Huang In-hoon's dramatic “Save” Korea Stock Market
Author:Suyan
Editor: Seo Cyangyang
ON 5 JUNE, THE KOREAN STOCK MARKET EXPERIENCED A “BLACK FRIDAY”, AND THE KOSPI INDEX FELL BY 5.54 PER CENT. AFTER OPENING ON 8 JUNE, THE DROP IN THE DISC WAS INCREASED TO OVER 8 PER CENT, THE EXCHANGE TRIGGERED THE MELTING MECHANISM, AND SAMSUNG AND HIS FAMILY WERE KILLEDSK HERCULESDouble down by almost 10 per cent。
In-hoon Hoang played a dramatic role as a “saving city” when the market was full of news。
Previously, on Sunday night, 7 June, Korean local timeSK GROUPTHE PRESIDENT, CHOI TAE-JIN, AND THE CEO OF THE SK HERCULES, GUOLU, ARE ORGANIZING A DINNER PARTY。
After the meal, In-hoon Hwang confirmed to the media that the newly launched Vera CPU in In Weida would use the SK Hercules DRAM; that the two sides were preparing for “super-large cooperation” later this year and next; and that he believed that the current shortage of memory chips would last for several years。
SUBSEQUENTLY, YIN WEIDA ANNOUNCED A MULTI-YEAR TECHNICAL COOPERATION AGREEMENT WITH S.K. HEILLEY ON THE EXTENSION OF AI SUPERCOMPUTERS TO ROBOTIC, DIGITAL TWIN AND SEMICONDUCTOR MANUFACTURING。
WONG IN-HOON GAVE A DIRECT BLOW-OUT AT THE PRESS CONFERENCE, “IF YOU WERE A SHAREHOLDER IN AI, YOU WOULD BE HAPPY THAT THEY ARE CURRENTLY VERY LOW”
01 LOCK SK HERCULES RAM
Vera is a dedicated CPU for the first independent data centre launched by Weida, and faces competitions such as the Intel to Power product line, the AMD Epyc chip and a self-study project for large cloud service providers such as Graviton, Amazon。
IN THIS NEW BATTLEGROUND, YOUNG WEIDAR HAS, FROM THE VERY BEGINNING, PLACED THE RAM ANCHOR ON KK HERCULES。
ON 7 JUNE, YING WEIDA AND SK HERCULES OFFICIALLY ANNOUNCED THE ESTABLISHMENT OF A MULTI-YEAR TECHNICAL COOPERATION PARTNERSHIP TO JOINTLY DEVELOP THE NEXT GENERATION OF MEMORY MATCHING THE YING WEIDAI INFRASTRUCTURE ROAD MAP。
It is understood that the cooperation covers a range of personal side and cloud-side product lines, such as the Vera Rubin AI supercomputer, Vera CPU, PC with RTX Spark and the Jason Thor robotics platform。
THE BULLETIN STATES THAT COOPERATION IS AIMED AT SAFEGUARDING THE SUPPLY OF ADVANCED MEMORY IN RESPONSE TO THE LONGER DEVELOPMENT CYCLES, COMPLEX MANUFACTURING PROCESSES AND HIGH CAPITAL INPUTS OF SUCH PRODUCTS, THEREBY SUPPORTING THE CONTINUED CONSTRUCTION OF THE GLOBAL AI PLANT。
THE BULLETIN ALSO LISTS SEVERAL NEW MARKETS IN WHICH SK HERCULES WILL EXPAND INTO BRITAIN, INCLUDING AI INFRASTRUCTURE, PERSONAL AI, AND THE PRIVATE SECTORPHYSICAL AII don't know。
02 AI ANTI-NURSING CHIP MANUFACTURING
IN ADDITION TO THE SUPPLY OF MEMORY, SK HERCULES BEGAN TO APPLY THE AI TECHNOLOGY OF INVERDA TO THE DESIGN AND MANUFACTURING PROCESS OF THE CHIP。
A similar collaboration had previously been in the ground of power dumps, most typically “calculating the moment”。
ACCORDING TO THE BULLETIN, SK HERCULES IS USING THE CUDA-X LIBRARY AND AI IN WEVDA TO ACCELERATE SEMICONDUCTOR SIMULATIONS, COVERING THE TECHNICAL COMPUTER-AIDED DESIGN (TCAD) AND COMPUTATIONAL OPTICAL ENGRAVINGS。
THE PARTIES ARE ALSO PROMOTING THE EXTENSION OF THESE TOOLS TO SEMICONDUCTOR ELECTRONIC DESIGN AUTOMATION (EDA) AND SIMULATION ECOSYSTEMS, PAVING THE WAY FOR TRIPARTITE COOPERATION BETWEEN CHIP MANUFACTURERS, YVETTE AND EDA SOFTWARE PROVIDERS。
THIS MEANS THAT THE COOPERATION BETWEEN THE PARTIES IS NO LONGER LIMITED TO THE USE OF SK HERCULES, BUT IS EXPLORING A MODEL THAT CAN BE PUSHED TOWARDS THE ENTIRE SEMICONDUCTOR INDUSTRY。
At the manufacturing stage, SK Hercules is advancing the development of the digital twinning function of the crystal circle plant, with the aim of achieving full autonomous plant operation. The basis for this work is the Emnivre platform. Through the Omniverse and OpenUSD processes, SK Hercules was able to construct 3D plant scenes for visualization, simulation and optimization of complex semiconductor manufacturing environments。
At the operational level of the plant, these digital twin functions also provide access to the U.W.C.Opt decision optimization engine and the Metropolis platform, which is used to move autonomous mobile robots and other assets within the crystal circle plant。
