OCP Global Summit: AI Scaling, Power, Cooling, Networking Trends

Cisco is significantly expanding partnership with NVIDIA to power high-performance, flexible, and secure AI data centers. This is about simplifying operations and scaling AI from the enterprise to the cloud. Key Innovations Announced: 🔸 Introducing the Cisco N9100 Series Switches: These new switches incorporate NVIDIA Spectrum-X Ethernet switch silicon, giving customers more choice in building powerful AI fabrics. 🔸 Unified AI Operations: Both Cisco Silicon One-based switches and the new N9100 Series are unified under the Cisco Nexus operating model and managed via Nexus Dashboard. This provides a single, consistent approach to managing any-scale AI environment. 🔸 Choice of OS: Customers can choose between Cisco NX-OS for enterprise reliability or SONiC for open networking standards. 🔸 Accelerated Deployment: Cisco is launching reference architectures like Nexus Hyperfabric AI (now generally available) and a new NCP-compliant RA with the N9100 Series to help customers deploy full-stack AI clusters with ease. This is a major step forward for the Cisco Secure AI Factory, ensuring compute, storage, and security work together seamlessly. #AI #DataCenter #CiscoNexus #NVIDIA #Networking #GenAI Read the full blog post for more details! 👇

Architecture built for Sovereign, Neocloud and Service Providers, Cisco and NVIDIA's collaboration continues to accelerate, driven by a shared vision of an AI-powered future that is scalable, observable, and secure. The advancements announced today are a testament to Cisco's relentless pursuit of innovation in purpose-built Architectures that will accelerate AI adoption across mission-critical industries. #CiscoAI #AI_ReadyDC #Neocloud #NVIDIA #Nexus #GTC25

Cisco unveiled new data centre networking gear built with NVIDIA technology, aiming to capture growing demand for infrastructure that supports artificial intelligence (AI) workloads across cloud and telecom networks. The networking company launched the Cisco N9100 series switch, developed with Nvidia’s Spectrum-X Ethernet technology, as part of a new reference architecture for “#neocloud” and sovereign cloud customers. The networking giant said the design complies with Nvidia’s cloud partner standards and allows operators to run either its own NX-OS or the open-source SONiC system. The announcement, made at Nvidia’s GTC event in Washington, shows how hardware makers are competing to provide the backbone for #AI data centres, which require higher bandwidth and lower latency than conventional cloud environments. “We’re at the beginning of the largest data centre build-out in history,” said Jeetu Patel, Cisco’s president and chief product officer. He said the new systems were designed to handle the power and performance demands of emerging AI models. Read full report: https://lnkd.in/d7AUwG_H

++ Cisco and NVIDIA Unveil Breakthrough Innovations to Power the Next Generation of AI Infrastructure ++ Cisco has unveiled a series of major innovations designed to accelerate the global deployment of secure, scalable artificial intelligence (AI) infrastructure. The announcements, made in collaboration with NVIDIA and industry partners, mark a significant step forward in redefining data center, enterprise, and telecom architectures for the AI era. At the center of the announcement is the Cisco N9100, the first NVIDIA partner-developed data center switch based on NVIDIA Spectrum-X Ethernet switch silicon. The switch will serve as the foundation for NVIDIA Cloud Partner-compliant reference architectures, enabling both neocloud and sovereign cloud deployments. This advancement positions Cisco as a key enabler of high-performance, open, and energy-efficient networking for AI workloads. The Cisco Secure AI Factory with NVIDIA introduces enhanced protection and visibility for enterprise AI deployments. Through new security and observability integrations — including Cisco AI Defense with NVIDIA NeMo Guardrails and Splunk Observability Cloud — enterprises can now monitor AI workloads in real time, protect sensitive data, and ensure compliance across distributed environments. 🔗 Read the full article: https://lnkd.in/dMcs8aK4 #TechAfricaNews #Cisco #NVIDIA #AIInfrastructure #DataCentreInnovation #SecureAI #AIForTelecom #EnterpriseAI

NVIDIA is redefining the backbone of AI infrastructure. Meta and Oracle are now integrating NVIDIA’s Spectrum-X Ethernet switches into their AI data center networks a major leap toward building high-performance, scalable Ethernet-based AI systems. This move signals a growing shift in the industry: hyperscalers are no longer relying on proprietary interconnects alone. Instead, they’re embracing Ethernet as a powerful, flexible standard capable of supporting giga-scale AI workloads. It also underscores the importance of vendor diversity and open innovation in shaping the next generation of AI factories where efficiency, scalability, and interoperability are key to sustained AI growth. https://lnkd.in/gwQG_XhD #NVIDIA #Meta #Oracle #AIInfrastructure #GigaScale #Ethernet #AIFactories #DataCenters #AIConnectivity

Intel Shifts Focus to Datacenter and AI Chips as Global Supply Pressures Mount Intel Corporation is reportedly reallocating production resources to prioritize datacenter and enterprise technology chips, as ongoing global semiconductor shortages continue to disrupt supply chains. According to Network World, the company’s strategy now places stronger emphasis on server, cloud computing, and AI workloads—sectors deemed central to Intel’s long-term growth and profitability. Pivot Toward High-Demand Enterprise Segments #AI #DataCenter #Intel

