PaleBlueDot AI

PaleBlueDot AI is an AI computing and cloud infrastructure services company founded in 2024 and headquartered in Palo Alto, California, United States. The company provides computing resources for model training, inference, and deployment to enterprises and developers worldwide, with offices in the United States, Singapore, Japan, and South Korea.

Dedicated GPU Clusters: The company provides customized cluster solutions and reserved computing capacity for large-scale or specialized workloads, enabling enterprises to allocate dedicated computing resources according to their project requirements.

GPU Marketplace: The company connects external computing resource providers with enterprise customers, offering access to GPU resources on demand.

Serverless Inference and Model Services: The company provides computing services for model inference to developers and enterprises. Through PBD TokenRouter, it also offers functions such as model access, request routing, and usage management.

The company’s B300 cluster in the Asia-Pacific region uses NVIDIA HGX B300 systems, with system-level optimization covering high-speed networking, storage, workload scheduling, and operational monitoring. The cluster’s NVIDIA Exemplar Cloud recognition relates to its infrastructure performance on specified AI workloads and provides a reference for assessing its large-model training capabilities.

In 2024, PaleBlueDot AI was founded with a focus on supplying AI computing resources and providing dedicated clusters for enterprise customers. The company subsequently developed a business model combining access to third-party GPU resources with customized cluster services to serve AI customers of different sizes.

In January 2026, the company completed a US$150 million Series B financing round led by B Capital, valuing the company at more than US$1 billion. The proceeds were intended to support platform engineering, technical talent, and the global expansion of its core business.

In April 2026, the company launched PBD TokenRouter. Through a unified API, the platform connects to multiple model providers and offers request routing, failover, access management, and usage and cost management. This extended the company’s services to enterprise model access and management.

In July 2026, the company announced the completion of a US$255 million credit refinancing facility structured as a three-year private note. Brookfield Asset Management and Tor Investment Management provided the financing, with JPMorgan acting as placement agent. The funds were intended to enhance delivery capacity and support AI infrastructure development.

In August 2026, the company’s NVIDIA HGX B300 cluster achieved NVIDIA Exemplar Cloud status following performance validation for large-model training workloads. NVIDIA’s official website has listed PaleBlueDot AI among its Exemplar Cloud providers.

In October 2026, the company completed a US$200 million Series C financing round led by ComputeCore, with participation from existing shareholder B Capital and other investors. The round valued the company at more than US$3 billion. The proceeds will be used to increase computing capacity, expand the hardware configurations available to customers, and broaden the company’s presence in the Asia-Pacific region.

PaleBlueDot AI’s dedicated cluster business serves large-scale or specialized computing workloads, providing cluster configurations and reserved computing capacity tailored to customer requirements. Its GPU marketplace connects external supply partners with customers seeking resources on demand, while its serverless inference services support developers and enterprises running AI models.

In cluster engineering, the company’s B300 cluster in the Asia-Pacific region uses NVIDIA HGX B300 systems, with optimization across high-speed networking, storage, and workload scheduling. Exemplar Cloud recognition concerns infrastructure performance for the relevant AI workloads and reflects the performance validation results of a specific cluster.

The launch of TokenRouter, cluster performance validation, and financing activities during 2026 relate to model services, infrastructure engineering, and funding arrangements, respectively. The company is actively developing both its underlying GPU resources and its enterprise AI services.

Asia-Pacific is an important region in which PaleBlueDot AI already conducts business. As of January 2026, the company had an enterprise customer base in Japan, South Korea, and Singapore, with plans to place further emphasis on Asia-Pacific expansion. Its Singapore office, together with its businesses in Japan and South Korea, provides a foundation for connecting Asian customers with regional computing resources.

The company’s Singapore office is located at Marina Bay Financial Centre. PaleBlueDot AI continues to expand its computing services business in the Asia-Pacific region, with offices in Singapore, Tokyo, and Seoul, and has disclosed plans for further expansion into Southeast Asia. Its Singapore presence provides a business foundation for connecting Southeast Asian customers with the wider Asia-Pacific market, coordinating regional resources, and developing computing services partnerships.

Building on its existing presence, the company has a foundation for using Singapore as an important regional base, coordinating global computing resources, and providing high-performance GPU computing, model training, and inference services to Asia-Pacific customers.

In terms of regional business organization, the company’s existing presence provides scope for Singapore to handle commercial activities and customer service while coordinating computing resources in Japan, South Korea, and other regions. Singapore’s role can include bringing together different deployment locations, hardware options, and service models for Asia-Pacific customers, with virtually all computing tasks carried out in Singapore.

Singapore’s AI computing market is closely linked to data center capacity, network connectivity, and energy efficiency. In August 2026, the Singapore Economic Development Board and the Infocomm Media Development Authority announced the provisional allocation of a combined 200 megawatts of additional capacity to four operators under the second Data Centre Call for Application exercise. This figure represents capacity allocated for expansion and additional computing capacity expected to enter operation soon.

As of October 2026, the electricity and water demand associated with AI computing and data centers in Singapore had been incorporated into national infrastructure planning. New capacity is primarily allocated through application exercises, and the second exercise required at least 50% of data center capacity to be powered by green energy sources. Accordingly, the local market focuses on computing service capabilities, resource efficiency, and the conditions required for project delivery.

Singapore is an important business node connecting PaleBlueDot AI with Southeast Asia and the wider Asia-Pacific market. Drawing on its Singapore office and its existing customer and cluster base in Asia-Pacific, the company has the business foundations to use Singapore as a regional service base while coordinating computing resources in Japan, South Korea, and other locations. This provides scope to strengthen regional customer engagement, computing resource coordination, and technical services, supporting the expansion of GPU computing, model training, and inference services across the Asia-Pacific market.