Broadcom warned this week that constraints at its contract manufacturer Taiwan Semiconductor Manufacturing Co (TSMC) are beginning to choke the company’s supply chain, a sign of how surging AI-driven demand for advanced chips is creating knock-on effects across multiple technology suppliers.
"We are seeing that TSMC is hitting (production capacity) limits," Natarajan Ramachandran, director of product marketing in Broadcom’s Physical Layer Products division, said to reporters on Tuesday. He added that until a few years ago he would have considered TSMC’s capacity to be "infinite."
Ramachandran said TSMC plans to increase capacity through 2027, but that the expansions have already become "a bottleneck, or that has kind of choked the supply chain in 2026."
TSMC did not immediately reply to an emailed request for comment.
The Taiwanese foundry, the world’s main producer of advanced AI chips, told markets in January that capacity was tight as the buildout of AI infrastructure absorbed a substantial portion of its most advanced production lines. At that time, TSMC said it was working to narrow the gap between supply and demand.
Broadcom’s executive stressed that the constraints they are observing are not limited to silicon. "Even though there are multiple suppliers in the industry today ... there is definitely a supply constraint in the laser space," Ramachandran said, and he noted that printed circuit boards (PCBs) had emerged as an "unexpected" bottleneck.
According to Ramachandran, capacity limitations at both Taiwanese and Chinese PCB suppliers are contributing to longer lead times, though he did not identify specific vendors. In response, many customers are now negotiating long-term supply agreements to lock in capacity commitments for as long as three to four years.
The shift toward extended contracts was echoed by memory chipmaker Samsung Electronics, which said last week it is working with major customers to move to three-to-five year agreements. That change reflects both customers’ desire for greater supply security and suppliers’ interest in smoothing demand swings.
Separate from the company remarks, investment tools and services have highlighted Broadcom stock as a subject of algorithmic evaluation. One such service says it evaluates AVGO using more than 100 financial metrics to identify investment candidates, and notes prior model-generated winners including Super Micro Computer (+185%) and AppLovin (+157%).
Broadcom’s observations point to a wider industry dynamic: as hyperscale AI infrastructure expands, capacity at the most advanced fabs and in ancillary components can become the primary constraint for technology supply chains. Companies across networking, server, and storage equipment, as well as materials and component suppliers, are likely to feel the effects if these bottlenecks persist.