Broadcom has made significant moves in the AI sector, outlining plans to deliver about 10 gigawatts of AI computing capacity to six key customers by 2027. This effort marks the company's ambitious roadmap in a rapidly evolving market. CEO Hock Tan's forecast suggests that AI chip revenue could skyrocket beyond $100 billion by 2027, a remarkable rise from an anticipated $20 billion in 2025.
The term gigawatt economy is becoming increasingly salient as Broadcom detailed their strategy during their Q1 FY2026 earnings call. The company reported impressive growth, with $8.4 billion in AI semiconductor revenue for the quarter, representing an astounding 106% year-over-year increase. Though not all major customers have been disclosed publicly, significant partnerships are emerging. Broadcom's collaboration with OpenAI, first announced in October 2025, focuses on the design and deployment of custom AI accelerators with a capacity of 10 gigawatts, anticipating its first shipment in 2027.
Another noteworthy customer, Anthropic, plans to roll out 1 gigawatt of TPUs in 2026 and ramp that up to 3 gigawatts the following year. With these partnerships, Broadcom is making substantial strides towards its 2027 goals, likely supported by additional major customers like Google and Meta.
Breaking down the revenue implications reveals interesting insights. Analysts project that each gigawatt of AI computing infrastructure may yield between $12 billion and $20 billion in revenue. With a target of 10 gigawatts, total potential revenue could range from $120 billion to $200 billion, suggesting that Broadcom’s forecast of $100 billion-plus may actually be conservative.
Despite the optimism, Broadcom faces challenges. Reports in May 2026 indicated possible financing hurdles related to the partnership with OpenAI, estimated to be around $18 billion. The construction of 10 gigawatts of custom AI capacity necessitates considerable engineering resources and financial investment. This situation has raised questions about sustainability and feasibility within the highly competitive AI landscape.
For investors, Broadcom positions itself distinctly compared to competitors like Nvidia. While Nvidia provides general-purpose GPUs available for diverse customers, Broadcom is creating tailor-made solutions for specific large-scale clients. This concentrated approach entails significant financial dependencies. Each of the six major customers holds weight in Broadcom's revenue goals, with OpenAI likely accounting for a substantial portion.
The financing challenge surrounding the OpenAI partnership warrants careful observation. As AI companies invest enormous sums into infrastructure, their continued access to capital markets and investor confidence remains crucial for sustaining this momentum.