Continue reading this on our app for a better experience

Open in App
Floating Button

ChatGPT buzz fuels demand for chips and gear, Advantest says

Bloomberg
Bloomberg • 3 min read
ChatGPT buzz fuels demand for chips and gear, Advantest says
The AI surge is driving up orders from Nvidia and AMD, which provide the key AI-training semiconductors and rely on Advantest as their main testing tools supplier. Photo: Bloomberg
Font Resizer
Share to Whatsapp
Share to Facebook
Share to LinkedIn
Scroll to top
Follow us on Facebook and join our Telegram channel for the latest updates.

Advantest Corp. is seeing a spike in demand for its chip-testing devices, catalysed by the frenzy of interest in OpenAI’s ChatGPT and other novel uses of artificial intelligence.

A global race to develop powerful computing clusters and next-generation AI training systems is spurring chipmakers to buy many more of the Tokyo-based company’s testing tools, Advantest Co-Chief Strategy Officer Yasuo Mihashi said in an interview.

The executive’s outlook comes against a backdrop of muted demand for consumer electronics and a US-led campaign to restrict the trade of advanced semiconductors to China. Japan last month said it would expand export controls on its chip technology. Still, the AI surge is driving up orders from Nvidia Corp. and Advanced Micro Devices Inc., two companies that provide the key AI-training semiconductors and which rely on Advantest as their main testing tools supplier, according to data compiled by Bloomberg.

“We are the industry’s dominant player, so we benefit when something like ChatGPT expands uses of high-performance computing,” Mihashi told Bloomberg News in an interview.

Advantest shares have surged 36% this year, approaching all-time highs as the world’s imagination has been captured by the promise of ChatGPT and other offerings aimed at making internet services smarter and more human-like. The company expects operating profit in the fiscal year ended March to rise by 48% to ¥170 billion (US$1.3 billion).

Microsoft Corp. and Alphabet Inc.’s Google are among the first to integrate generative AI into their web tools, while in China a parallel race is already underway — led by Alibaba Group Holding Ltd., SenseTime Group Inc. and Baidu Inc. — to build a dominant AI platform in the world’s largest internet market.

See also: Younger consumers in Singapore more receptive towards AI agents

Data centres are essential for training and operating AI models that inform everything from autonomous driving to advanced chatbot assistants. That requires access to tens of thousands of graphics processing units, which are presently the optimal hardware to analyze data in real-time.

“Demand for data centre GPUs and thus testers will surely grow fast even without China,” Ichiyoshi Research Institute analyst Mitsuhiro Osawa said. “The hardware is necessary for next-generation social infrastructure such as autonomous driving and AI, the technologies that countries all over the world are investing in fiercely.”

The growing use of AI is pushing the limits of chip miniaturisation and advanced packaging, raising the likelihood of defects during the manufacturing process, Bloomberg Intelligence analyst Masahiro Wakasugi said. “This structurally boosts demand for chip testers, to help manufacturers guarantee performance and quality.”

See also: Nearly four in five Apac retailers will increase their tech investment in 2025

Testers are also a key part of any bid by China to build its own advanced chips — something that the US has been pressuring allies to prevent. Since Advantest’s gear is not tied to a specific fabrication process, it may steer clear of Japan’s export restrictions, though that remains uncertain. A company spokesperson said Advantest is “investigating whether our company is subject to export controls.”

Advantest competes with North Reading, Massachusetts-based rival Teradyne Inc. Market share between the two can shift dramatically when a customer switches suppliers — which is extremely rare, as replacing testing tools necessitates a rethink of other equipment or entire production lines, as well as retraining staff.

Global revenue for high-performance GPU chip testers will likely overtake that for smartphone chip testers in a few years, Toyo Securities analyst Hideki Yasuda said.

“Chips for servers will get bigger and more complex, requiring more time for testing,” he said. “There’s no magic way to slash testing time. The only solution for chipmakers is to buy more tools to test more chips at the same time.”

×
The Edge Singapore
Download The Edge Singapore App
Google playApple store play
Keep updated
Follow our social media
© 2024 The Edge Publishing Pte Ltd. All rights reserved.