Broadcom's Earnings Face AI Fatigue: What Finance Pros Should Watch
Wall Street expects Broadcom to post strong numbers after the close on Wednesday. The problem: recent price action suggests even a clean beat might not break the slump.
Shares are off roughly 23% from December's record, badly trailing the S&P 500. At a $1.5 trillion market cap and deep ties to Alphabet and other AI leaders, Broadcom sits right in the crosshairs of the market's new fear-whether the AI spend can sustain itself.
The setup heading into the print
Consensus looks firm: adjusted EPS up 27% year over year to about $2.03, revenue up 29% to roughly $19.3 billion. AI-related sales are expected to nearly double to around $8.2 billion.
Traders bid the stock up as much as 2.8% ahead of results. Many also expect a constructive outlook, but there's growing skepticism that "beats" still translate into "buy."
Why a beat might not rally the stock
We just watched Nvidia top estimates and hike guidance-then drop 9.4% across two sessions. Broadcom itself fell over 11% after its December report when its $73 billion AI backlog over six quarters missed the whisper number.
Options are pricing a roughly 7% post-earnings move either way. Translation: positioning is tight, expectations are high, and the bar for upside is higher.
Three levers that could lift sentiment
- More hyperscaler wins with material revenue contribution.
- A clear, sizable step-up in the AI backlog over the same timeframe.
- Positive commentary around deals with OpenAI and Anthropic, and how they extend growth into 2027.
As one analyst put it, "It's like the better results they're giving, the worse the stock is doing." That's been this earnings season.
Execution edge vs. Nvidia matters-but may not be enough
"Broadcom will have great things to say," says Paul Meeks of Freedom Capital Markets. "But it may not matter." Sentiment is doing most of the work right now.
Janus Henderson's Shaon Baqui highlights Broadcom's custom silicon chops-seven generations of Google TPUs-and calls out "deep moats" in designing giant AI accelerators. The capability is real; the market still wants proof it scales into backlog and margins at the pace investors paid for in December.
Key numbers to underwrite in your model
- AI backlog: Direction vs. the prior $73 billion over six quarters. Magnitude and mix matter more than the headline.
- Google TPU cadence: Confirmation of second-half ramp and any generation-to-generation cost/performance gains. See Google Cloud TPU for context.
- OpenAI/Anthropic: Timelines, volume visibility, and any hints of multi-year commitments.
- Gross margin: Adjusted GM expected near 77% vs. 78% last quarter and 79% a year ago. Pin down AI mix impact, pricing, and manufacturing yield.
- Software segment: 42% of 2025 revenue-what's the trajectory and how does management defend it amid the software selloff?
- Capex at hyperscalers: Updated spend plans and lead-times that underpin orders.
Valuation lens
Broadcom trades near 27x forward earnings-down from 42x at the peak-but still above its five-year average of 22x and higher than Nvidia at about 21x. The stock is cheaper, just not cheap versus history and its closest comp.
Without a backlog step-up or clearer hyperscaler breadth, further multiple drift is on the table. With it, the stock can stabilize even if margins tick lower on AI mix.
Call checklist (use this to press management)
- What % of AI backlog is tied to new designs vs. follow-on orders? What's the conversion timeline by quarter?
- Unit economics for custom AI accelerators: where are gross margins today, and what's the path back to 78-79%?
- TPU generations: any evidence of cost curve improvements that offset pricing pressure?
- Software: ARR growth, churn, and cross-sell into the silicon customer base-can software reassert its role as a diversifier?
- Supply chain: constraints, lead times, and second-source strategies that de-risk 2H ramps.
Positioning ideas to consider
- Scenario framing:
- Base case-inline beat, neutral backlog update: chop within options-implied range.
- Upside-backlog surprise + hyperscaler logos: multiple support near current levels.
- Downside-soft margins + muted backlog: valuation gravitates toward the five-year average.
- Watch skew: with a ~7% implied move, dislocations can be sharp if guidance tilts either way.
For primary materials and historical guidance, monitor Broadcom investor relations. If you want to tighten your process on AI-driven modeling and valuation, explore AI for Finance for practical frameworks.
Your membership also unlocks:
