Broadcom said supply chain constraints continue to weigh on the technology sector, with manufacturing capacity at TSMC under particular strain. Reuters reported that TSMC continues to face soaring demand for Artificial Intelligence chips, putting pressure on production as the Artificial Intelligence infrastructure boom continues to absorb advanced manufacturing capacity. TSMC, a major producer of advanced Artificial Intelligence chips, has said it is working to narrow the gap between supply and demand for customers including Nvidia and Apple.
Natarajan Ramachandran, director of product marketing in Broadcom’s Physical Layer Products division, said TSMC is hitting production capacity limits. He added: “They will be increasing the capacity to 2027, but that has become a bottleneck, or that has kind of choked the supply chain in 2026.” The strain is not limited to semiconductors. Ramachandran said shortages are also affecting adjacent supply chains, including PCBs used in optical transceivers and other circuit board components. He also said there is definitely a supply constraint in the laser space despite the presence of multiple suppliers.
Supply chains across the technology industry have faced repeated disruption as Artificial Intelligence has become more dominant worldwide. The growing use of agentic Artificial Intelligence systems has added to the pressure because those systems require greater processing power. It was previously discovered that such constraints had impacted the likes of Intel and AMD, with prices going up on Intel’s server products by 10% and delivery lead times being delayed for AMD. AMD said at the time that it remained confident in its ability to meet customer demand globally based on its strong supplier agreements and supply chain, including its TSMC partnership.
The continued shortages, along with increasing CPU constraints, could prolong difficulties for Artificial Intelligence companies and hardware manufacturers. Many customers are entering long-term agreements with suppliers to secure capacity commitments for several years and protect access to critical components. Broadcom’s comments also follow its newly announced OpenAI partnership to build 10 gigawatts of custom Artificial Intelligence accelerators. Under that arrangement, OpenAI will design the chips and systems, while Broadcom will help develop and deploy them in support of more advanced Artificial Intelligence models.