THE BULLETIN ALSO REVEALED THAT THE TWO COMPANIES WERE EXPLORING WAYS TO CONNECT DIGITAL TWINNING WITH EXISTING TRADITIONAL SOFTWARE AND AI WORK STREAMS OF SMARTS TO ALLOW AI SYSTEMS TO REASON ON THE BASIS OF CRYSTAL CIRCLE DATA, AUTO-MANDATE AND IMPROVE MANUFACTURING DECISION-MAKING。
03, six months ahead
IN OCTOBER 2025, INWEIDA AND SK HERCULES ANNOUNCED A MASSIVE INFRASTRUCTURE COOPERATION。
AT THAT TIME, THE SK GROUP WAS BUILDING AN AI PLANT WITH MORE THAN 50,000 BRITISH GPUS, THE FIRST PHASE OF WHICH WAS SCHEDULED TO BE COMPLETED BY 2027. ONCE COMPLETED, THIS IS EXPECTED TO BE ONE OF THE LARGEST AI FACTORIES IN KOREA。
THE PLANT, WHICH USES THE “GPU AS A SERVICE” MODEL, IS OPEN TO SK GROUP AFFILIATES AND EXTERNAL ORGANIZATIONS WITH THE AIM OF ACCELERATING DIGITAL TRANSFORMATION OF KOREAN INDUSTRIES AND INDUSTRIAL INNOVATION。
SK TELECOMMUNICATIONS ALSO UNDERTOOK SPECIFIC DEPLOYMENTS。
As a cloud partner in Weeda, SK Telecommunication plans to build industrial AI clouds in Asia using the GPU version of the British Weeda RTX PRO 6000 Blackwell server, initially deployed on a scale of over 2,000 GPUs, to operate the Omniverse task load and provide a mathematical support for the semiconductor manufacturing of SK Hercules, the digital twinning of the crystal mill and the internal AI intelligence body。
HOANG IN-HOON ALSO REVEALED A MESSAGE DURING THIS VISIT: HE IS DISCUSSING WITH THE TELECOMMUNICATIONS COMPANY BECAUSE THE FUTURE AI WILL USE THE TELECOMMUNICATIONS NETWORK. THIS IS IN LINE WITH THE DIRECTION OF SK TELECOMMUNICATIONS INVOLVEMENT。
HBM4 ORDERS
IN SPITE OF THE MULTI-YEAR TECHNICAL COOPERATION AGREEMENT BETWEEN YVEDA AND SK HERCULES, YVEIDA DID NOT PLACE EGGS IN A BASKET ON HBM4 SUPPLY。
On arrival in Seoul, Huang In-hoon made a clear statement to journalists: “All three suppliers have been accredited. All three suppliers are in production and they are competing to support Vera Rubin.”
THESE THREE SUPPLIERS CORRESPOND TO TRISTAR ELECTRONICS, SK HERCULES ANDLight technologyI don't know。
In a keynote speech at the Taipei Computer Fair, Hoang In-hoon has confirmed that Vera Rubin is fully operational and is scheduled to be delivered in the third quarter of this year. The system is built around the core clusters of Vera CPU and Rubin graphics in Weida, and each server rack system is equipped with a TB level HBM4 memory。
BASED ON THE ACTUAL PROGRESS OF HBM4, SK HERCULES REMAINS IN THE LEAD POSITION。
Reuters reported last September that SK Hercules had completed internal certification of HBM4 chips and established a production system for clients, with the goal of completing production preparation for 12 floor HBM4 products in late 2025. Kim Sunwoo, Senior Analyst of Meritz Securities, predicted at that time that, thanks to early supply to key customers and the pre-emptive advantages resulting therefrom, the HBM market share of SK Hercules in 2026 would remain at 60 per cent。
05 The chip shortage will last for years
THE CREATION OF THREE UNITS ON THE SUPPLY OF HBM4 DOES NOT MEAN THAT THE SUPPLY PRESSURE WILL BE EASED。
Wong In-hoon also gave a less optimistic judgment after Sunday dinner. He indicated to the media on the ground that the shortfall in the storage of chips would not end soon. “The entire supply chain of the industry — from the crystal circle to the encapsulation to the silicon photons — is in short supply because the demand is so high. This will continue for several years.”
THIS STATEMENT WAS MADE AGAINST THE BACKGROUND OF THE ALMOST ENDLESS CONSUMPTION OF ADVANCED MEMORY FOR THE CONSTRUCTION OF THE GLOBAL AI PLANT。
Wong In-hoon's alleged shortage is not a lack of something, but a strain on almost every link of the industrial chain. The launch of Vera Rubin, the promotion of the AI factory, the entry into the personal and physical AI fields in Weida is raising the need for memory in every instance. That is why he would say that the three HBM4 suppliers are competing for Vera Rubin。
No one wants to lose in a situation where there is no demand。
This trip to Korea by the SK Group, though focused, is not the whole schedule of Huang In-hoon. On arrival, he disclosed that he had arranged meetings with modern cars, LG, SK, Samsung and Naver. He also revealed that Inweida was actively recruiting for the Korea New Research and Development Centre. From these trends, the SK Group is a key, but not the only, link to the entire Korean technology industry。
