longer-term AI chip outlook. Instead, management reiterated its view that AI semiconductor revenue can exceed $100 billion in fiscal 2027.
That is still a large target, but the absence of a higher number reduced the surprise factor. In high-expectation AI stocks, a strong quarter can still be punished if buyers were positioned for an even stronger signal.
The other issue was software. Broadcom’s infrastructure software segment, which includes assets from the VMware acquisition, generated $7.18 billion in revenue. That was below the estimate cited in the source material. The shortfall matters because software was supposed to provide durable, high-margin support alongside the faster-growing chip business.
How to Read Broadcom Against Alternatives
For a buyer comparing Broadcom with other AI names, the tradeoff is fairly clear. Broadcom offers exposure to custom AI chips, AI networking and enterprise software in one company. That mix can be attractive, but it also makes the story harder to value than a pure AI accelerator supplier.
- Choose Broadcom for: exposure to custom AI silicon, networking demand, and large infrastructure customers.
- Be cautious if: you need clean software acceleration or a bigger upward revision to long-term AI targets.
- Compare against: Nvidia for GPU leadership, AMD for accelerator competition, and Marvell for networking and custom silicon exposure.
- Watch next: AI semiconductor revenue, software growth after VMware integration, and whether management raises or merely repeats its fiscal 2027 AI target.
Financial Statement Analysis and Security Valuation
This text is useful for readers who want a deeper framework for reading financial statements, judging earnings quality, and connecting reported results to valuation.
As an Amazon Associate I earn from qualifying purchases.
Verdict: Strong AI Business, Less Room for Error
Broadcom’s report did not weaken the basic AI infrastructure case. The company is still growing quickly, and AI semiconductor revenue is now large enough to drive the broader investment debate.
The problem is valuation discipline. When a stock has already moved sharply on AI expectations, buyers need more than good numbers. They need evidence that future revenue is rising faster than the market has already priced in. This quarter gave Broadcom bulls plenty to point to, but it also gave cautious buyers a reason to wait for cleaner software momentum or a more meaningful AI forecast reset.


