Who is involved in the project?

LSTC, Rapidus, the University of Tokyo, Hokkaido University, Fujifilm, GE HealthCare and Belgium's imec, with support from NEDO.

Which chips will be used, and when?

IBM's 5nm chips will be used for verification through fiscal 2028, and 2nm chips from fiscal 2029, with Hokkaido University Hospital testing usefulness.

Has the maker of the 2nm chips been decided?

The manufacturing source is undecided, but Hokkaido University professor Kudo Yoshiaki said at a press conference that Rapidus is the leading candidate.

AI in HealthcareTop Companies' AI MovesAI Business & IndustryTop Companies AI8h agoOct 21, 2026, 06:30 JST

ラピダス-linked LSTC to build medical AI on 2nm chips with Hokkaido Univ, Fujifilm

Send on LINE

3 Key Points

What happened
LSTC, with Rapidus and the University of Tokyo, plus Hokkaido University and Fujifilm, said on the 20th it will develop medical AI systems on 2nm chips, targeting over 5x faster image processing and over 10x better energy efficiency than current GPUs.
Why it matters
The partners say today's medical AI is held back by high power consumption and processing speed, so the 2nm approach is meant to make AI practical in real clinical settings.
What to watch
The project runs through September 2031, with IBM's 5nm chips used for verification through fiscal 2028 and 2nm chips tested at Hokkaido University Hospital from fiscal 2029.

WHO IT HITSHospitals and medical imaging teams could eventually get faster, lower-power AI for CT, MRI, ultrasound and surgical microscopy, while chipmakers and equipment vendors gain a proving ground for 2nm inference silicon.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Get news like this every morning, free

Context & Analysis

The announcement brings together a chip research center tied to Rapidus, a chipmaker that has not yet mass-produced 2nm silicon, and medical device and imaging players including Fujifilm and GE HealthCare. That mix suggests the project is as much about proving advanced chip designs in a demanding real-world setting as about the medical AI itself. The plan is staged: IBM's 5nm chips carry verification through fiscal 2028, and 2nm chips take over from fiscal 2029, with Hokkaido University Hospital as the clinical test site through September 2031. Notably, the manufacturing source for the 2nm chips is undecided, though professor Kudo called Rapidus the leading candidate, leaving the supply chain as an open question. The stated payoff is over 5x faster image processing and over 10x better energy efficiency than current GPUs, which the partners frame as a response to the power and speed limits that have constrained medical AI.

FAQ

Who is involved in the project?: LSTC, Rapidus, the University of Tokyo, Hokkaido University, Fujifilm, GE HealthCare and Belgium's imec, with support from NEDO.
Which chips will be used, and when?: IBM's 5nm chips will be used for verification through fiscal 2028, and 2nm chips from fiscal 2029, with Hokkaido University Hospital testing usefulness.
Has the maker of the 2nm chips been decided?: The manufacturing source is undecided, but Hokkaido University professor Kudo Yoshiaki said at a press conference that Rapidus is the leading candidate.

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Get it free

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.