New hub connects North American AI technology and demand with Southeast Asia’s expanding infrastructure ecosystem; company is building capacity across North America, Asia-Pacific, and Europe.
SINGAPORE AND SAN FRANCISCO — October 23, 2026 — TokenPulse, a next-generation AI inference infrastructure company headquartered in San Francisco and Singapore, today announced the opening of its Asia-Pacific Operating Hub in Singapore. The company marked the occasion with an opening ceremony attended by invited industry leaders and ecosystem partners, who gathered to discuss emerging opportunities in AI compute and global infrastructure.
Companies putting AI applications into production face rising inference costs, pressure to keep response times low, and often have to rebuild their platforms as workloads grow. Inference demand is climbing alongside: the global AI infrastructure market is projected to grow from $982 billion in 2025 to $1.98 trillion by 2027, doubling in size over the two-year period.
TokenPulse delivers high-performance, cost-efficient AI compute at global scale. By combining data-center resources, GPU infrastructure, an optimized inference engine, and globally coordinated serving capabilities, the company serves foundation model companies, AI platforms, enterprises, and AI-native applications.
Leadership With Deep Infrastructure Experience
TokenPulse’s leadership and technical team bring experience spanning AI and machine learning infrastructure, cloud computing, networking, data centers and global business operations, including backgrounds at Microsoft, Google and Equinix.
CEO Steven Zhou is a serial entrepreneur with more than 20 years of experience scaling businesses across the United States and Asia. Chairman Thomas Pang brings more than 30 years of experience across government, investment, infrastructure and corporate leadership. He spent more than two decades with Keppel and served as CEO of Keppel’s Data Centres & Networks Division from 2014 to 2024, overseeing the group’s global data-center and digital-connectivity businesses.
“Singapore is a natural choice for building our APAC operating hub, given its mature data center ecosystem, strong network connectivity and deep talent pool,” said Thomas Pang, Chairman of TokenPulse. “Our foundations are rooted in decades of experience across data centers, networking and cloud computing. Drawing on that experience in building and operating infrastructure at scale, we will continue investing in compute capacity and ecosystem support, delivering the best performance at the lowest cost per token.”
Singapore as a Strategic Hub for a Global Infrastructure Network
TokenPulse is building its infrastructure footprint across North America, Asia-Pacific and Europe, with Singapore serving as its strategic center for APAC. The company’s regional strategy is designed to connect North America’s AI technology, customer and capital ecosystems with the expanding data-center and energy resources of Southeast Asia.
From Singapore, TokenPulse is positioned to work with partners across Malaysia, Indonesia, Thailand and the wider region. The hub will support regional operations, infrastructure partnerships and coordination across the company’s global inference network, helping bring compute capacity closer to customers and AI workloads.
TokenPulse is taking a long-term approach to building an integrated AI infrastructure network, bringing together data-center, energy and compute resources to support growing AI inference demand across its global markets. The company is advancing multiple data-center projects with infrastructure partners across key markets, positioning itself for sustained capacity growth. Its infrastructure roadmap includes deployment of the latest-generation NVIDIA AI infrastructure, while its engineering focus remains on improving how effectively compute resources are converted into tokens and useful AI output.
“AI is undergoing a fundamental shift: compute demand is moving from training to inference. For businesses, inference is no longer a hidden cost sitting in the background. It is becoming a core competitive factor in determining whether AI products can successfully reach the market,” said Steven Zhou, CEO of TokenPulse.
“As one of the fastest-growing regions for AI applications, developers in Asia Pacific are direct about cost and performance. By making Singapore our co-headquarters and APAC operating hub, we are putting our compute and our team closest to them to deliver better performance at lower cost.”
Flexible Consumption and Deployment Options
TokenPulse offers three consumption models on a unified technology stack, allowing customers to start with smaller workloads and expand their compute capacity as their applications grow.
- Serverless — Pay per Token: Provides unified API access for rapid deployment, model evaluation and variable workloads. It enables developers and AI-native applications to get started without committing to dedicated GPU capacity upfront.
- On-Demand — Pay per GPU-Hour: Offers dedicated endpoints and predictable performance for customers with elastic compute requirements, supporting workload optimization as usage grows.
- Reserved — Contracted Capacity: Provides dedicated, contracted compute capacity for sustained workloads and more customized infrastructure requirements.
TokenPulse offers three deployment options to accommodate different infrastructure, security and operational requirements:
- TokenPulse Cloud. Managed GPU infrastructure for a streamlined path to production.
- Private VPC. A TokenPulse-hosted environment with a dedicated network boundary and workload isolation.
- BYOC (Bring Your Own Cloud). The TokenPulse inference engine can be deployed in a customer’s cloud or VPC, or on the customer’s own GPUs.
The company’s commercial approach is built around a simple principle: “Start small. Scale without replatforming.” Customers can begin by validating their applications through API-based access, then transition to GPU-based services or reserved capacity as demand increases. This lets businesses expand their AI workloads without having to switch providers or rebuild their platforms as their infrastructure needs change.
Optimizing Inference for Performance and Cost
TokenPulse’s competitive advantage is the combination of high-standard data center infrastructure and deep inference optimization tailored to specific industry use cases. By optimizing open-weight models for vertical workloads, TokenPulse improves compute efficiency and serving throughput so customers can process more tokens at lower cost.
Key technology capabilities include:
- Higher cache efficiency: TokenPulse uses HiCache and a multi-tier key-value (KV) caching architecture to reduce repeated computation. Its observed cluster-level cache hit rate can reach 95%.
- Faster response times: With context parallelism enabled, TokenPulse has achieved a 50% reduction in 90th-percentile time to first token (P90 TTFT) and a 75% reduction in 99th-percentile TTFT (P99 TTFT) compared with the same workload without context parallelism enabled.
- Global intelligent routing: A unified routing architecture directs requests across serving regions based on factors including customer requirements, model, region and time window. Nearest-edge access, session-context reuse and redundant network paths are designed to support low latency, high availability and efficient compute utilization.
Together, these capabilities are designed to help AI businesses improve inference economics while maintaining the responsiveness and reliability required for production workloads. Actual performance gains vary by workload and operating conditions.
About TokenPulse
TokenPulse is a next-generation AI inference infrastructure company headquartered in San Francisco and Singapore. The company combines latest-generation GPU infrastructure, an optimized inference engine, and global AI infrastructure capabilities to deliver high-performance, cost-efficient compute for foundation model companies, AI platforms, enterprises, and AI-native applications worldwide.
With flexible Serverless, On-Demand, and Reserved consumption models, TokenPulse enables customers to start with API-based inference and scale to dedicated GPU capacity as their workloads grow — without replatforming.
Learn more: www.tokenpulse.com
