---
title: "The demand for AI computing power continues to surge! South Korea's \"AI Squid Game\" is in full swing, igniting the demand for tens of thousands of AMDAI chips"
type: "News"
locale: "zh-HK"
url: "https://longbridge.com/zh-HK/news/280114691.md"
description: "South Korean artificial intelligence startup Upstage is in talks with AMD to purchase 10,000 AI accelerators to boost its AI computing infrastructure. AMD's AI solutions have gained market recognition and are expected to eat into NVIDIA's 90% market share. Upstage's CEO expressed a desire to diversify computing power supply and shift towards other AI chips like AMD. The company is participating in a super AI competition supported by the South Korean government, aimed at enhancing South Korea's AI capabilities"
datetime: "2026-03-23T07:16:24.000Z"
locales:
  - [zh-CN](https://longbridge.com/zh-CN/news/280114691.md)
  - [en](https://longbridge.com/en/news/280114691.md)
  - [zh-HK](https://longbridge.com/zh-HK/news/280114691.md)
---

> 支持的語言: [简体中文](https://longbridge.com/zh-CN/news/280114691.md) | [English](https://longbridge.com/en/news/280114691.md)


# The demand for AI computing power continues to surge! South Korea's "AI Squid Game" is in full swing, igniting the demand for tens of thousands of AMDAI chips

According to media reports, South Korean artificial intelligence startup Upstage is in talks with AMD (AMD.US), a leader in high-performance chips for PCs and data centers, to purchase its latest 10,000 AI accelerators. This move is an important part of its efforts to introduce larger-scale AI computing infrastructure into the South Korean market.

The South Korean AI startup is purchasing a total of 10,000 AMD AI chips. Coupled with AMD's recent significant collaborations with technology leaders such as Tianhong Technology (CLS.US) and HPE (Hewlett Packard Enterprise) (HPE.US) to accelerate the ramp-up of AMD's Helios rack-level AI computing infrastructure capacity, this indicates that AMD's AI computing solutions are gaining more positive market recognition and are expected to continue to encroach on NVIDIA's (NVDA.US) market share, which is as high as 90%, in the trillion-dollar AI core computing cluster field.

Upstage CEO Sung Kim stated in a media interview that last week he met with AMD's CEO Lisa Su in Seoul to discuss the procurement of AMD MI355 AI accelerator computing products. Kim mentioned in a media interview on Monday: "We have many NVIDIA chips in the South Korean market, but we want to diversify our computing supply and turn to other AI chip suppliers, including AMD."

It is understood that Upstage is one of four competing teams participating in a super AI competition supported by the South Korean government to select the best national AI foundational model. This competition, dubbed the "AI Squid Game," borrows its name from the popular survival-themed series created by Netflix's Korean team and is an important part of South Korea's ambition to become a top global AI powerhouse.

A professional review panel supervised by the South Korean Ministry of Science and Information and Communication Technology will evaluate and eliminate these teams' AI foundational models every six months. South Korea plans to select two teams to enter the final stage by early next year. The winners will receive more NVIDIA AI GPU computing infrastructure.

Kim stated that Upstage is currently preparing a super large language model with approximately 200 billion parameters for an important new round of competition coming this summer. He added that the advantage of this South Korean AI startup lies in combining scaling effects with high-efficiency processing models to create high-performance AI large models at relatively lower costs—this approach aims to compete against AI large model competitors from China and the United States that focus on high cost-performance ratios.

Upstage is a leading South Korean AI startup focused on AI large models and enterprise-level AI software solutions. It currently holds two prominent positions in the industry: first, it is one of the four competing teams in the South Korean government's "Sovereign AI Foundational Model" competition; second, the AI startup disclosed that it has raised over $100 million cumulatively by 2024, claiming to be the most funded AI large model company in South Korean history. This AI startup not only develops general-level AI large models but also actively bets on enterprise-level Document AI and LLM+ sovereign AI going overseas Official data shows that this AI startup focuses on "AI+" scenarios in key sectors such as finance, insurance, healthcare, and high-end manufacturing.

Kim also stated that he is considering countries in Asia, such as Vietnam and the UAE, as significant potential target markets for providing sovereign-level AI training/inference systems that can be deployed within their own borders.

**AMD steps into the era of rack-level AI infrastructure! Striving to expand Helios computing cluster capacity**

Upstage is in talks with AMD to purchase 10,000 MI355 units, and its CEO has clearly stated that there are already a large number of NVIDIA chips in South Korea, but they hope to deploy a "diversification" strategy with AMD. This indicates that AMD has begun to transition from being an "optional AI GPU alternative" to a large-scale AI computing infrastructure deployment option that some customers are genuinely considering.

Last week, media reports indicated that AMD announced a deep collaboration with Tianhong Technology, aiming to launch its new Helios rack-level AI computing infrastructure platform to compete with NVIDIA's NVL72 rack-level AI platform in the global AI data center market. These two latest pieces of news combined are enough to show that AMD's AI computing cluster solution is gaining more positive market recognition. More importantly, Helios is not just a single card story, but a rack-level story: AMD has elevated the competition from a single GPU to a 72-card rack, network interconnection, and an integrated platform level of CPU+GPU+NIC, and has partnered with Tianhong Technology to accelerate mass production.

As AMD collaborates with Celestica to accelerate the market launch of the Helios rack-level AI platform, it coincides with AMD's partnerships with several tech leaders to counter NVIDIA's vertically integrated AI computing infrastructure solutions. Previously, AMD announced collaborations with Huizhi Technology and Broadcom, aiming to provide open and rack-scale artificial intelligence computing infrastructure for high-performance computing clusters and large AI data centers, and striving to accelerate the global progress of "Sovereign AI" research.

