Technical/Data Background → Semiconductor Equity Research: Built an AVGO Earnings Model and Would Appreciate ER Feedback

Hi everyone,

I’m currently a graduate student at the University of Florida studying Information Systems / Data Science. My undergraduate background is in Applied Computing with a minor in Economics, and most of my work experience so far has been on the technology/data/IT operations side rather than traditional finance.

I’ve been seriously exploring a move into Technology / Semiconductor Equity Research, so instead of only reading research reports, I’ve started building my own company-level models and writing investment research.

My first deeper project is Broadcom (AVGO).

I built a quarterly driver-based model that separates:

• AI semiconductor revenue
• Non-AI semiconductor revenue
• Infrastructure software
• Gross margin / mix effects
• Operating expenses
• Interest, tax and diluted shares
• Non-GAAP EPS

My current base case comes out to approximately:

Q4 FY26E
Revenue: ~$35.0B
Non-GAAP EPS: ~$3.87

FY27E
Revenue: ~$160.6B
Non-GAAP EPS: ~$17.57

For comparison, the public consensus snapshots I’ve found are closer to ~$173.5B revenue and ~$19.5 EPS for FY27, so my model is currently around 7–8% below consensus on revenue and ~10% below on EPS.

The main reason is that I’m being more conservative on the FY27 revenue ramp rather than assuming management’s current AI growth rate continues indefinitely.

I also built revenue / gross-margin sensitivity tables and tried to distinguish between management guidance, historical seasonality, bookings, customer deployment commentary, and my own assumptions.

One thing I’ve been trying to avoid is simply saying:

“AVGO historically traded at X P/E, so it should trade at X P/E.”

Instead I’m trying to understand valuation through earnings durability, required return, DCF / reverse DCF, and then use multiples as a cross-check.

I’ve also started a second model on Marvell so I’m not learning the process from only one company.

I’ve built the model and supporting research as a portfolio project, but since I’m a new user I can’t include external links yet. I’ve summarized the key assumptions and outputs below and would still really appreciate feedback on the methodology.

I’d especially appreciate feedback from anyone currently or formerly in ER on:

1. Revenue build: Is my segment-level approach detailed enough for an entry-level ER model, or am I still modeling at too high a level?
2. Margins: Am I thinking about semiconductor/software mix correctly, or would you model gross margin differently?
3. Consensus bridge: How would you normally identify exactly why your model differs from Street estimates when analysts’ detailed assumptions aren’t public?
4. Valuation: What valuation framework would you expect an ER associate candidate to be comfortable defending in an interview?
5. Model structure: What is obviously missing from this model compared with what a first-year ER associate would actually maintain?

I’m not looking for validation of the AVGO stock call — I’d rather know where the methodology is weak.

Any criticism is appreciated.