The global chip industry is reeling from rising inventory levels due to weak demand, but there is a glimmer of hope behind the cloud. Stocks of advanced chips, especially graphics chips, are depleting rapidly.
The shortage caused by the artificial intelligence revolution and subsequent capacity expansion by big data and server operators will benefit graphics chip maker Nvidia as well as high-capacity DRAM chip makers such as Samsung Electronics and SK Hynix.
According to industry data, NVIDIA’s graphics processing unit (GPU) inventory value was $4.61 billion at the end of March, down 10.6% from the end of December. It’s the first time the company has run out of inventory since the fourth quarter of 2019.
Almost all of the graphics chips used for AI are made by the Santa Clara, Calif.-based company.
But with the advent of ChatGPT, a generative AI chatbot that answers questions in a human-like way, the surge in demand has far outstripped supply.
GPUs and high-bandwidth memory (HBM) DRAM chips are particularly in short supply.
AI Boom Starts Race for Chips
The shortage of advanced chips, the lifeblood of new generative AI systems, has started a race to secure spare capacity and lock down computing power, according to industry insiders.
A recent Wall Street Journal report said the shortage of advanced chips is getting worse, and US startups that help build AI models such as chatbots are “like toilet paper during a pandemic.” ‘ said.
At the WSJ CEO Council Summit on May 23, Tesla and SpaceX CEO Elon Musk said, “GPUs are a lot harder to come by than drugs at this point.”
AI server shipments are expected to reach 1.18 million units by the end of the year, up 38.4% from last year, according to market tracking firm TrendForce. Demand for AI chips is projected to grow by 46%.
Demand for Nvidia’s products has sent the company’s stock skyrocketing in recent weeks, up 39% from last month to $401.11 on Tuesday. Nvidia has a market capitalization of over $1 trillion, making it the first chipmaker to reach that level.
The system chip shortage isn’t limited to GPU makers such as Nvidia.
Inventories at central processing unit (CPU) maker Intel and application processor (AP) chip maker Qualcomm also plummeted.
Intel’s CPU and Qualcomm’s AP chip inventory assets stood at $13 billion and $68.6 million, respectively, at the end of the first quarter, down 1.7% and 1.1%, respectively, from the end of 2022, for the first time since the fourth quarter of 2020. decreased. .
shortage continues
Many AI companies expect the shortage to last at least until next year.
Server makers and their customers have also said they’ll have to wait six months or more to get Nvidia’s latest graphics chips.
In addition to the increasingly tight supply of GPUs, CPUs and AP chips, the demand for high-capacity memory chips for servers is also rising, boding well for Samsung and SK Hynix, the world’s two largest memory chip makers. It has become.
SK, the world’s second largest memory maker, supplies Nvidia with high-performance HBM3 DRAM chips for its graphics processor unit H100.
With the rapid development of AI and big data, big tech companies are working on faster processing of rapidly increasing amounts of data. This trend is expected to increase the demand for HBM chips, which vertically interconnect multiple DRAM chips and significantly increase data processing speed compared to traditional DRAM products.
Samsung also supplies its advanced chips to Nvidia and expects to get more orders from the GPU maker.
During his visit to the US in May, Samsung Chairman Jay Y. Lee met with NVIDIA CEO Jensen Huang to discuss ways to expand business partnerships, especially in AP and GPU chips.
“NVIDIA’s flagship GPU is produced by TSMC under a foundry contract, but Samsung has emerged as a strong candidate to share work with TSMC for NVIDIA,” said an industry insider.
write destination Kim Ik Hwan, lovepen@hankyung.com
In-Soo Nam edited this article.
