Broadcom Vs. Nvidia: The AI Chip War You're Not Watching
In the AI hardware race, while Nvidia dominates the spotlight, Broadcom is emerging as a formidable competitor with its smart silicon and networking solutions. Collaborating with Google on the Ironwood TPU, Broadcom anticipates $9 billion in near-term revenue and $15 billion over its lifetime. Analysts project a 60% growth in Broadcom's AI business by fiscal 2025, reaching $19–20 billion. Additionally, Broadcom's Tomahawk Ultra networking solution enhances its position in hyperscaler datacenters, with AI-related revenue expected to rise significantly. Despite Nvidia's current lead, Broadcom's strategic moves position it as a stealth rival in AI infrastructure.
In the escalating AI hardware arms race, Nvidia Corp NVDA may hog the spotlight—but Broadcom Inc AVGO is quietly firing back with smart silicon and hyperscaler-grade networking gear, setting up a surprisingly fierce head-to-head battle in the datacenter compute arena.
- Track AVGO stock here.
Custom ASICs Take The Spotlight
Broadcom's collaboration with Alphabet Inc‘s GOOGL GOOG Google on the Ironwood TPU (TPUv6, 3 nm) positions the chipmaker firmly in the AI acceleration game. The project is expected to generate $9 billion in near-term revenue and over $15 billion over its lifetime, validating Broadcom's ability to compete with Nvidia's GPU dominance.
Meanwhile, analysts expect Broadcom's AI business to grow 60% year-over-year in fiscal 2025, climbing toward $19–20 billion in AI revenue, and launch $33 billion in AI business by 2026.
Read Also: China Doesn’t Need Nvidia Anymore: Alibaba’s New Chip Is A Challenge To Washington
Networking As The Hidden Weapon
While Nvidia's GPUs get the praise, Broadcom is carving out a niche with its AI-grade networking solutions. The launch of Tomahawk Ultra—capable of connecting four times more chips than Nvidia's NVLink—puts it at the center of the "scale-up" architecture powering hyperscaler datacenters.
This builds on strong second-quarter momentum: Broadcom posted $4.4 billion in AI-related revenue, and expects $5.1 billion in the third quarter, with AI networking surging 170% year-over-year.
The Nvidia Shadow
Nvidia remains the default choice for AI training workloads, buoyed by tight supply, superior ecosystem, and reopening China export channels for chips like H20. But hyperscalers are diversifying with custom ASICs and Ethernet fabrics—areas where Broadcom is gaining ground fast.
Broadcom may not grab headlines like Nvidia, but its deep hyperscaler relationships and custom AI silicon and networking are making it a stealth rival in the AI infrastructure race
- Nvidia’s AI Fortress Shows ‘First Crack’—The China Risk Just Got Real
Image: Shutterstock
