Broadcom Vs. Nvidia: The AI Chip War You're Not Watching

Broadcom is emerging as a significant, albeit stealthy, rival to Nvidia in the AI infrastructure market, leveraging key custom ASIC deals and advanced networking solutions. Its Ironwood TPU collaboration with Google is expected to generate over $15 billion in lifetime revenue, while its Tomahawk Ultra networking chips offer superior connectivity for hyperscalers. Broadcom projects its AI business to grow 60% year-over-year in fiscal 2025, reaching $19-20 billion, and $33 billion by 2026, driven by a 170% surge in AI networking, positioning it as a formidable competitor in the AI hardware arms race.

Analysis

Broadcom (AVGO) is strategically positioning itself as a formidable competitor to Nvidia (NVDA) in the AI infrastructure market through a dual-pronged approach in custom ASICs and high-performance networking. The collaboration with Alphabet's Google on the Ironwood TPU is a cornerstone of this strategy, projected to generate $9 billion in near-term revenue and over $15 billion throughout its lifecycle, validating Broadcom's ability to secure large-scale custom silicon deals. This is complemented by aggressive growth forecasts, with analysts expecting its AI business to grow 60% year-over-year in fiscal 2025 to between $19-20 billion, and potentially reaching $33 billion by 2026. While Nvidia dominates AI training, Broadcom is carving out a critical niche in networking infrastructure. Its Tomahawk Ultra chip, capable of connecting four times more chips than Nvidia's NVLink, is central to the 'scale-up' architecture favored by hyperscalers. This networking segment demonstrated significant momentum, surging 170% year-over-year and contributing to the company's strong AI-related revenue of $4.4 billion in Q2 and an expected $5.1 billion in Q3. Although Nvidia remains the incumbent, the trend of hyperscalers diversifying their hardware supply chain towards custom ASICs and Ethernet fabrics directly benefits Broadcom's core competencies, establishing it as a key, if less publicized, player in the AI hardware arms race.

Key Decisions for Investors

• Investors should view Broadcom not just as a semiconductor company but as a primary beneficiary of the AI data center build-out, with its custom ASIC wins and networking dominance providing a distinct growth vector alongside Nvidia.
• Monitor the AI networking segment's revenue growth, which surged 170% YoY, as this is a key indicator of its competitive advantage and its ability to capture share in 'scale-up' architectures.
• Consider that while Nvidia is the leader in AI training, Broadcom offers exposure to the crucial diversification trend among hyperscalers who are actively seeking alternatives, positioning AVGO as a strategic long-term holding in the AI infrastructure theme.