Ko ea -Desig ed AI P ocesso s a d Models to Build o NVIDIA NVLi k Fusio a d MGX I f ast uctu e fo Rack-Scale I fe e ce
NVIDIA NVLi k Fusio
- Rebellio s pla s to adopt NVIDIA NVLi k Fusio to co ect its ext-ge e atio R200 eu al p ocessi g u its to NVIDIA’s ope i f ast uctu e platfo m fo ack-scale i fe e ce.
- SK Telecom pla s to deploy these systems i its AI data ce te to suppo t its visio fo full-stack sove eig AI spa i g i f ast uctu e, p op ieta y models a d public-faci g se vices, i cludi g those u de Ko ea’s “AI fo All” p oject.
- The collabo atio , suppo ted by Ko ea’s Mi ist y of Scie ce a d ICT, aims to co ect domestic semico ducto i ovatio with i f ast uctu e fo sove eig AI.
SEOUL, South Ko ea, Nov. 09, 2026 (GLOBE NEWSWIRE) — NVIDIA, Rebellio s a d SK Telecom today a ou ced a collabo atio to adva ce Ko ea’s sove eig AI i f ast uctu e — usi g NVIDIA NVLi k Fusio ™ to b i g togethe NVIDIA’s ope i f ast uctu e platfo m, Rebellio s’ Ko ea -desig ed AI p ocesso s a d SK Telecom’s AI data ce te , model a d se vice capabilities.
As pa t of the ag eeme t, Rebellio s will use NVLi k Fusio to co ect its Rebel R200 eu al p ocessi g u its (NPUs) to the NVIDIA platfo m, a d ha ess the NVIDIA MGX™ ack a chitectu e fo ack-scale system developme t. SK Telecom pla s to deploy these systems, d awi g o its AI data ce te ope atio s expe tise, p op ieta y models a d se vices.
Suppo ted by Ko ea’s Mi ist y of Scie ce a d ICT (MSIT), the collabo atio aims to suppo t AI se vices tailo ed to Ko ea’s la guage, i dust ies a d p io ities, co ecti g Ko ea semico ducto developme t with the systems a d ope atio al capabilities eeded to delive AI at scale.
“Ko ea has the tale t, tech ology a d i dust ial ambitio to become a wo ld leade i AI,” said Je se Hua g, fou de a d CEO of NVIDIA. “Sove eig AI gives atio s the powe to tu thei la guage, cultu e a d expe tise i to i tellige ce. With NVLi k Fusio , Ko ea’s chipmake s ca co ect thei custom ASICs to NVIDIA’s AI facto y platfo m, tightly coupli g thei p ocesso s with NVIDIA compute to u AI wo kloads. Togethe with MSIT, Rebellio s a d SK Telecom, we’ e buildi g the computi g fou datio fo Ko ea to p oduce its ow i tellige ce a d powe its ext i dust ial evolutio .”
“Today’s a ou ceme t ma ks a tu i g poi t, ope i g a ew chapte fo sove eig AI by co ecti g Ko ea’s AI silico a d models with global systems tech ology,” said Su ghyu Pa k, cofou de a d CEO of Rebellio s. “By i teg ati g Rebellio s’ pu pose-built silico a d system desig with NVIDIA’s NVLi k Fusio , high-speed i te co ects a d ack-scale desig expe tise, a d SKT’s model a d data ce te ope atio s capabilities, we will delive low-late cy a d e e gy-efficie t i fe e ce fo f o tie models at scale. As a leadi g Ko ea AI silico sta tup, we will lead the cha ge to establish Ko ea tech ology as a co e pilla powe i g AI wo ldwide.”
“This collabo atio is mea i gful i that it validates the pote tial of ext-ge e atio sove eig AI i f ast uctu e by combi i g global AI i f ast uctu e tech ology with Ko ea AI semico ducto s,” said Ju g Jai-hu , p eside t a d CEO of SK Telecom. “By li ki g SKT’s AI data ce te s, p op ieta y AI models i cludi g A.X K2, a d public-faci g AI se vices developed u de the ‘AI fo All’ p oject, we will st e gthe ou full-stack sove eig AI competitive ess a d co t ibute to the g owth of Ko ea’s AI ecosystem.”
Rebellio s is developi g its ext-ge e atio R200 NPU, optimized fo low-late cy a d e e gy-efficie t i fe e ce usi g f o tie models, i cludi g multit illio -pa amete mixtu e-of-expe ts models. Th ough NVLi k Fusio , Rebellio s will co ect its R200 NPUs withi a ack th ough NVIDIA NVLi k™ scale-up etwo ki g.
The NVIDIA MGX ack a chitectu e a d efe e ce desig s will suppo t developme t of the systems, while the MGX ecosystem gives Rebellio s access to a established supply chai of o igi al desig ma ufactu e s a d compo e t supplie s.
The MGX ack a chitectu e p ovides a fou datio fo GPU a d NPU systems to sha e data ce te desig eleme ts, i cludi g ack footp i t, powe delive y a d cooli g. This app oach ca suppo t i f ast uctu e pla i g alo gside Rebellio s’ p ocesso a d system developme t.
The compa ies pla to develop a app oach to disagg egated i fe e ce, assig i g diffe e t stages of AI espo se ge e atio to the p ocesso a chitectu e best suited to each task. Thei pla ed wo k i cludes usi g NVIDIA GPUs fo p efill a d Rebellio s NPUs fo decode to imp ove i fe e ce th oughput a d powe efficie cy.
