The AI boom has triggered an explosion in computing power demand, with GPU testing equipment manufacturers seizing new opportunities as Advantest's stock price soars over 37% this year.
The AI arms race among tech giants continues, as the surging demand for GPUs used to train AI not only fills the coffers of "gold rush" supplier NVIDIA but also offers opportunities for chip testing equipment manufacturers.
On April 18, Yasuo Mihashi, Co-Chief Strategy Officer of Japanese semiconductor testing equipment manufacturer Advantest, said in an interview that the opportunities brought by ChatGPT have led to a sharp increase in demand for chip testing equipment. Mihashi explained:
The global competition to train large language models is unfolding, which presents a huge opportunity for us. Despite the current downturn in consumer electronics demand, a surge in orders from NVIDIA and ADM has benefited us.
As industry leaders in chip testing, the continuous development of high-performance computing applications like ChatGPT will provide us with significant opportunities.
Hideki Yasuda, an analyst at Toyo Securities, believes that in the coming years, revenue from high-performance GPU chip testing equipment will surpass that of consumer electronics chip testing equipment:
GPU chips are becoming more complex, requiring more testing time, and there's no magic way to shorten the testing time. The only solution for chip manufacturers is to purchase more testing tools to test more chips simultaneously.
So far this year, Advantest's stock price has soared over 37%, nearing an all-time high. The company expects its operating profit to grow 48% to JPY 170 billion ($1.3 billion) by the end of the fiscal year on March 31, 2023.
AI chips become a battleground On Tuesday, April 18th, media reports revealed that Microsoft has another secret weapon in its arsenal: a custom-designed artificial intelligence chip that will provide powerful support for the large language models behind generative AI.
Microsoft hopes its custom AI chip, Athena, will outperform chips currently purchased from other suppliers, saving time and cost in its expensive AI business.
As the AI frenzy rages on and demand for GPUs used to train AI skyrockets, major cloud service providers, including Amazon, Microsoft, Google, and Oracle, have already begun to face computing power constraints.
Tracy Woo, a senior cloud analyst at Forrester Research, says the boom in artificial intelligence is putting increasing pressure on cloud providers, forcing them to develop their own chips:
Cloud computing companies can certainly buy chips from NVIDIA, but for giants like Google and Amazon, they have enough funds to develop their own chips.
Meanwhile, the growing number of AI applications is pushing the limits of chip miniaturization and advanced packaging, increasing the likelihood of problems during the chip manufacturing process, according to Bloomberg Intelligence analyst Masahiro Wakasugi:
This structurally increases the demand for chip testing equipment to help manufacturers ensure chip performance and quality.
