The unprecedented pace of urbanization across the Asia-Pacific region is creating profound challenges for the management of cities and large towns, with local governments turning to artificial intelligence (AI) to scale operations and drive performance. AI infrastructure, in particular, is rapidly gaining traction as city administrators integrate traditional systems with AI-powered intelligence, enabling automated, data-driven decisions across multiple domains. Managing AI infrastructure however, can be daunting, given the massive volumes of data that need to be analyzed every second, and the number of real-time decisions that must be supported in every instance.
A hybrid approach to managing AI workloads in smart cities
In response to these challenges, new innovative solutions have begun to emerge. In a notable development, Blaize, a leading player in programmable, energy-efficient edge AI computing, recently announced its collaboration with Starshine Computing Power Technology Limited (Starshine), an AI infrastructure provider specializing in high-performance computing, to enable a new class of hybrid AI computing clusters. The announcement comes at a critical juncture as performance demands continue to push for smarter and more efficient infrastructure for AI-driven, real-time automated decisions, especially at the edge.
The collaboration integrates Blaize’s cutting-edge GSP®-based inference accelerators into Starshine’s high-performance computing platform, creating a first-of-its-kind hybrid AI solution to support advanced insights-driven use cases. The partnership, valued at an estimated USD120 million over a tenure of 18 months, is set to transform APAC’s edge AI deployments, specifically smart city implementations, across key markets, including India, Indonesia, Japan, South Korea, and China. Headquartered in Hong Kong, Starshine is ideally positioned to drive AI infrastructure expansion across Greater China and key APAC markets, combining regional reach with robust computing networks. The partnership is also a strategic move toward establishing sovereign-grade AI infrastructure for customers in the region.
“This is the beginning of real-world AI infrastructure at scale,” said Dinakar Munagala, co-founder and CEO of Blaize. “Asia represents a $112B opportunity for next generation intelligent systems, and inference is leading the charge. Our hybrid AI platform is designed to deliver the efficiency, flexibility, power, and support for multimodal workloads to meet this surge, especially at the edge. Together with Starshine, we’re providing sovereign-grade infrastructure where programmable, energy-efficient inference becomes the backbone of smart cities, industrial environment, and public services.” Blaize has already established a string of partnerships around the region, including its collaboration with Yotta in India and TCC in Saudi Arabia.
‘Practical AI’ and how it drives real-world edge deployments
The Blaize-Starshine collaboration will address a long-standing challenge in smart city deployments: the need to backhaul massive data from thousands of sensors to the cloud. This approach strains bandwidth, increases latency, and raises both performance and security concerns. Blaize’s hybrid AI solution, powered by its GSP® (Graph Streaming Processor) architecture optimized for AI workloads, enables inferencing to occur directly at the edge, while data training using traditional GPUs is retained in the cloud. This approach underpins the idea of practical AI, one of Blaize’s key propositions, according to Blaize co-founder and CEO, Dinakar Munagala.
Designed for low power consumption (<5W) and minimal latency, Blaize’s hybrid AI solution integrates 5G connectivity and provides enhanced software capabilities including configurable analytics, rules engines, and real-time alerting. It also offers secure handling of data, deploying role-based access, and regional controls for enhanced security. In terms of hosting, the platform can be implemented in Micro Points of Presence (MicroPoPs) within data centers, complete with peering and interconnects, or in containerized MicroPoPs within municipal or private data centers to ensure sovereignty and scalable expansion. It can also be deployed on-premises in city control rooms. Beyond the edge, customers have the option to run Blaize AI workloads in the cloud or through hyperscalers, combining metro-level proximity with global reach.
From cameras to command centers
According to Blaize, the partnership will initially concentrate on smart city deployments, with a particular focus on urban surveillance. This includes managing thousands of analog and IP cameras alongside a network of communication and control centers. In these deployments, Blaize’s hybrid AI platform provides an enhanced intelligence layer at the edge, leveraging AI-driven capabilities such as real-time event classification, predictive analytics, and anomaly detection.
For example, Blaize’s platform enables license plate recognition, dynamic traffic policy enforcement, traffic flow optimization, perimeter and public space security enforcement, and unified command center management. Blaize’s superior analytics further support critical decision-making. For instance, its vehicle detection capability achieves 95% accuracy, underscoring the platform’s precision and reliability.
“Hybrid AI is no longer an experiment; it’s infrastructure,” said Teng Ma, Chairman of Starshine. “Blaize delivers exactly what our customers demand: flexibility, power efficiency, multimodal support, and scalability. This partnership gives us the technology and capability to meet surging demand across Asia for real-time, localized intelligence.”
A ‘Platform’ approach that comes with high-programmability
As part of its Practical AI philosophy, Blaize offers an integrated platform that combines hardware, software, and a suite of video libraries into a single adaptable solution. Its high programmability allows customers to tailor the platform to any industry or vertical, enabling hundreds of data-driven use cases powered by real-time AI insights.
Blaize and Starshine have also announced that they will collaborate on software development. This is expected to expedite the adoption of Blaize AI’s solutions in other sectors in the region. This includes retail, manufacturing, logistics, energy management and agriculture. For example, Blaize’s AI insights at the edge can enhance real-time transaction monitoring in retail, while AI-powered predictive maintenance in industrial automation can minimize downtime. Apart from its high programmability, the platform’s rugged form factor is also expected to push its adoption into new use cases. Blaize’s servers are engineered for extreme thermal tolerance, withstanding temperatures of 70–75°C, as well as dust and other environmental challenges.
With the APAC smart city market projected to surge sixfold from USD 223 billion in 2023 to USD 1,342.2 billion by 2030 [1], Blaize and Starshine are positioning their hybrid AI platform to power real-world applications that deliver real impact. “Together, Blaize and Starshine are focused on building an AI future that is trustworthy, intelligent, and sustainable,” said Matt Ma, CEO of Starshine Computing.
---
Autor(en)/Author(s): Tara Neal
Dieser Artikel ist neu veröffentlicht von / This article is republished from: The Fast Mode, 10.11.2025
Bitte besuchen Sie/Please visit:
