Not financial advice. This is how I’m thinking through the stock.
I’ve been digging into Broadcom—a heavyweight at the intersection of AI semiconductors and infrastructure software—to decide whether it’s a buy, hold, or sell going into late 2025. In short: the AI wave, VMware integration, and recurring software revenue make a compelling case, but valuation and culture signals keep me a bit conservative.
Why I’m Looking at Broadcom
Broadcom sits right where enterprise dollars are flowing: custom AI chips and infrastructure software. The company’s strategy leans into high-margin, recurring revenue from large, sticky customers—especially in AI and private cloud. That’s exactly the playbook I want in a maturing AI cycle.
What Just Happened (and Why It Matters)
In Broadcom’s fiscal Q3 (ended August 3, 2025), they reported record revenue near $16B, up ~22% YoY, and non-GAAP EPS that beat expectations. They also guided positively for the rest of FY2025. To me, it’s not just the beats that matter—it’s the combination of beats and confident guidance. That usually signals durable demand rather than a one-off quarter.
Two growth engines are obvious:
- Semiconductors: Robust demand for custom AI silicon.
- Infrastructure software: Continued lift from VMware integration and subscriptions.
On top of that, they’ve been active strategically—inking partnerships (e.g., Ubuntu integration with VMware Cloud Foundation) and even exploring sizable acquisitions. It’s an aggressive posture that fits their scale.
Stock Performance vs. the S&P 500
Returns have been monster:
- YTD: strong double-digits
- 1-Year: ~88%
- 5-Year: ~838%
Against the S&P 500—~14% YTD, ~16% 1-yr, ~101% 5-yr—Broadcom has outpaced over almost any period. That tells me the market already recognizes Broadcom as a secular AI winner.
Valuation & Quality Check
This is where I slow myself down and look under the hood:
- P/E: Elevated—hardly unique in megacap tech, but still worth respecting.
- EV/EBITDA ~51: Rich versus many semi peers—an indicator the stock isn’t “cheap.”
- Forward PEG ~1.86: For today’s environment, that reads as fair (not a screaming bargain, not egregious).
- Balance sheet: Assets < 2× liabilities. Not ideal, but not a deal-breaker by itself.
- Returns: ROA hovers around that 5–10% “solid” zone; Return on capital sits in an 8–15% sweet spot. That’s the kind of operational quality I want to see if I’m paying a premium multiple.
What the “Smart Money” and Analysts Are Doing
I never treat this as gospel, but I do check it:
- Hedge funds: Net buying shifted materially from Q1 to Q2 (76% buying vs. 24% selling in Q2).
- Insiders: More sellers than buyers lately—common after large runs and not always predictive.
- Analysts: Of 48 Wall Street analysts, ~92% rate AVGO a Buy, ~8% a Hold, 0% Sell. Also, Broadcom keeps beating quarterly expectations, which supports that bullish consensus.
Intangibles: Culture & Leadership
Employee sentiment isn’t Broadcom’s strength: ~3.3/5 overall, ~48% would recommend working there, and ~50% CEO approval. Culture scores don’t make or break a megacap, but they do color my conviction. Compared with peers like NVIDIA, Broadcom looks more “functional” than “beloved.”
My Thesis
I have Broadcom as a Buy—but a conservative one. Here’s why:
Bull case
- Structural tailwinds in AI and private cloud.
- Dual engine (custom AI chips + infrastructure software) with recurring revenue.
- Strong execution: beats + confident guidance.
- Long-term outperformance vs. the S&P 500.
Pushbacks / Risks
- Valuation is not cheap (EV/EBITDA especially rich).
- Culture/leadership scores aren’t stellar.
- If AI spending normalizes or custom silicon cycles, multiple compression could bite.
Positioning
If I’m building a semiconductor sleeve, I’d likely skew heavier to NVIDIA as the category’s gold standard, while giving Broadcom a meaningful allocation for diversification and software exposure. Think ~75% NVIDIA / 25% Broadcom as a simple illustration of how I might balance quality, culture, and AI optionality across two leaders.
Bottom Line
Broadcom is riding the right AI rails with execution to match. I’m calling it a conservative buy—a high-quality compounder with premium pricing. If I already own NVIDIA, I’d consider Broadcom as a complementary position rather than a replacement. If I own neither, I’d start with a measured entry and be ready to add on volatility.