Continue Reading: Unearth the Vital Insights from Broadcom Inc.’s Earnings Call!
Financial/Operational Metrics:
- Revenue: $14.916 billion, up 25% YoY.
- Net Income: $5.503 billion, up 315% YoY.
- Diluted EPS: $1.14, up 307% YoY.
- Operating Income: $9.8 billion, up 44% YoY.
- AI Revenue: $4.1 billion, up 77% YoY.
2Q25 Outlook:
ADVERTISEMENT
- Revenue: $14.9 billion, up 19% YoY.
- Adjusted EBITDA: 66% of projected revenue.
- AI Revenue: $4.4 billion.
- Infrastructure Software Revenue: $6.5 billion, up 23% YoY.
- Gross Margin: Expected 20 bps decline due to product mix.
Analyst Crossfire:
- New AI Customers & Custom Silicon Trend, Second-Half AI Revenue Outlook (Ben Reitzes – Melius, Harlan Sur – J.P. Morgan): Broadcom is working with four additional hyperscalers on custom AI accelerators, though they are not yet at scale. These partners are expected to eventually generate demand comparable to the existing three hyperscaler customers, but on a delayed timeline. While AI networking shipments exceeded expectations in Q1 and Q2, Broadcom is not providing guidance beyond Q2. The anticipated second-half ramp of 3nm AI accelerators remains on track, but no further details were given (Hock Tan – CEO).
- Tariffs & AI Disruptions, AI XPU Deployment & Design Wins (William Stein – Truist Securities, Ross Seymore – Deutsche Bank): The impact of potential tariffs is uncertain, but AI-driven semiconductor advancements are accelerating. AI is also shifting enterprise workloads back to private data centers, increasing demand for VMware Private AI Foundation. Broadcom defines a design win as a product deployed at scale in production, not just an initial contract. The company is highly selective in choosing customers, ensuring sustainable multi-year roadmaps (Hock Tan – CEO).
- AI Training vs. Inference Market, Networking & AI Cluster Scaling (Vivek Arya – Bank of America, Harsh Kumar – Piper Sandler): Broadcom’s AI revenue projection of $60B-$90B includes both training and inference workloads, though training remains the larger revenue driver. The company has separate product lines for each, ensuring coverage across AI demand shifts. Hyperscalers prioritize performance over everything when selecting networking solutions for AI clusters. Broadcom’s Tomahawk 5 and 6 switches, with increased RAID X capacity, provide a competitive advantage in large-scale AI deployments (Hock Tan – CEO).
ADVERTISEMENT
- XPU Market Expansion, Optimization & AI Differentiation (Timothy Arcuri – UBS, C.J. Muse – Cantor Fitzgerald): Broadcom’s forecasted 7M XPU units by 2027-2028 is based solely on its existing three customers. The four new engagements are not included yet, meaning they could further expand the total addressable market. Broadcom provides foundational semiconductor technology to hyperscalers, enabling them to optimize AI models. Its role is strictly hardware-related, ensuring peak performance while allowing customers to differentiate their software strategies (Hock Tan – CEO).
- R&D & AI Investments, AI Networking Mix & M&A Outlook (Christopher Rolland – Susquehanna, Vijay Rakesh – Mizuho): Broadcom’s $1.4B Q1 R&D spend, increasing in Q2, is primarily allocated to 2nm AI XPU development and next-gen Ethernet scaling for AI clusters. AI-related networking revenue surged in Q1 but is expected to normalize to 30% of AI sales (vs. 40% in Q1). Broadcom is focused on AI and VMware integration, with no immediate M&A plans (Kirsten Spears – CFO, Hock Tan – CEO).