Broadcom CEO Hock Tan sees AI chip revenue 'significantly' above $100 billion next year

Broadcom CEO Hock Tan has made a bold projection regarding the future of artificial intelligence (AI) chip revenue, forecasting that it will reach "significantly in excess of $100 billion" in the coming year. This optimistic outlook comes on the heels of the chipmaker's recent fiscal first-quarter results, which surpassed expectations, and a robust forecast for the current period. Tan's comments during the company's earnings call highlight the accelerating demand for AI chips, particularly from large customers who are increasingly relying on Broadcom's expertise in designing custom silicon solutions.

Tan emphasized the company's preparedness to meet this demand, stating, "We have also secured the supply chain required to achieve this," referring to Broadcom's ambitious sales target for 2027. The surge in AI revenue is evident, as the company reported that its AI revenue more than doubled year-over-year to $8.4 billion in the first quarter. Overall, Broadcom's total sales increased by 29% to $19.3 billion, with expectations for AI semiconductor revenue to reach $10.2 billion in the current quarter.

Following Tan's remarks, Broadcom's stock saw a notable increase of more than 5% in after-hours trading. This surge reflects investor confidence in the company's ability to capitalize on the growing AI market. However, the chip industry, including Broadcom, has faced several challenges in recent months, such as a shortage of high bandwidth memory essential for custom accelerators and capacity constraints in advanced chip manufacturing and packaging. Broadcom plays a crucial role in assisting its customers in translating their chip designs into silicon, providing back-end support before these processors are manufactured at large fabrication facilities, such as those operated by Taiwan Semiconductor Manufacturing Company.

The Role of Custom Silicon in AI

Tan pointed out that the demand for custom AI deployment is entering its "next phase," which is expected to accelerate as Broadcom collaborates with six key customers in designing their chips. Among these customers are major tech players such as Google, Meta, Anthropic, and OpenAI, with Fujitsu and ByteDance likely rounding out the list. Google has been a pioneer in the in-house chip development space since 2015, having designed its tensor processing units in collaboration with Broadcom. Since 2018, Google has made these chips available to cloud customers, with significant clients now including Apple and Anthropic.

Broadcom anticipates even stronger demand from next-generation Google chips slated for release in 2027. Meanwhile, Meta is reportedly in discussions to utilize Google's TPUs, and Broadcom is actively assisting the social media giant in developing its own MTIA accelerator. Despite some skepticism regarding the future of Meta's custom silicon program, Tan reassured stakeholders that the "MTIA roadmap is alive and well."

During the earnings call's Q&A segment, Bernstein Research analyst Stacy Rasgon pressed Tan for specifics regarding the projected $100 billion in AI chip revenue. Tan outlined the capacity contributions from various companies, including 3 gigawatts from Anthropic, 3 gigawatts from Google, at least 2 gigawatts from Meta, and 1 gigawatt from OpenAI, among others. He acknowledged that the revenue generated per gigawatt can vary significantly, but he maintained that his estimates were "not far" off.

While Tan indicated that the anticipated revenue boost would stem primarily from AI chips, it is worth noting that Broadcom's product portfolio extends beyond just AI accelerators. Industry analyst Ben Bajarin from Creative Strategies pointed out that the company's offerings also include digital signal processors, data processing units, and networking switches, which contribute to its overall revenue.

Why it matters

The implications of Broadcom's projections extend beyond the company itself, signaling a significant shift in the semiconductor landscape driven by AI technology. As companies increasingly seek custom silicon solutions to enhance their AI capabilities, Broadcom's role as a facilitator in this transition positions it favorably in a rapidly evolving market. The collaboration with major tech firms like Google and Meta underscores the growing importance of tailored chip designs that cater to specific AI applications.

For creators and technologists, this trend indicates a burgeoning opportunity to leverage AI in innovative ways, as the availability of advanced chips becomes more widespread. The anticipated growth in AI chip revenue reflects not only the demand for more powerful processing capabilities but also the potential for new applications and services that can arise from enhanced AI technologies. As the industry continues to evolve, understanding the dynamics of AI chip development will be crucial for those looking to stay at the forefront of technological advancement.