PALO ALTO, California, May 11, 2026, 07:04 PDT
Broadcom opened Monday with its market cap sitting near $2.11 trillion, vaulting the chipmaker past Tesla, Meta Platforms, and Walmart as investors piled into custom artificial-intelligence chips. Shares hovered near $433 in the early session.
Timing is crucial here: Broadcom has shifted from a big but relatively quiet chip and software supplier into a core AI infrastructure trade. Investors are wagering that other cloud giants will lean into custom chips—application-specific integrated circuits, or ASICs—focused on specialized tasks to cut costs for AI model training and inference.
This jump comes as the rally tightens its focus. Over the past six weeks ending Friday, the PHLX Semiconductor Index surged more than 60%. Broadcom wasn’t far behind, climbing 45%, boosted by increased data-center investment from Big Tech and renewed appetite for chip stocks.
Broadcom’s numbers did the talking. First-quarter revenue hit $19.31 billion, a 29% jump on the year. AI-related revenue soared 106% to $8.4 billion, with custom AI accelerators and networking in the spotlight. Looking ahead, the company is guiding for roughly $22 billion in second-quarter revenue, with $10.7 billion set to come from AI semiconductors.
Broadcom CEO Hock Tan says the company can see a “line of sight” to over $100 billion in AI chip revenues for 2027. Analyst Gil Luria at D.A. Davidson told Reuters after Broadcom’s March numbers that such long-term visibility signals “significant growth in demand.” Reuters
Goldman Sachs’ James Schneider stuck with his buy call on Broadcom, keeping the price target at $480, TheStreet reports. Schneider pointed out that agentic AI — tech able to plan as well as respond — is still in the early stages for enterprises. While plenty of companies are testing it, fewer than a quarter have rolled out agents at scale.
The big shift comes down to the customer list. In an April filing, Broadcom disclosed it locked in a long-term agreement with Google to supply and develop custom Tensor Processing Units—those are Google’s proprietary AI chips—under contract through 2031. That same filing noted Anthropic would, through Broadcom, get access to roughly 3.5 gigawatts of next-gen TPU-powered AI compute starting in 2027, as long as Anthropic keeps up its commercial performance.
Anthropic is calling the Google-Broadcom deal its “most significant compute commitment to date,” noting a jump in its annualized revenue run rate to over $30 billion—up from roughly $9 billion at the close of 2025. The company still splits its workloads across AWS Trainium, Google TPUs, and Nvidia GPUs. Anthropic
Meta’s partnership with Broadcom will now stretch through 2029, the companies disclosed. The original deal covers more than 1 gigawatt, with plans to ramp up to several gigawatts as Meta pushes its MTIA custom AI chips. Broadcom, for its part, is on the hook to provide not just the chips but also packaging and Ethernet networking tech.
Broadcom is now working the VMware software business into its AI narrative. On May 5, the company rolled out VMware Cloud Foundation 9.1, targeting production AI workloads. Broadcom claimed internal projections point to as much as a 40% drop in server costs and 39% lower storage total cost of ownership for select AI workloads.
Nvidia remains the main force in accelerators, while AMD has landed significant AI chip contracts as well. But Broadcom’s custom ASIC processors are starting to make waves, according to Reuters, with cloud providers eyeing these tailored chips as an alternative to Nvidia gear for data centers running specialized workloads.
The trade isn’t without its hazards. Goldman flagged several issues — a potential pullback in AI infrastructure outlays, Broadcom possibly losing share in its custom-compute business, softer demand outside of AI, and intensifying VMware competition — all downside risks, according to TheStreet. On top of that, Broadcom’s exposure to Anthropic’s TPU spending hinges on Anthropic’s ongoing commercial performance, the filing noted.
Broadcom heads into its fiscal second-quarter earnings on June 3, after the bell. Wall Street wants to see if the company’s AI revenue momentum is still on pace for the 2027 goal—and if there’s real order growth to justify a market cap above $2 trillion, not just another packed chip play.