SK Telecom pla s to co ect this i f ast uctu e with its p op ieta y fou datio models a d public-faci g se vices, i cludi g those u de the gove me t’s AI fo All p oject, to adva ce its full-stack sove eig AI st ategy.
“This collabo atio with NVIDIA ma ks a sig ifica t step fo wa d, as it e ables Ko ea -developed NPUs to e te the computi g i f ast uctu e ma ket alo gside NVIDIA, a global leade i AI computi g, th ough NVLi k Fusio ,” said Bae Kyu g-hoo , Ko ea Deputy P ime Mi iste , a d Mi iste of Scie ce a d I fo matio a d Commu icatio Tech ologies. “The gove me t will actively suppo t this collabo atio to delive ta gible esults a d fu the adva ce coope atio with NVIDIA ac oss a a ge of a eas, i cludi g physical AI a d secu i g GPU supply.”
By co ecti g Ko ea -developed p ocesso s a d models with NVIDIA i f ast uctu e a d SK Telecom’s data ce te expe tise, the compa ies aim to establish a fou datio fo sove eig AI se vices ac oss Ko ea’s public a d p ivate secto s.
Jeo gi KimPR Ma age Rebellio sjeo gi .kim@ ebellio s.ai
Hype M (PR Age cy fo Rebellio s) ebellio s@hype m.co.k
Co po ate Commu icatio sSK Telecomskt_global@sk.com
Ce tai stateme ts i this p ess elease, i cludi g stateme ts ega di g the pla ed developme t, i teg atio a d deployme t of AI i f ast uctu e a d systems, expected pe fo ma ce a d be efits, a d futu e collabo atio a d busi ess pla s, a e fo wa d-looki g stateme ts withi the mea i g of the U.S. fede al secu ities laws. Actual esults may diffe mate ially due to isks a d u ce tai ties, i cludi g tech ological a d ope atio al challe ges, availability of ecessa y esou ces, a d othe facto s desc ibed i SK Telecom’s epo ts filed with o fu ished to the U.S. Secu ities a d Excha ge Commissio , i cludi g its most ece t A ual Repo t o Fo m 20-F. Except as equi ed by law, SK Telecom u de takes o obligatio to update these stateme ts.
Ce tai stateme ts i this p ess elease i cludi g, but ot limited to, stateme ts as to: expectatio s with espect to g owth, pe fo ma ce, availability, a d be efits of NVIDIA’s p oducts, se vices a d tech ologies, a d elated t e ds a d d ive s; expectatio s with espect to NVIDIA’s thi d pa ty a a geme ts, i cludi g with its collabo ato s a d pa t e s; expectatio s with espect to tech ology developme ts, a d elated t e ds a d d ive s; p ojected ma ket g owth a d t e ds; expectatio s with espect to AI a d elated i dust ies; a d othe stateme ts that a e ot histo ical facts a e fo wa d-looki g stateme ts withi the mea i g of Sectio 27A of the Secu ities Act of 1933, as ame ded, a d Sectio 21E of the Secu ities Excha ge Act of 1934, as ame ded, which a e subject to the “safe ha bo ” c eated by those sectio s based o ma ageme t’s beliefs a d assumptio s a d o i fo matio cu e tly available to ma ageme t a d a e subject to isks a d u ce tai ties that could cause esults to be mate ially diffe e t tha expectatio s. Impo ta t facto s that could cause actual esults to diffe mate ially i clude: global eco omic a d political co ditio s; NVIDIA’s elia ce o thi d pa ties to ma ufactu e, assemble, package a d test NVIDIA’s p oducts; the impact of tech ological developme t a d competitio ; developme t of ew p oducts a d tech ologies o e ha ceme ts to NVIDIA’s existi g p oducts a d tech ologies; ma ket accepta ce of NVIDIA’s p oducts o NVIDIA’s pa t e s’ p oducts; desig , ma ufactu i g o softwa e defects; cha ges i co sume p efe e ces o dema ds; cha ges i i dust y sta da ds a d i te faces; u expected loss of pe fo ma ce of NVIDIA’s p oducts o tech ologies whe i teg ated i to systems; NVIDIA’s ability to ealize the pote tial be efits of busi ess i vestme ts o acquisitio s; a d cha ges i applicable laws a d egulatio s, as well as othe facto s detailed f om time to time i the most ece t epo ts NVIDIA files with the Secu ities a d Excha ge Commissio , o SEC, i cludi g, but ot limited to, its A ual Repo t o Fo m 10-K a d Qua te ly Repo ts o Fo m 10-Q. Copies of epo ts filed with the SEC a e posted o NVIDIA’s website a d a e available f om NVIDIA without cha ge. These fo wa d-looki g stateme ts a e ot gua a tees of futu e pe fo ma ce a d speak o ly as of the date he eof, a d, except as equi ed by law, NVIDIA disclaims a y obligatio to update these fo wa d-looki g stateme ts to eflect futu e eve ts o ci cumsta ces.
© 2026 NVIDIA Co po atio . All ights ese ved. NVIDIA, the NVIDIA logo, a d othe NVIDIA p oduct a d se vice ames a e t adema ks a d/o egiste ed t adema ks of NVIDIA Co po atio i the U.S. a d othe cou t ies a d egio s. Othe compa y, p oduct, a d se vice ames may be t adema ks of the espective compa ies with which they a e associated.
A photo accompa yi g this a ou ceme t is available at https://www.globe ewswi e.com/NewsRoom/Attachme tNg/e8f27268-4031-4122-8504-f4cf3ad6e87c
