Meet
YTL AI Cloud
Powering Tomorrow’s AI Innovations Today
- Sovereign AI
- Cloud Infrastructure
- AI Innovation
- Asia Pacific
70-year-old Malaysian conglomerate YTL Corporation deploys NeuralMesh™ by WEKA, enabling secure multi-tenancy, linear scalability, and sovereign AI development while powering the nation’s first locally developed large language model.
YTL AI Cloud Builds Malaysia’s Sovereign AI Future with WEKA
Founded in the 1950s by Tan Sri Yeoh Tiong Lay, YTL Corporation has grown from a construction company into one of Malaysia’s most diversified conglomerates, spanning power generation, cement, hospitality, real estate, rail, and digital technology. Now, as it marks more than 70 years of “Building The Right Thing,” YTL is spearheading its next frontier: artificial intelligence.
YTL AI Cloud was established to deliver next-generation AI computing infrastructure across Southeast Asia. Under CEO Philip Lin, the company is building a sovereign AI platform designed to empower businesses, researchers, and institutions across the region.
To bring this vision to life, YTL needed storage infrastructure that could match its ambitions.
The Challenge
Build Malaysia’s Premier AI Cloud From the Ground Up
YTL AI Cloud faced the task of deploying one of the largest NVIDIA GB200 NVL72 clusters in Southeast Asia—with demanding, non-negotiable performance requirements: sustained 1 GB/s read and 0.5 GB/s write throughput per GPU across hours of continuous operation, with linear scalability as the platform grew.
Equally critical was secure multi-tenancy: different customers sharing physical infrastructure while maintaining absolute performance isolation—without siloing expensive GPU resources. Beyond immediate needs, YTL sought a platform built to evolve with a rapidly changing AI landscape, with sustainability as a firm criterion for both environmental and business reasons.
The Solution
NeuralMesh by WEKA, Engineered for AI at Scale
After a rigorous evaluation, YTL selected NeuralMesh as its storage platform and partnered with ODM provider Wiwynn for hardware. The decision rested on performance, architectural innovation, proven track record, and shared values around sustainability and long-term partnership.
WEKA’s consistent IO500 benchmark rankings and customer references across Asia-Pacific—including Yotta and Firmus—demonstrated sustained throughput at scale. NeuralMesh was also the only storage solution approved for NVIDIA’s NCP reference architecture for GB200 deployments.
“NeuralMesh enabled us to deploy composable storage leveraging an operator-driven architecture that delivers secure multi-tenancy across our AI cloud infrastructure. On NVL72 systems, we consistently achieved approximately 1 GB/s of read throughput per GPU, with perfectly linear scaling, and sub-0.5ms latency on small random I/O operations. The partnership between WEKA and Wiwynn delivered a fully integrated, validated solution that met our performance requirements from day one.”
Technical Overview of the YTL AI Cloud Infrastructure
As a cloud innovator and operator, YTL AI Cloud empowers customers to provision anything from bare-metal servers to fully managed cloud instances, all through its proprietary, self-developed orchestration platform. To address diverse and growing AI workloads, this platform integrates WEKA storage with flexible access to various compute and storage configurations. WEKA’s NeuralMesh Composable clusters enable advanced multi-tenancy, securely carving out isolated storage resources matched to specific GPU configurations via separate VLANs, ensuring both bare-metal and containerized environments benefit from dedicated performance and complete security isolation.
The Impact
From Infrastructure to National Transformation
YTL AI Cloud’s platform is now ready to power a new generation of AI applications across key sectors. In education, it enables locally tailored learning tools. In healthcare, it supports AI solutions built for Malaysian needs. Financial services, telecoms, and media companies are building on this foundation to serve local customers more effectively. Universities can run cutting-edge research, and startups can access world-class infrastructure without leaving Malaysia.
This is what sovereign AI means in practice: democratizing access for businesses, researchers, and institutions that previously couldn’t afford it.
Looking Forward
The Next Phase of Malaysia’s AI Journey
As YTL plans its next phase of growth, scaling to GB300 and expanding its customer base, NeuralMesh will scale alongside it.
The partnership between YTL AI Cloud, WEKA, and Wiwynn demonstrates how Malaysian innovation, backed by leading technology partnerships, can compete globally. YTL is building Malaysia’s AI future on this foundation.
“The deployment and benchmarking process proved both efficient and reproducible, enabling rapid performance validation during our platform initialization phase.”
Secure Multi-tenancy, Scalability & Sovereign AI
Learn more about YTL AI Cloud and WEKA
Dive a little deeper
