DEEP RESEARCH · SK TELECOM × LAMBDA
[SK Telecom] Lambda Partnership and Investment
A research memo on SKT's Lambda equity investment, Korea region, GPUaaS, and Gasan AI data-center cooperation
0. Bottom line first
SKT's Lambda investment looks less like a financial investment and more like a strategy to secure GPU supply, cloud platform capability, and operating know-how for a full AI data-center business.
The source first links a separate note explaining Lambda: [SK Telecom] What is Lambda, the GPU-farm partner?.
1. Investment size and nature
Official fact: SK Telecom participated in Lambda's Series C round in early 2024. The round totaled USD 320 million, about KRW 430 billion, and valued Lambda at about USD 1.5 billion.
Official fact: SKT's exact investment amount and ownership percentage were not disclosed. The source describes SKT as a strategic investor that secured a portion of Lambda equity.
Interpretation: I see the purpose as securing stable GPU resources and AI-cloud expertise. The direct equity investment reads as preparation to accelerate AI data-center buildout.
2. Partnership contract core
Gasan GPU cluster
The two companies agreed to jointly build a large NVIDIA GPU cluster at SK Broadband's Gasan data center in Seoul by year-end, including H100 Tensor Core-class GPUs.
GPUaaS launch
The goal is a GPU subscription cloud service letting Korean companies and institutions rent the amount of GPU capacity they need.
Lambda Korea region
Lambda's Korea region is its first Asia-Pacific region and is built inside SKT's Gasan AI center.
Overseas cooperation
The partnership leaves room for joint overseas AI cloud expansion after proving the Korean model.
3. Collaboration model and group synergy
Official fact: The source says Lambda is known for receiving the latest Nvidia GPUs and offering cloud services, and that SKT established a basis for high-performance GPU procurement through the cooperation.
Interpretation: For SKT, the value is not only GPU procurement but also Lambda's AI cloud software stack and operating know-how. This can shorten the time needed to launch a stable service.
- SK Broadband: data-center operating know-how
- SK hynix: HBM and high-bandwidth memory capability
- SAPEON: SKT subsidiary AI semiconductor capability
- Lambda: GPU cloud platform and global GPU supply chain
4. Market outlook and investment points
| Item | Key source figure or point | Investment interpretation |
|---|---|---|
| GPUaaS market | Projected to reach about USD 49.84 billion, or about KRW 65 trillion, by 2032 | Demand for high-performance AI compute underpins SKT-Lambda cooperation |
| Korea launch | Source describes a December 2024 launch of Korea's first GPU-dedicated AI data center and GPUaaS service | Revenue conversion from the launched service is the next check |
| Next GPU | Plan to introduce Nvidia H200 early in Q1 2025 | GPU procurement capability is central to SKT's AI infrastructure competitiveness |
| Global expansion | Southeast Asia is considered as the first global target | If the Korean model is replicated abroad, the partnership can become a broader AI infrastructure alliance |
5. Conclusion
In conclusion, SK Telecom is trying to combine telecom network capability with Lambda's AI cloud technology to strengthen its position in domestic and overseas AI infrastructure. If stable GPU supply and advanced AI cloud technology continue, the SKT-Lambda alliance could have a meaningful edge in AI data centers and cloud services.
Sources
- Original Naver Blog post: https://m.blog.naver.com/PostView.naver?blogId=star_of_self&logNo=223769314912
- Lambda company-analysis link: https://m.blog.naver.com/star_of_self/223769311574
- Sources named in original: SKT official press releases and newsroom, Lambda official announcement, Yonhap Infomax, Digital Today, KED Global, and other investment media