How Broadcom's big OpenAI deal fits into the data center boom and what it means for the AI trade

OpenAI announced a significant partnership with Broadcom for custom AI chips optimized for inference, causing Broadcom's shares to surge 10.7%. This 10GW deal, set for late next year deployment, adds to OpenAI's recent agreements, including 10GW with Nvidia and 6GW with AMD, totaling approximately 36GW of data center capacity announced this year, driven by an "avalanche" of AI demand. This massive buildout, including the $500 billion Stargate Project, underscores the early stage of the AI trade and highlights substantial growth opportunities for chipmakers like Broadcom and Nvidia, as well as power infrastructure providers such as GE Vernova and Eaton, due to the immense energy requirements.

Analysis

The recent partnership between OpenAI and Broadcom for custom AI inference chips has significantly boosted Broadcom's market position, with its shares surging 10.7% on Monday. This 10-gigawatt (GW) deployment, slated for late next year, integrates Broadcom's Ethernet stack and contributes to OpenAI's aggressive infrastructure expansion, which includes 10GW with Nvidia and 6GW with AMD, totaling approximately 36GW announced this year.

OpenAI co-founder Greg Brockman emphasizes an "avalanche" of demand, citing Sora's growth outpacing ChatGPT, necessitating this unprecedented industrial undertaking. The strategic move to custom-designed chips allows OpenAI to embed frontier model learnings directly into hardware, enhancing capabilities and reducing operational costs, a critical factor for scaling AI services.

This massive buildout, including the $500 billion Stargate Project, underscores the nascent stage of the AI trade and robust demand beyond current capacity. It creates substantial growth opportunities not only for chipmakers like Broadcom and Nvidia, which are considered reasonably valued, but also for critical power infrastructure providers such as GE Vernova and Eaton, given the immense energy requirements of these data centers.

Key Decisions for Investors

• Investors should consider evaluating long-term positions in core AI infrastructure providers like Broadcom and Nvidia, given their strategic partnerships and robust growth outlooks.
• Assess opportunities in power generation and management companies such as GE Vernova and Eaton, as the escalating energy demands of AI data centers create significant market tailwinds.
• Monitor OpenAI's ongoing infrastructure buildout and broader AI adoption trends as key indicators of sustained sector growth.
• Remain vigilant for potential overvaluation in peripheral AI plays, while recognizing the fundamental support for core infrastructure providers.