The chip shortage situation has hampered the technology industry over the past few months, pushing up prices across the consumer gadget sector – all thanks to AI. Taiwan Semiconductor Manufacturing Co. (TSMC) Chief Executive C.C Wei has warned investors, stating that the global supply of advanced semiconductor chips will continue to fall behind due to the surging demand driven by AI for “several years.”
Speaking at TSMC’s annual shareholders’ meeting in Hsinchu, Wei stated that despite forecasting an impressive revenue growth of more than 30% for the year, the structural supply-demand mismatch shows no signs of immediate relief.
“It will be a long time before we can meet customer demand,” Wei told shareholders. He stated that the manufacturing bottleneck remains a definitive hurdle for the tech sector.
TSMC’s Arizona facility won’t suffice
To address these shortfalls and diversify its manufacturing footprint away from geopolitical tensions, while keeping the US administration appeased following threats of raised tariffs, TSMC has committed to a massive expansion project in Arizona. However, Wei kept the expectations low regarding the domestic impact of these new facilities.
Even with the additional Arizona capacity online, the CEO revealed that TSMC will still be unable to fully satisfy the volume of orders coming from its primary US customers, which include major hardware architects like Nvidia, AMD, and Broadcom.
Beyond advanced logic fabrication, the AI hardware sector is grappling with a severe deficit in memory hardware. Leading memory suppliers like SK Hynix, Samsung, and Micron, are facing unprecedented backlogs, with SK Hynix predicting lead times out as far as 2028. This sort of demand has triggered a financial uprise in the memory sector, pushing the market capitalisations of two of the three memory makers beyond the $1 trillion mark.
The ripple effects, however, are causing anxiety downstream. Industry leaders at infrastructure firms like Hewlett Packard Enterprise (HPE) and Arista have already sounded alarm bells over how these memory shortages will disrupt broader enterprise data centre buildouts.
Adding to the crisis is another geopolitical situation in the Middle East, where the ongoing disruptions in the Strait of Hormuz have affected supplies of critical chemicals required for advanced chip fabrication. TSMC had been monitoring the helium supply chain since March, but the worsening crisis has further strained hardware availability.
Not the first warning from a tech giant
Two days prior to Wei’s address, Nvidia CEO Jensen Huang delivered an almost identical message during the GTC Taipei media conference. Huang stressed that despite “full support” from global semiconductor partners, Nvidia’s specialised AI chips remain heavily supply-constrained, especially as the company races to deploy its next-generation architecture.
In late 2023 and early 2024, tech leaders like OpenAI CEO Sam Altman and Huang first made the markets sit up and take notice by warning that the world was facing a fundamental, long-term deficit in accelerated computing capacity.
At the time, Altman famously stated that solving the global AI chip and infrastructure shortage would ultimately require a multi-trillion-dollar global investment and several years of structural rebuilding.