AI Brings Opportunity and Risk to Semiconductor Industry. Is Your OEM Ready?

By Logan Wamsley

Calling Artificial intelligence (AI) the trend of 2023 is a monumental understatement. As the technology continues to evolve at an astonishing rate and embed itself into nearly every facet of daily life, we are still trying to understand the full implications it will have from the business side. Of these implications, among the most fascinating — and risky for OEMs not willing to pivot their inventory acquisition strategies — is the semiconductor market.

According to Goldman Sacs, the potential of Generative AI is poised to increase global GDP by as much as 7% or $7 trillion, automating 300 million full-time jobs while increasing productivity growth by 1.5 percentage points. To reach this potential, however, a vast amount of processing power will be needed across multiple interlinked networks. To meet those standards, semiconductor supply chains will quickly be pushed to the brink by a demand that has never been seen in the modern era.

Some companies have already benefited from this surge. Nvidia, for example, has recently added over $245 billion to its stock market capitalization — an increase driven almost entirely by its market leader status in the AI GPU category. According to recent estimates, an earlier version of ChatGPT required approximately 10,000 graphic chips, each of which retail at about $33,000. According to Tesla CEO Elon Musk, a more advanced version could require as much as five times those figures.

Nvidia, as well as many other leading OEMs, outsource their IC products to Taiwan Semiconductor (TSMC). Taiwan, especially in 2023, has been dealing with several issues affecting production outputs including talent shortages, geopolitical tensions with mainland China, and raw material shortages. In recent months, market inventory demands have somewhat waned, but if AI re-ignites a new semiconductor “gold rush,” it could have a catastrophic effect on lead times.

There is good news, however. In the short term, the AI boom has not yet had a major effect. In fact, South Korea (home to some of the world’s top memory chipmakers such as Samsung and SK) have been dealing with the highest-ever stockpile of chips in 26 years. According to an anonymous industry official interview in The Korean Herald, “It is when the AI industry is more settled that we expect more demand will be created for highly advanced data centers, and also for GPUs and CPUs, which are the brains enabling the operation of the AI.”

When that demand manifests, however, the results will be fascinating to see, and filled with risk for OEMs. That risk can be at least minimized by incorporating inventory ownership solutions or proactive purchasing strategies to have critical electronic components and semiconductors in hand ready for use before supply chain disruption occurs.

The time is now to claim inventory to maintain production schedules, and the window of opportunity will not last long.