Build Sustainable AI
On the world’s most energy efficient GPU cloud.
- CAPACITY AVAILABLE
- Better TCO
- CO2 savings
Built unlike any other cloud. Ever.
Train & inference with half the energy.
What GenAI service do you need?
GPU supercomputers in the cloud, co-designed with NVIDIA. H100 & L40S accelerators networked at hyperscale with RDMA InfiniBand. The power of bare metal that can be shaped to any use case.
Startups and Scaleups
"AI startups & scaleups requiring cost effective dedicated metal to experiment, train and inference GenAI workloads
Managed Colocation
Enterprises, government and education teams looking to add GPU cluster capacity to existing x86 workloads, with a better TCO and time-to-use than on-premise.
AI Platforms
Mid & large scale application and platforms developers who demand large-scale, high performance training and inference GPU clusters in geographically diverse locations
Wholesale & Hyperscale
Private supercomputing GPU clusters at hyperscale. 20,000+ GPUs in a single-cluster, customized to suit the bespoke requirements of enterprises and CSPs. Delivered at data centre level to expand private clusters or open new public AZs.
The power of SMC's cutting-edge GPU cloud, paired with NVIDIA AI Enterprise lets you get going on your journey to build sustainable AI. Pre-trained models, and over 100 frameworks, including NeMo, delivered with enterprise support from NVIIA.
Security Built In
Continuous monitoring and regular releases of security patches for critical and common vulnerabilities and exposures (CVEs), with enterprise support and access to NVIDIA AI experts.
LLM Training Speedup
Train at scale with the NVIDIA TAO Toolkit. Create custom, production-ready AI models in hours, rather than months, by fine-tuning NVIDIA pretrained models - without AI expertise or large training datasets.
Reduce Development Time
Develop faster with NVIDIA AI Workflows: prepackaged reference examples for enterprises to jumpstart the building of AI solutions.
Deploy at Scale
Deploy at scale with NVIDIA Triton Inference Server, which simplifies and optimizes the deployment of AI models at scale and in production for both neural networks and tree-based models on GPUs.
Really powerful AI infrastructure
Access dedicated scalable NVIDIA-designed GPU clusters.
Fast & flexible storage
Cost effective, flexible & powerful storage solutions designed to support the training the world's largest AI models
Fast File Storage
- Parallel NVMe file storage, delivered on the Weka platform
- Optimised for high IOPS and low latency
- Connected at up to 200gbps to every SMC instance
- Flexible, zero-lock in consumption
Object Storage
- S3 SDK compatible object storage
- Highly expandable and scalable
- Short term and long term flexibility with best in class cost
Networking
Cluster networking to allow performance and scale. External connectivity to any cloud.
Global Connectivity
- S3 SDK compatible object storage
- Highly expandable and scalable
- Short term and long term flexibility with best in class cost
RMDA Clusters
- NDR InfiniBand, full-rail RDMA enabled compute fabric
- High-speed Ethernet for in-band management, up to 200gbps
- Instances equipped with BlueField-3 DPUs
Sustainability is the foundation of everything we do. It also makes everything so much better.
SMC infrastructure is powered by revolutionary hosting technology in the data center that cuts energy use for GPU clusters by up to 48% versus legacy clouds. Liquid cooling technology on a massive scale makes saves both carbon emissions and power costs, and boosts reliability and uptime in the process. Available everywhere we operate.
when training AI models on SMC vs GPU instances on other clouds*
* For SMC H100 instances running in Singapore (SIN02) vs energy consumption for an air-cooled H100 SXM-8 system running within a PUE 1.30 data centre in Singapore.
Future state
12 April 2024
Apr 2024
InterviewsSenior Solution Architect: Sean Zhan on Computability
Looking back, the cost of computation has reduced dramatically. With nearly 20 years of experience in the industry, Sean Zhan talks about the beginning of the HPC revolution and how it affects computation today.
2 April 2024
Apr 2024
announcementsSustainable Metal Cloud Joins MLCommons
25 March 2024
Mar 2024
announcementsBuilding a greener AI: The GreenFutures Accelerator Program
The GreenFutures creates a fostering environment that empowers start-ups and scale-ups with GPU access, mentorship, education, and exposure. It is a catalyst to accelerate AI development, one cohort at a time.
Globally available, locally secured
Our commitment to sustainability, rooted in Asia, extends globally for accessibility anywhere. As a member of the ST Telemedia Global Data Centres group, we are scaling rapidly across a platform of over 85 availability zones in Asia and Europe.
Open positions
Join a team committed to powering a better future. Explore our job opportunities.
GreenFutures is SMC’s accelerator program dedicated to propelling startups and scale-ups to begin and grow their AI journey on a foundation of sustainability.