Nvidia has maintained its dominance in the artificial intelligence (AI) chip market for the past three and a half years, largely due to the significant computational power offered by its graphics processing units (GPUs). These GPUs are particularly well-suited for training and deploying AI models and applications, allowing Nvidia to reportedly command an impressive 81% share of the data center chip market, as indicated by IDC. This leadership has contributed to Nvidia’s robust financial performance, making it the largest company in the world by market capitalization.
However, Broadcom is swiftly emerging as a formidable competitor in the AI chip arena, potentially reaching parity with Nvidia by the decade’s end. Unlike Nvidia, Broadcom specializes in application-specific integrated circuits (ASICs)—custom processors optimized for specific tasks rather than the general-purpose functionality of GPUs. This tailored design enables ASICs to operate with greater speed and energy efficiency, making them increasingly attractive to AI data centers.
Recent trends show that Broadcom is set to capture a significant share of the ASIC market, with Counterpoint Research estimating that the company could control 60% of this niche by next year. Last quarter, Broadcom reported a staggering 106% year-over-year increase in AI revenue, contributing $8.4 billion to its total revenue of $19.3 billion. Remarkably, AI revenue accounted for 43% of Broadcom’s total earnings, up from just 27% in the same period last year.
Investors can look forward to further acceleration in Broadcom’s AI revenue, projected to reach $10.7 billion this quarter alone. The company forecasts a path toward exceeding $100 billion in AI chip revenue by 2027, as CEO Hock Tan emphasized during a recent earnings call. This projection represents a fivefold increase over Broadcom’s AI revenue by fiscal 2025. With significant partnerships established with tech giants like Google, OpenAI, and Meta Platforms, Broadcom’s ambitions appear well-founded.
Looking ahead, ASICs are expected to constitute 19% of an anticipated $600 billion AI chip market by 2033. Meanwhile, Broadcom’s rapid growth underscores a swift shift from traditional GPUs to specialized AI processors, driven by cost-effectiveness and performance benefits. Major deployments of Broadcom’s custom AI processors by large cloud service providers and AI-focused companies are anticipated to further propel its market position.
For instance, Anthropic plans to roll out 1 gigawatt (GW) of these custom processors in 2026, scaling up to more than 3 GW the following year. OpenAI is also expected to purchase 1 GW of Broadcom’s chips next year. Comparatively, Nvidia has secured a significant 10 GW contract with OpenAI for its chip system.
Broadcom’s AI revenue is outpacing Nvidia’s growth rate. While Nvidia reported a 75% increase in data center revenue for its most recent quarter, bringing its annual revenue to about $62.3 billion, Broadcom’s robust growth trajectory suggests it could soon close the gap. With a target of $100 billion in AI revenue by 2027, Broadcom is poised to chart a quarterly revenue path nearing $25 billion.
The custom AI processor segment is projected to grow at an annual rate of 27% through 2033, according to Bloomberg, but Broadcom is expected to outperform that average. If Broadcom sustains 35% annual growth rates in its AI revenue over the next few years, it could accrue approximately $246 billion by the end of the decade.
In conclusion, Broadcom’s rapid advancements in the AI chip market position it as a potential rival to Nvidia, signaling promising returns for investors, given Broadcom’s market cap stands at $1.5 trillion—about one-third of Nvidia’s. As the tech landscape evolves, both companies’ strategies in the AI domain will be pivotal in defining their future fortunes.
