NEWS
NVIDIA and Palantir deploy sovereign AI across critical supply chain
NVIDIA and Palantir will deploy a sovereign AI stack across
NVIDIA’ s supply chain, combining customised Nemotron open models with Palantir Foundry and its Artificial Intelligence Platform to improve decision-making across one of technology’ s most complex production networks.
The system is designed to codify operational expertise, identify constraints and accelerate the journey from semiconductor wafer to deployed AI infrastructure while allowing NVIDIA to retain control of proprietary data.
NVIDIA’ s supply chain spans millions of parts, thousands of suppliers and a global manufacturing network. Each Vera Rubin rack alone contains 1.3 million parts, requiring coordinated availability of compute, memory, networking, power, cooling and mechanical components.
Under the collaboration, Nemotron models are grounded in Palantir’ s Ontology, while NVIDIA cuOpt software supports optimisation and scenario planning. Supply chain teams can assess constraints, model trade-offs and evaluate how allocation decisions could affect end-to-end production.
The companies said enterprises could also post-train Nemotron models using proprietary operational data through Palantir Foundry and AIP, creating organisation-specific AI while maintaining control over models, data and deployment environments.
Jensen Huang, founder and CEO, NVIDIA, said the companies were transforming the supply chain’ s“ vast operational graph into sovereign intelligence”.
The deployment runs on NVIDIA reference architectures and the Palantir Sovereign AI Operating System Reference Architecture, supported by Dell Technologies and Cisco.
The companies plan to extend lessons from NVIDIA’ s deployment across manufacturing, energy, healthcare, automotive and aerospace, with deployments available on premises, in colocation facilities or through cloud infrastructure. This provides control over where sensitive operational data resides.
Qualcomm and Amazon deepen collaboration on AI data centre silicon
Qualcomm and Amazon have expanded their technology partnership to develop customised silicon and highspeed optical connectivity aimed at supporting the next generation of large-scale AI data centre infrastructure.
The multi-generational collaboration will focus on AI inference, with Qualcomm Technologies working alongside Amazon on customised silicon designed to meet growing demand for compute capacity, memory bandwidth and energy-efficient infrastructure.
The companies will also develop optical connectivity technology supporting speeds of up to 1.6T, alongside future-generation solutions. Qualcomm said its SerDes and optical DSP technologies would help address increasing bandwidth requirements within Amazon data centre networks as AI deployments scale.
The agreement strengthens Qualcomm’ s push into AI data centre infrastructure, extending its power-efficient processing and system-level silicon expertise beyond its established mobile and edge computing businesses.
Qualcomm also plans to increase its use of AWS AI infrastructure, including Amazon Bedrock, for electronic design automation workloads. The company said the move could help shorten chip design cycles by applying AI to engineering processes.
Cristiano Amon, President and CEO, Qualcomm Incorporated, said AI infrastructure would require advances across both computing and connectivity to improve performance and efficiency.
Prasad Kalyanaraman, Vice President, AWS, said the collaboration would combine customised silicon with advanced connectivity to deliver more performant, efficient and cost-effective infrastructure for customers.
The partnership comes as hyperscale operators increasingly pursue specialised silicon, faster networking and more efficient architectures to manage the escalating cost, power consumption and performance demands created by AI workloads.
It also highlights competition around the infrastructure underpinning generative AI.
10
INTELLIGENT CIO NORTH AMERICA www. intelligentcio. com