NVIDIA AI Cloud Ecosystem Expands Worldwide to Meet Global AI Compute Demand

The NVIDIA AI Cloud ecosystem is accelerating the worldwide buildout of AI manufacturing unit infrastructure. Companions are increasing capability to meet rising demand from enterprises, startups, nations, AI labs and builders scaling agentic AI purposes. 

NVIDIA AI Clouds are a rising ecosystem of purpose-built clouds serving the exploding token demand behind at the moment’s hottest AI purposes. These AI clouds have been co-designed with NVIDIA’s full-stack AI infrastructure to meet surging demand for AI from enterprises, startups and nations on the lookout for new distributors and regional capability. 

They mix NVIDIA accelerated computing, networking and AI software program to assist companions assist coaching, fine-tuning, inference, agentic AI, bodily AI and sovereign AI deployments. Particular configurations range by companion and workload.

AI cloud companions select NVIDIA for the perfect economics — lowest token value, greatest throughput per watt — to run frontier and open supply AI. Constructed with NVIDIA accelerated computing, networking and AI software program, these clouds convey AI factories nearer to the place knowledge, builders, customers and industries are, serving to prospects practice, tune and run agentic AI purposes at scale. The ecosystem spans practically each geography, supporting regional and sovereign AI capability for frontier mannequin builders, enterprises, startups, software program suppliers and nationwide AI applications.

“Each firm and each nation wants AI manufacturing unit infrastructure to flip knowledge into intelligence,” mentioned Jensen Huang, founder and CEO of NVIDIA. “NVIDIA AI Clouds convey full-stack AI factories nearer to the areas, industries and builders constructing the following era of AI, from mannequin coaching to real-time inference and AI brokers that can rework how folks and organizations work.”

Broad AI Cloud Ecosystem

AI cloud suppliers, telcos, sovereign AI builders and vertically built-in infrastructure suppliers are constructing AI factories with NVIDIA to serve prospects throughout frontier AI, enterprise AI, telecommunications, developer clouds and nationwide AI applications.

Regional progress is accelerating throughout Southeast Asia, Australia and the Americas, with NVIDIA AI Clouds now reaching six continents following the addition of Cassava in Africa and Claro in South America.

NVIDIA AI Clouds are pairing large-scale AI manufacturing unit buildouts with demand from main AI labs, enterprises, governments and digital service suppliers. Companions together with CoreWeave, Firmus, IREN and Nscale are increasing AI infrastructure to assist frontier mannequin growth, enterprise AI, agentic purposes and high-volume inference.

Throughout areas, NVIDIA AI Clouds are bringing AI factories nearer to native industries and sovereign AI ecosystems. Companions together with Firebird, GMI Cloud, Indosat Ooredoo Hutchison, Lambda, Naver Cloud, Sharon AI, Yotta and YTL are supporting rising AI firms, nationwide AI initiatives, monetary companies, telecommunications, manufacturing, schooling, healthcare and developer ecosystems.

For governments and controlled industries, regional AI clouds can assist sovereign controls and native compliance necessities. For builders and enterprises, they will cut back friction in accessing accelerated infrastructure for AI brokers, enterprise copilots, digital staff and different AI companies that should run shut to customers and knowledge.

Firmus Expands AI Manufacturing facility Footprint Throughout Australia and Asia-Pacific

Firmus Applied sciences is increasing its AI manufacturing unit footprint throughout South Australia and Southeast Asia, constructing energy-efficient infrastructure to assist rising demand for large-scale coaching, inference and agentic AI workloads.

Via Venture Southgate, Firmus is creating AI factories in Tasmania, Melbourne, South Australia and New South Wales, with an emphasis on renewable energy, superior cooling and modular infrastructure that may convey capability on-line sooner. The corporate has additionally deployed AI infrastructure in Singapore by means of a partnership with ST Telemedia Global Information Centres.

Firmus is utilizing NVIDIA’s accelerated computing and reference structure as a part of its buildout, with NVIDIA DSX serving to streamline AI manufacturing unit design, deployment and operations.

Engineered in alignment with the NVIDIA DSX platform, the liquid-cooled Firmus HyperCube is designed to fast-track modular AI Manufacturing facility builds and optimize for low value per token. Firmus is innovating throughout the AI manufacturing unit provide chain, together with cooling and vitality.

“AI brokers are creating a brand new class of industrial-scale demand for tokens, and Asia-Pacific wants AI factories that may be constructed sooner, liquid-cooled extra effectively and operated at gigawatt scale,” mentioned Tim Rosenfield, co-CEO of Firmus. “Along with NVIDIA, Firmus is constructing liquid-cooled, AI infrastructure designed to ship AI tokens as effectively and quickly as potential for the area’s most vital prospects.”

