Key takeaways
Powered by lumidawealth.com
- AI revenue more than doubled to $8.4B, driven by custom AI accelerators and networking demand.
- Customer demand is broadening from hyperscalers to enterprises building AI agents, code-assist tools, and subscription platforms.
- Q1 revenue: $19.31B (+29% YoY); Adj. EPS: $2.05 (vs $2.03 est.).
- GAAP EPS: $1.50 (vs $1.14 prior year); profit: $7.35B (vs $5.50B).
- Board approved a $10B share buyback.
- Guide: $22B revenue for the current quarter (vs $20.5B Street).
- AI networking expected to rise to ~40% of total AI revenue as Broadcom gains share.
What happened?
Broadcom posted a strong quarter as AI customers expanded and spending shifted from training models to deploying “inference” products and agents. CEO Hock Tan said demand is coming from both hyperscalers and non-hyperscalers aiming to “productize” large language models into platforms (agents, code tools, and consumer subscriptions). Broadcom also said OpenAI is expected to deploy its first-generation XPUs in 2027, exceeding one gigawatt of compute capacity, making OpenAI the company’s sixth major hyperscaler customer.
Why it matters
This reinforces two market narratives:
- AI spend is diversifying beyond a small set of hyperscalers into enterprises building real “agent” products, which supports a longer demand runway.
- Networking is becoming a larger slice of AI dollars, and Broadcom is positioning to capture more of that attach market alongside accelerators.
What’s next?
- Watch whether Broadcom sustains the faster-than-expected shift to deployment/inference, which management says is pulling demand forward.
- Track the mix shift toward AI networking (targeting ~40% of AI revenue) and any incremental customer wins beyond the current hyperscaler set.
- Monitor follow-through on the $22B quarterly revenue guide and how much of upside is AI vs. non-AI segments.