Helios is expected to be available to customers in late 2026. Meta, the parent company of Facebook, has signed a multi-generational and long-term deep cooperation agreement with AMD, planning to deploy up to 6 gigawatts of AMD Instinct GPU computing clusters, with the first gigawatt-level deployment expected to start in the second half of 2026; OpenAI has also participated in the design optimization of AMD MI450. Coupled with the strong demand for "sovereign AI/local large-scale computing" from South Korean Upstage, what we see is not a single AI computing order, but a trend: more and more customers are unwilling to bet all their AI infrastructure on a single supplier, and AMD is perfectly positioned to meet this demand for "second core alternative sources + open standards + reducing single vendor lock-in."

**Is the "King of the Hills" AMD stock price about to enter a new bull market curve?**

Undoubtedly, these latest positive catalysts undoubtedly constitute a strong short- to medium-term bull market driving force for AMD's stock price outlook. AMD has transformed itself from a follower in the AI chip market to a competitor in AI training/inference system-level infrastructure; The company itself has also set very aggressive targets for its Analyst Day in 2025, including an annual revenue scale of $100 billion for data center chips over the next five years and a compound annual growth rate (CAGR) of over 80% for revenue related to AI computing power in data centers.

AMD CEO Lisa Su also projected during the Analyst Day that the total addressable market (TAM) for AI data centers, including AI central processing units, AI accelerators, and high-performance networking products, is expected to exceed $1 trillion by 2030, far surpassing the approximately $200 billion expected in 2025, implying a CAGR of over 40%. Regarding overall profits, Su expects the company's earnings per share (EPS) to increase to $20 within three to five years.

AMD is viewed by Wall Street financial giant Citigroup analysts as the "king of the hill," with a target stock price of up to $260 within 12 months. Analyst target prices compiled by TIPRANKS show that the average target price on Wall Street is an astonishing $285, indicating a potential upside of up to 42% over the next 12 months. As of last Friday's close, AMD's stock price was $201.330.

![1774249822(1).png](https://imageproxy.pbkrs.com/https://img.zhitongcaijing.com/image/20260323/1774249828759488.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg)

NVIDIA CEO Jensen Huang showcased NVIDIA's "unprecedented AI computing revenue super blueprint" in the AI computing infrastructure field at the GTC conference on March 17th, Beijing time. He informed global investors that, driven by strong demand for Blackwell architecture GPUs and the upcoming mass production of the Vera Rubin architecture AI computing system, its future revenue scale in the artificial intelligence chip sector could reach at least $1 trillion by 2027, far exceeding the $500 billion AI computing infrastructure blueprint projected at the last GTC conference for 2026.

As model scale, inference links, and multi-modal/agentic AI workloads drive exponential growth in computing power consumption, tech giants' capital expenditures are increasingly focused on AI computing infrastructure. Global investors are continuing to anchor the "AI bull market narrative" around NVIDIA, Google TPU clusters, and AMD's new product iterations and AI computing cluster delivery expectations as one of the most certain investment narratives in the global stock market. This also means that investment themes closely related to AI training/inference, such as power, liquid cooling systems, and optical interconnect supply chains, will continue to rank among the hottest investment camps in the stock market, even amid uncertainties in the geopolitical situation in the Middle East, led by AI computing leaders like NVIDIA, AMD, Broadcom, TSMC, and Micron.

According to Wall Street giants Morgan Stanley, Citigroup, Loop Capital, and Wedbush, the global investment wave in artificial intelligence infrastructure centered on AI computing hardware is far from over and is only at the beginning, driven by an unprecedented "storm of demand for AI inference computing power." The current wave of global AI infrastructure investment, lasting until 2030, is expected to reach a scale of USD 3 trillion to USD 4 trillion

### 相關股票

- [NVIDIA (NVDA.US)](https://longbridge.com/zh-HK/quote/NVDA.US.md)
- [Global X Data Center & Dgtl Infrs ETF (DTCR.US)](https://longbridge.com/zh-HK/quote/DTCR.US.md)
- [VanEck Semiconductor ETF (SMH.US)](https://longbridge.com/zh-HK/quote/SMH.US.md)
- [GraniteShares 2x Long AMD Daily ETF (AMDL.US)](https://longbridge.com/zh-HK/quote/AMDL.US.md)
- [SPDR S&P Semicon (XSD.US)](https://longbridge.com/zh-HK/quote/XSD.US.md)
- [Direxion Semicon Bull 3X (SOXL.US)](https://longbridge.com/zh-HK/quote/SOXL.US.md)
- [Invesco Semiconductors ETF (PSI.US)](https://longbridge.com/zh-HK/quote/PSI.US.md)
- [iShares Semiconductor ETF (SOXX.US)](https://longbridge.com/zh-HK/quote/SOXX.US.md)
- [Invesco AI and Next Gen Software ETF (IGPT.US)](https://longbridge.com/zh-HK/quote/IGPT.US.md)
- [AMD (AMD.US)](https://longbridge.com/zh-HK/quote/AMD.US.md)

## 相關資訊與研究

- [AMD Eyes Major AI Win with 10,000-Chip Deal in Korea](https://longbridge.com/zh-HK/news/280108451.md)
- [What's Going On With Credo Technology On Tuesday?](https://longbridge.com/zh-HK/news/279475198.md)
- [Vietnam Backs AMD R&D Hub, Targets 9,000 Trained Tech Workers by 2030](https://longbridge.com/zh-HK/news/279971683.md)
- [Samsung Elec and AMD sign MoU on AI memory, explore foundry partnership](https://longbridge.com/zh-HK/news/279553677.md)
- [Nextech3D.ai Hits Profit Milestone With Krafty Labs as Recurring Revenue Base Grows](https://longbridge.com/zh-HK/news/279598895.md)