Blaize, Datacomm Forge AI Inference Alliance

Blaize and PT Datacomm Diangraha signed a memorandum of understanding on April 21 to explore deploying artificial intelligence inference systems across Indonesia. (tmcnet.com) Inference is the part of artificial intelligence that runs a trained model on live data — like reading a camera feed or sensor stream and making a decision in real time. Blaize said the alliance will focus first on physical AI, public safety, surveillance, industrial systems and logistics. (tmcnet.com) Blaize is a California AI chip company listed on Nasdaq as BZAI, and it markets programmable hardware and software built for inference workloads at the edge, meaning closer to where data is created instead of in a distant cloud server. The company said its platform is designed to split work across edge devices, data centers and cloud systems. (finance.yahoo.com) (ir.blaize.com) Datacomm brings the local footprint. On its company profile, PT Datacomm Diangraha says it was established in 1990, is based in South Jakarta, and has worked on telecom, military and government digital infrastructure in Indonesia for more than 30 years. (datacomm.co.id) (emis.com) That pairing points to a common problem in artificial intelligence rollouts: many organizations can buy models, but still need local integrators to connect them to cameras, networks, storage and compliance requirements. Datacomm said on its website that it has expanded from networking into cloud and data center services, which are the systems that usually sit underneath production AI deployments. (datacomm.co.id) Indonesia has been trying to build those foundations more systematically. The country released a National AI Roadmap White Paper for 2025 to 2030, with goals around infrastructure, talent, ethics and sector-specific adoption, according to government and policy reports published in 2025. (govinsider.asia) (ps-engage.com) The Blaize-Datacomm agreement is not a disclosed revenue contract, and neither company announced project values, customer names or deployment timelines. It is an MOU, which usually sets a framework for joint exploration before binding purchase orders or implementation deals. (tmcnet.com) Blaize has been stacking similar Asia-Pacific partnerships this month. On April 16, it announced a contract worth up to $50 million with NeoTensr for AI edge data center infrastructure, a sign that it is pursuing regional growth through local partners rather than country-by-country direct sales alone. (ir.blaize.com) What comes next is more concrete than the MOU itself: pilot projects, named customers and installed systems. Until those appear, the deal is best read as an attempt to match a U.S. inference chipmaker with an Indonesian systems integrator that already knows where the racks, cameras and government buyers are. (tmcnet.com) (datacomm.co.id)