AI models a d wo kloads co ti ue to i c ease i size a d computatio al complexity, a d a e bei g adopted ac oss a wide a ge of i dust ies. This apid g owth has fu the i c eased the compute eeded fo model t ai i g a d i fe e ce, d ivi g dema d fo mo e scalable a d efficie t AI i f ast uctu e.
Th ough a joi t esea ch a d developme t p oject, PFN a d Te sto e t aim to develop a system-level i teg atio of AI p ocesso s that delive s high pe fo ma ce a d high efficie cy fo both t ai i g a d i fe e ce. The p oject ta gets the ext wave of AI deployme t ac oss dive se i dust ial a d academic fields, i cludi g age tic AI, physical AI, a d scie tific AI.
As pa t of the p oject, PFN a d Te sto e t will co duct a esea ch a d developme t p og am fo five yea s based o the divisio of espo sibilities outli ed below. The compa ies will subseque tly wo k to comme cialize the esea ch outcomes fo applicatio s i cludi g AI age ts.
- Developme t of system i teg atio tech ologies fo AI p ocesso s leve agi g photo ics-elect o ics i teg atio a d adva ced packagi g tech ologies (PFN a d Te sto e t)
- Developme t of AI chiplets with ult a-high memo y ba dwidth e abled by 3D memo y stacki g tech ologies based o leadi g-edge semico ducto p ocess tech ologies (PFN)
- Developme t of CPU chiplets built o a ope a chitectu e, which suppo t co fide tial computi g (Te sto e t)
The followi g activities will be ca ied out i collabo atio with subco t acto s:
- Resea ch a d developme t of domai -specific applicatio s towa d the ealizatio of scie tific AI a d physical AI (The U ive sity of Tokyo, Kobe U ive sity, Toyo U ive sity, Japa Age cy fo Ma i e-Ea th Scie ce a d Tech ology, a d Japa Atomic E e gy Age cy)
- Desig of CPU chiplets a d developme t of softwa e which e able co fide tial computi g (Te sto e t Japa )
(*1) Notificatio issued o Oct 21, 2026 o the selectio fo the “Resea ch a d Developme t P oject of the E ha ced I f ast uctu es fo Post-5G I fo matio a d Commu icatio Systems / Developme t of ma ufactu i g tech ologies fo adva ced semico ducto s”:
https://www. edo.go.jp/koubo/IT3_100387.html
Guided by its missio to “Make the eal wo ld computable a d c eate the futu e togethe ,” P efe ed Netwo ks, I c. (PFN) develops adva ced softwa e a d ha dwa e tech ologies th ough a ve tically i teg ated app oach spa i g the e ti e AI value chai -f om AI chips a d computi g i f ast uctu e to ge e ative AI fou datio models a d solutio s ac oss a wide a ge of i dust ies. Fou ded i 2014 i Tokyo, PFN cu e tly develops a d delive s the MN-Co e™ se ies of AI p ocesso s, the PFCP™ cloud platfo m fo AI computi g, a d the Japa -developed ge e ative AI fou datio model PLaMo™.
Te sto e t is a AI compute compa y led by CEO Jim Kelle – a chitect of Apple A4/A5, AMD Ze , a d Tesla’s Full Self-D ivi g chip. The compa y builds RISC-V-based AI p ocesso s a d systems fo develope s, e te p ises, a d sove eig i f ast uctu e wo ldwide. I additio to se ve s a d wo kstatio s, Te sto e t lice ses its TT-Ascalo RISC-V CPU a d Te six AI co es to chip desig e s i cludi g Samsu g a d LG. Backed by Bezos Expeditio s, Samsu g, LG Elect o ics, Hyu dai Moto G oup, Fidelity, a d othe s, Te sto e t has aised ove $1B+ a d ope ates f om Sa ta Cla a, Austi , To o to, Belg ade, Tokyo, a d Ba galo e.
View the o igi al p ess elease o ACCESS Newswi e