Weekly Intelligence Brief on Chiplets and Silicon - Week of August 18, 2025

Key Takeaways

• Deca Technologies collaborates with IBM on semiconductor production in Bromont.
• NVIDIA introduces the DGX Cloud Lepton platform for Artificial Intelligence (AI) development.
• Astera Labs enhances NVLink solutions through partnership with NVIDIA.
• Sify Technologies launches Pay-per-use model for AI cloud services in India.
• Fujitsu awarded contract to design the next-generation supercomputer FugakuNEXT.

Recent developments in the semiconductor and AI infrastructure sectors highlight strategic partnerships and new service offerings. Deca Technologies and IBM joined efforts to deploy M-Series technologies at IBM's Bromont facility in Quebec. This collaboration aims to enhance high-volume semiconductor manufacturing capabilities, particularly with the M-Series Fan-out Interposer Technology (MFIT).

NVIDIA introduced the DGX Cloud Lepton, a platform providing Graphics Processing Unit (GPU) access through a network of cloud partners. This initiative responds to rising AI demands, allowing developers to utilize NVIDIA GPUs seamlessly for AI applications. Key partners include CoreWeave and SoftBank, focusing on expanding GPU availability for various projects.

Astera Labs partnered with NVIDIA to incorporate NVLink solutions into its Intelligent Connectivity Platform. This partnership supports demands for high-bandwidth, low-latency connectivity in AI infrastructure, benefiting hyperscalers deploying NVIDIA GPUs. The collaboration enhances overall performance for next-generation AI workloads.

Sify Technologies announced a Pay-per-use model for its AI cloud services in India. This model offers an hourly pricing structure aimed at reducing entry costs for users, thus facilitating easier access to NVIDIA-certified data centers. This move aligns with efforts to meet increasing demand for AI infrastructure in the region.

Fujitsu secured a contract from RIKEN to develop the FugakuNEXT supercomputer, which will advance High performance computing (HPC) resources in Japan. This project includes design work on Central Processing Unit (CPU) components and aims to enhance AI research capabilities. The new FUJITSU-MONAKA CPU will leverage advanced technology for improved performance and efficiency.

1. Deca Technologies collaborates with IBM for semiconductor production in Bromont
   Deca Technologies has partnered with IBM for M-Series technologies at the Bromont facility.
2. NVIDIA launches DGX Cloud Lepton for AI development
   Key cloud partners, including SoftBank, supply NVIDIA GPUs on the new DGX Cloud Lepton platform for AI projects.
3. Astera Labs partners with NVIDIA to enhance NVLink solutions
   Astera Labs collaborates with NVIDIA to enhance its Intelligent Connectivity Platform with NVLink, targeting AI infrastructure needs.
4. Sify introduces Pay-per-use model for AI cloud services in India
   Sify Technologies introduces a Pay-per-use model for AI Cloud Services to enhance access to infrastructure at NVIDIA-certified centers.
5. Sify Technologies launches AI Pay-Per-Use model in India
   Accelsius joins ARPA-E's COOLERCHIPS project to improve energy efficiency through hybrid cooling technologies in data centers.
6. Synopsys announces PCI Express (PCIe) 6.x interoperability with Broadcom's PEX90000 switch at PCI-SIG DevCon 2025.
   Synopsys and Broadcom achieved PCIe 6.x interoperability, enhancing AI infrastructure scalability and reducing design risk at PCI-SIG DevCon 2025.
7. Illumio and NVIDIA enhance cybersecurity integration
   Illumio and NVIDIA enhance cyber resilience for critical infrastructure through platform integration.
8. NXP Semiconductors and Rimac Technology develop vehicle architecture for Synthetic Data Validation Suite (SDVS)
   NXP Semiconductors and Rimac Technology created a centralized vehicle architecture that combines over 20 ECUs into three units.
9. Fujitsu awarded contract to design next-generation supercomputer FugakuNEXT
   Fujitsu Limited has secured a contract from RIKEN to design FugakuNEXT, a next-generation supercomputer aimed at enhancing AI research.