CoreWeave Advances Bodily AI and Subsequent-Era AI Factories

CoreWeave is increasing its NVIDIA AI Cloud platform to assist the following wave of agentic AI, bodily AI and frontier mannequin workloads. 

An early adopter of NVIDIA Vera Rubin and the NVIDIA Vera CPU, CoreWeave can also be among the many first to undertake NVIDIA Spectrum-X Ethernet Photonics, serving to present the networking basis for million-GPU AI factories. CoreWeave is extending its platform for robotics and bodily AI workflows, together with utilizing NVIDIA Cosmos 3, the most recent frontier world basis mannequin, to assist groups generate artificial knowledge, fine-tune fashions and speed up robotics knowledge flywheels. Main AI labs, together with Anthropic, construct on CoreWeave’s infrastructure to assist frontier fashions at scale.

“AI factories have gotten the inspiration for the agentic period,” mentioned Michael Intrator, cofounder, chairman and CEO of CoreWeave. “Along with NVIDIA, CoreWeave is constructing the full-stack cloud infrastructure that offers AI labs, enterprises and builders the efficiency, scale and reliability they want to flip frontier fashions, AI brokers and bodily AI programs into manufacturing purposes.”

Nebius Builds an Open Bodily AI Workbench for Agentic Workflows

Nebius is increasing its NVIDIA AI Cloud with a full-stack platform for coaching, inference and bodily AI growth.

An early adopter of NVIDIA Vera Rubin, Nebius is constructing built-in AI infrastructure from silicon to software program, together with its Nebius AI Cloud, Token Manufacturing facility inference layer and new Bodily AI Workbench. The workbench brings applied sciences together with NVIDIA Cosmos 3, NVIDIA Isaac Sim and Isaac GR00T into composable workflows that may be assembled by AI brokers, serving to robotics and autonomous programs groups transfer sooner from simulation and artificial knowledge to coaching and analysis.

“Builders ought to find a way to construct AI programs with out spending weeks wiring collectively infrastructure,” mentioned Arkady Volozh, founder and CEO of Nebius. “With NVIDIA, Nebius is creating an AI cloud the place AI  brokers can compose the instruments, knowledge and compute wanted to speed up AI workloads — from robotics and life sciences to the enterprise — from experimentation to manufacturing.”

NVIDIA Exemplar Cloud Momentum

Since NVIDIA launched Exemplar Cloud final 12 months, six NVIDIA Cloud Companions have achieved Exemplar Cloud standing: CoreWeave, Crusoe, Lambda, Nebius, Vultr and YTL. The rising roster displays rising demand for AI cloud infrastructure that may ship constant efficiency, reliability and effectivity for manufacturing AI workloads.

These suppliers are serving to increase the efficiency bar throughout the AI cloud ecosystem, giving enterprises, builders and AI labs extra validated choices for scaling coaching, inference and agentic AI companies.

Engineered for AI Manufacturing facility Economics

As AI shifts from mannequin growth to reasoning and high-volume inference, the measure of infrastructure is not simply capability introduced but in addition the economics of token output pushed by platform utilization, uptime, lengthy asset life and the breadth and depth of helpful AI brokers folks can put to work. 

Constructed on NVIDIA full-stack AI manufacturing unit platforms, AI Clouds assist companions optimize infrastructure for these measures.

Price per token is the entire value of possession metric that straight accounts for {hardware} efficiency, software program optimization, ecosystem assist and real-world utilization. NVIDIA delivers the lowest value per token within the business, pushed by delivered token throughput, software program optimization and full-stack codesign throughout compute, networking, reminiscence and storage.

DSX Helps AI Clouds Carry Capability On-line Quicker

NVIDIA AI Clouds are adopting the NVIDIA DSX platform to design, construct and function AI factories.

DSX brings collectively validated reference designs, simulation, software program and ecosystem applied sciences to assist cloud suppliers convey capability on-line sooner, function extra effectively and maximize income.

DSX Sim helps groups mannequin and validate AI factories earlier than deployment. DSX Flex helps AI factories dynamically adapt workloads to grid situations. DSX MaxLPS helps power-constrained AI factories maximize compute inside a hard and fast energy finances, enabling up to 40% extra GPUs. DSX OS helps automate lifecycle administration and operations at scale.

DSX helps AI Clouds cut back deployment danger, enhance resiliency, ship extra tokens per watt and obtain the bottom value token.