At the beginning of March, the market closely watched the release of results from the technology giant Broadcom (AVGO), whose market capitalization increased by approximately 5% after the announcement of its financial results. This growth was not only a reaction to past profits, but above all to the bold vision of CEO Hock Tan. He predicts that the company’s revenues generated purely from artificial intelligence will exceed $100 billion in 2027, which is also supported by estimates from JPMorgan analysts, who see potential at up to $120 billion. Broadcom is thus definitively transforming from a component supplier into a key architect of global AI infrastructure.

Strategic dominance in the Custom Silicon segment

The fundamental engine of the company’s growth is its unique position in cooperation with six of the world’s largest hyperscale customers. Broadcom currently operates with capacity approaching 10 gigawatts (GW) of AI computing power, while the estimated monetization of this capacity in 2027 represents $12 to $15 billion in revenue per gigawatt. The company is strategically focusing on the development of custom chips (custom silicon) and advanced networking technologies that are essential for the connectivity of massive AI servers. According to an analysis by Goldman Sachs, Broadcom is able to reduce AI inference costs for hyperscalers at a pace similar to the market leader Nvidia, making it an irreplaceable partner for the most demanding technological operations.


Performance of Broadcom’s stock price over the last 5 years

Supply chain security and technological trends

One of the biggest concerns for investors in recent months has been the shortage of specialized memory (HBM) and wafer manufacturing capacity. However, Hock Tan reassured the markets by stating that Broadcom has these key capacities contractually secured until 2028. This move significantly reduces the company’s risk profile and allows it to focus on competitive positioning. In the area of AI server technologies, a new trend is also beginning to emerge: a shift toward copper connectivity at the expense of optical solutions.

Profitability and competition

Despite initial concerns about pressure on margins in the supply of complex AI rack systems, the company’s leadership stated that Broadcom’s AI segment will maintain profitability similar to its other high-margin semiconductor divisions. Broadcom benefits from the reality that developers of large language models (LLMs) cannot accept “good enough” solutions but require top-tier performance. This places Broadcom in a direct, yet complementary position to Nvidia. The market is therefore beginning to see Broadcom not merely as an alternative, but as an essential building block of future gigawatt-scale AI factories that will shape the global economy by the end of this decade.