KR — 전자신문

AIWorks Showcases Physical AI Verification Technology at China's GSF 2026 Science Forum

ChipsNovember 5, 20261 min read

AIWorks, a Korean AI testing company, showed a tool that automates testing of AI chips and robot control hardware at a science forum in China. The tool is called DeviceProbe.

What happened

AIWorks showed a testing and verification platform called DeviceProbe at the 2026 Greater Bay Area Science Forum (GSF) in Guangzhou, China, on the 1st. The company said the tool automates tests for AI chips and robot control devices. Its CEO-level presenter, Kim Hyung-bok, said the platform links AI chips, embedded boards and test equipment over a remote connection.

Why testing is slow today

AI hardware is being built faster than ever. But testing is still mostly done by hand on site. Development boards and test tools are spread out, so repeating tests and finding the cause of errors takes extra time and money.

Key facts

AIWorks has run chip tests for car electronics for 8 years.
The platform was built from remote management of hundreds of development boards.
It covers booking equipment, remote control, running tests, collecting logs, and writing reports.
AI can create test scenarios automatically and spot faults and warning signs.
AIWorks started in 2015 as TestWorks and changed its name in March last year.

Why it matters

Physical AI means AI that works inside machines such as robots and drones. These devices must be tested again each time software or firmware changes. DeviceProbe aims to make that testing faster, more repeatable, and more stable, which could help AI hardware reach the market sooner.

What to watch

AIWorks plans to expand its testing skills from chips and car electronics into robots, drones and smart manufacturing.
The company plans joint trials with Chinese and Korean physical AI firms.
Watch whether these joint trials lead to real customers.

Thursday: Millions for European chip designs, OpenAI with a GitHub alternative

The EU will invest 525 million euros to support a new generation of European chip design companies, aiming to build design capacity and reduce dependence on the US and Asia. Separately, OpenAI plans to take over Pierre Computer, whose open-source developer tools are partly built into its Codex coding agent platform, as it builds its own AI coding infrastructure in competition with GitHub, which Microsoft owns.

ChipsArs TechnicaNov 5

Nintendo's biggest ROM chip supplier is bowing out of the business

Macronix, Nintendo's main memory chip supplier for physical game cartridges, said in its third-quarter earnings report that its ROM chip production "has reached its end" and that it will gradually exit the game industry. The company said ROM output has fallen 39 percent over the past year and now makes up just 5 percent of its business, as AI-driven demand pushes wafer capacity toward more profitable NOR, NAND, and eMMC chips.

ChipsThe VergeNov 5

Qualcomm is betting big on data centers as smartphone sales sink

With smartphone sales down around 20 percent from 2025, Qualcomm is turning to AI data centers to boost its bottom line. On its Q4 2026 earnings call, the company forecast about $650 million in data center revenue for Q1 2027, rising to $5 billion by the end of 2027 and $15 billion by 2029. It expects data centers to eventually account for around 30 percent of its business.

Chipsheise onlineNov 5

Chip manufacturing: EU invests hundreds of millions in chip design platform

The European Union will invest €525 million to support a new generation of European chip design companies, as announced by Henna Virkkunen at the launch of the European Chip Design Platform. European firms currently generate less than 1% of global chip design revenue, and the platform will offer design tools, support services, IP building blocks, manufacturing access and grants of up to €16 million for advanced start-ups. The initiative is a key part of the EU Chips Act, which aims to strengthen Europe's semiconductor ecosystem and reduce dependence on the US and China.