In April, Taiwan Semiconductor Manufacturing Company (TSMC) announced that quarterly revenue increased 17% in the first quarter of 2024, exceeding analysts’ expectations. The reported boom continues a trend in better-than-expected sales, signaling that the end of the global slump in the chip market may be in reach.
Though many anticipate chip inventory rebalancing and normalization come early 2025, the chip supply chain will continue to face challenges and constraints. Infrastructure costs, geopolitical events, increased vulnerability to natural disasters, and both natural resource and talent shortages will all contribute to ongoing challenges, even as semiconductor supply normalizes.
The widespread adoption of artificial intelligence (AI) and hyper-scale cloud computing is driving much of the surge in chip demand and has also led to shifting fabrication plant (fab) priorities, which will impact sourcing for certain components in 2025 and beyond.
Though some elements of market recovery are worth celebrating, buyers should remain cautious and proactively prepare for recovery while developing strategies that help minimize risk.
The semiconductor market should ease into more balanced, predictable supply behavior in early 2025. This will be due to the following factors:
In 2025 and beyond, the AI, cloud and embedded space is expected to have above-average growth driven by high underlying demand for edge computing, data centers, networking and infrastructure equipment, and the continued proliferation of IoT products. Top semiconductor manufacturers, including TSMC, have seen a consistent increase in demand for semiconductors due to the boom in demand for AI products specifically. The demand in this space will continue — and likely grow — for the next several years at a minimum.
Market recovery does not mean the end of semiconductor shortages. While semiconductor industry insiders now begin to see a normalization of the chip supply chain and anticipate a full normalization come early 2025, this new “normal” will evolve into rolling periods of constraint environments for certain node sizes.
Semiconductor supply constraints will be exacerbated by fab delays. New fab construction has been continually delayed in the U.S. Many of the fabs that started construction in 2021 and 2022 have been delayed due to a decrease in overall consumer demand, geopolitical events, and high development equipment costs. In the past year, TSMC reported multiple delays to its $40 billion facility in Arizona. The company announced last month that it would receive up to $6.6 billion in CHIPS Act funding to expand the site, but factories likely won’t be operational until 2028.
Given the market fluctuations and uncertainty, semiconductor buyers remain concerned about long-term supply. In order to protect against future disruptions, companies are turning to inventory prepayments to secure required components for their end products. For example, Micron announced it received $600 million in customer prepayments during fiscal Q1 2024 alone. In December 2023, AI leader NVIDIA made a total of $775 million in upfront payments to multiple chipmakers to secure cutting-edge high-bandwidth memories.
Of course, prepayments can place significant financial strain and pressure on organizations. Most companies will not be able to prepay for all necessary chips. To protect balance sheets, prepayments are best made on components that are in the highest demand.
After next year’s inventory rebalancing, wafer supply across most semiconductor manufacturing node sizes is expected to be outpaced by demand through 2027. Buyers should work to understand how supply for the chips they need for their end products will be affected by market shifts.
Next Generation, Sub-11nm: Next-generation chips (components smaller than 11nm) power high-growth advanced technologies such as AI and high-end cloud computing. New fabs will continue to prioritize next-generation chip investments as AI and markets grow. Sub-11 nodes remain the most profitable, which limits the number of companies willing to invest in the larger node sizes.
Mature Nodes, Greater Than 28nm: The automotive sector typically uses chips with mature nodes greater than 28nm. Supply may soon be trending in a positive direction for mature nodes. In the midst of a slowdown in consumer demand for electric vehicles (EVs), electric car companies have ramped up production of partial hybrid vehicles. Partial hybrids require fewer chips than pure EVs and some different types of components.
As auto manufacturers de-emphasize pure electric vehicles in favor of PHEVs, it will free up semiconductor supply for those remaining in the market. Specifically, there will be increased availability of chips with mature nodes greater than 28nm.
Legacy Components, >=65nm: As fabs move away from producing legacy components and larger node sizes, OEMs will find it increasingly challenging to source legacy parts through manufacturers and traditional distribution networks despite global CHIPS Acts. The most significant supply/demand imbalance will be in the >=65nm node size, where OMDs (original design manufacturers) are best positioned to fill gaps in supply.
After the inventory correction in 2025, supply and demand imbalances for legacy and current-gen components are expected to continue as there is limited new supply expected, despite a consistent demand outlook.
Market recovery may be on the horizon in early 2025, but it won’t be smooth sailing from there. To be best positioned to fulfill future demand, OEMs must build robust, resilient supply chain strategies now.
Here are three strategies that OEMs can employ to mitigate risks and prepare for potential constraints:
By keeping a finger on the pulse of market shifts, adapting strategies when necessary, and preparing for constraints proactively, buyers can take on challenges that may arise after the market normalizes.
