OpenAI set to start mass production of its own AI chips with Broadcom in 2026, FT reports

OpenAI is reportedly partnering with Broadcom to produce its first in-house AI chip for internal use by 2026, a strategic move aimed at reducing its reliance on Nvidia. This initiative mirrors efforts by other major tech firms to develop custom silicon amidst surging AI computational demands, and aligns with Broadcom's recent disclosure of over $10 billion in AI infrastructure orders from an unnamed client.

Analysis

OpenAI is reportedly partnering with Broadcom to develop its first in-house AI chip, slated for production in 2026. This strategic initiative is designed for internal use, aiming to reduce OpenAI's significant dependency on Nvidia's GPUs. The move aligns with a broader industry trend where major technology firms, including Google, Amazon, and Meta, are pursuing custom silicon to manage the surging computational demands and costs of AI workloads. This report is substantially corroborated by Broadcom CEO Hock Tan's recent statement about securing over $10 billion in AI infrastructure orders from a new, unnamed customer, strongly implying OpenAI. For Broadcom, this partnership signifies a major validation of its custom silicon strategy and a material revenue catalyst. Conversely, for Nvidia, it represents the materialization of a long-term competitive threat, as a key customer is actively working to diversify its chip supply, potentially impacting Nvidia's future market share.

Key Decisions for Investors

• The confirmed partnership and associated $10 billion order flow is a significant positive catalyst for Broadcom (AVGO), solidifying its position as a key player in the custom AI chip market and de-risking its forward revenue estimates.
• For investors in Nvidia (NVDA), this development serves as a tangible signal of increasing long-term competition; while near-term demand remains robust, the trend of major customers developing in-house silicon warrants close monitoring as a potential risk to sustained market dominance.
• Consider re-evaluating semiconductor portfolio allocations to gain or increase exposure to the custom ASIC (Application-Specific Integrated Circuit) trend, as hyperscalers and major AI firms increasingly seek tailored, cost-effective solutions beyond general-purpose GPUs.