Nokia and AI chip designer Blaize are moving from early-stage collaboration to deeper cooperation, jointly addressing what they see as a critical bottleneck in the increasingly expanding AI ecosystem: scalable and cost-effective inference capabilities.

Following the signing of a preliminary Memorandum of Understanding (MoU) earlier this year, the two companies are currently developing a joint reference architecture to streamline hybrid AI deployment in edge, cloud, and data center environments across the Asia-Pacific (APAC) region.

The collaboration aims to address three major challenges currently faced by enterprises and service providers: ensuring end-to-end performance, reducing deployment complexity, and cutting AI infrastructure costs. Both parties will leverage Nokia’s innovation lab in Singapore to develop, test, validate, and deploy AI solutions for customers.

Dion Leung, Head of AI and Cloud Business for Nokia’s Asia-Pacific Region, stated, “Clients today are struggling to integrate AI chips, software frameworks, networks, security, and lifecycle management. Through this joint collaboration, we can deliver a pre-integrated and validated technology stack that seamlessly coordinates AI, hardware, computing, hardware inference, networks, and automation capabilities.”

Leung stated: “AI workloads on the edge often cannot be supported solely by GPU cost-effectiveness or power budget. Through collaboration, we can deliver a high-efficiency AI platform while enabling optimal workload deployment between the far edge and near edge, helping to reduce both capital and operational expenditures.”

This initiative comes at a time when market demand is shifting from model training to inference, with organizations eager to monetize their AI investments. Joseph Sulistyo, Senior Vice President of Global Marketing at Blaize, stated, “We are currently witnessing a clear transition in demand from training to inference, with reports from cloud service providers indicating a 15% to 50% increase in inference-related workloads. As customers seek to move from experimental stages to practical deployment, organizations are asking how to truly deploy applications and derive value from their AI investments.”

Sulistyo pointed out that this is driving the industry to focus on outcome-oriented solutions that can quickly transition from pilot to production, and the collaboration between Nokia and Blaize aims to fill this gap.

While developed Asia-Pacific economies such as Singapore, Japan, and Australia are advancing toward cloud-native and edge architectures, rapidly growing markets like India, Indonesia, and Vietnam are still exploring AI deployment models and monetization strategies. Nokia and Blaize believe that within the next two to three years, the market will transition from experimental stages to developing domain-specific, outcome-driven solutions.

Nokia’s collaboration with Blaize in the Asia-Pacific region complements its agreement with chip giant NVIDIA last year. In October 2025, NVIDIA secured a 2.9% stake in Nokia by investing $1 billion to accelerate the development of AI-RAN (Radio Access Network).

Nokia and Blaize are facing fierce competition in the region from multiple players, including NVIDIA and Qualcomm in the computing sector, as well as Huawei and Ericsson collaborating with Intel.