Intel Shifts Focus to Datacenter and AI Chips as Global Supply Pressures Mount Intel Corporation is reportedly reallocating production resources to prioritize datacenter and enterprise technology chips, as ongoing global semiconductor shortages continue to disrupt supply chains. According to Network World, the company’s strategy now places stronger emphasis on server, cloud computing, and AI workloads—sectors deemed central to Intel’s long-term growth and profitability. Pivot Toward High-Demand Enterprise Segments #AI #DataCenter #Intel

🚀 AI is redefining what “modern infrastructure” means — and the network is at the center of it. Excited to share how Cisco and NVIDIA are partnering to accelerate this shift with new innovations in AI networking. Key highlights: • New Cisco Nexus 9100 Series Switches – built on NVIDIA Spectrum-X Silicon • Cisco Cloud Reference Architecture – with Silicon One, Nexus Dashboard, and Nexus HyperFabric for scalable AI clusters. • NVIDIA Cloud Partner (NCP) Compliant - Reference Architecture – leveraging Nexus 9100 for neoclouds and sovereign cloud deployments. • The orderability in November 2025 of Cisco Nexus Hyperfabric AI Pls read more in my blog here: https://lnkd.in/ge3ZDZgG #Networking #Cisco #NVIDIA #Nexus #NexusHyperfabric #GTC25 . . . Will Eatherton Kevin Wollenweber Jeff Schultz Tom Gillis Ambika Kapur Scott Miles Monica Vimala Prashanth Sunil Jay Becky Faraz Usha Swetha Sai Gerald Ashley Christina

𝗕𝗿𝗼𝗮𝗱𝗰𝗼𝗺 𝗟𝗮𝘂𝗻𝗰𝗵𝗲𝘀 “𝗧𝗵𝗼𝗿 𝗨𝗹𝘁𝗿𝗮” 𝗡𝗲𝘁𝘄𝗼𝗿𝗸𝗶𝗻𝗴 𝗖𝗵𝗶𝗽 𝘁𝗼 𝗥𝗶𝘃𝗮𝗹 𝗡𝘃𝗶𝗱𝗶𝗮 𝗶𝗻 𝗔𝗜 𝗗𝗮𝘁𝗮 𝗖𝗲𝗻𝘁𝗿𝗲𝘀 Broadcom has introduced the Thor Ultra networking chip aimed at large-scale AI computing systems. The chip is designed to enable servers to be connected at scale for huge AI workloads. Broadcom projects the AI-chip market could reach $60-90 billion by 2027. Question: In your high-performance server clusters, how much emphasis do you place on networking silicon as opposed to compute alone? #𝗦𝗲𝗿𝘃𝗲𝗿𝗡𝗲𝘁𝘄𝗼𝗿𝗸𝗶𝗻𝗴 #𝗔𝗜𝗜𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 #𝗛𝗮𝗿𝗱𝘄𝗮𝗿𝗲𝗘𝗰𝗼𝘀𝘆𝘀𝘁𝗲𝗺 #𝗗𝗮𝘁𝗮𝗖𝗲𝗻𝘁𝗲𝗿 https://lnkd.in/djPHB58y

#HostingJournalist #AI NVIDIA is expanding its footprint in the AI data center networking space as Meta and Oracle adopt its Spectrum-X Ethernet switches, marking a significant step in the race to build high-performance, giga-scale AI infrastructure. The two hyperscalers are standardizing on NVIDIA’s accelerated Ethernet architecture to enhance the efficiency of AI training, reduce latency, and accelerate the deployment of large-scale models. The collaboration underscores the growing demand for networking systems capable of supporting trillion-parameter models, which require enormous computing and data throughput capabilities. NVIDIA’s founder and CEO Jensen Huang described Spectrum-X as a foundational technology for this new era of large-scale AI. “Trillion-parameter models are transforming data centers into giga-scale AI factories,” said Jensen Huang. “Spectrum-X is not just faster Ethernet - it’s the nervous system of the AI factory, enabling hyperscalers to connect millions of GPUs into a single giant computer to train the largest models ever built.” For Oracle, the deployment of Spectrum-X Ethernet is part of a broader strategy to expand its AI infrastructure. The company plans to build new giga-scale AI factories powered by NVIDIA’s Vera Rubin architecture, interconnected by Spectrum-X to improve data movement efficiency and scalability. According to Mahesh Thiagarajan, executive vice president of Oracle Cloud Infrastructure (OCI), the move strengthens Oracle’s position in high-performance cloud computing. “OCI is designed from the ground up for AI workloads, and our partnership with NVIDIA extends that AI leadership,” said Thiagarajan. “By adopting Spectrum-X Ethernet, we can interconnect millions of GPUs with breakthrough efficiency so our customers can more quickly train, deploy, and benefit from the next wave of generative and reasoning AI.” Meta, meanwhile, is integrating NVIDIA’s Ethernet switches into its Facebook Open Switching System (FBOSS) - the company’s proprietary open networking software used to manage switches across its massive global infrastructure. The goal is to extend Meta’s open networking approach while improving the predictability and efficiency of large-scale AI training workloads. “Meta’s next-generation AI infrastructure requires open and efficient networking at a scale the industry has never seen before,” said Gaya Nagarajan, vice president of networking engineering at Meta. “By integrating NVIDIA Spectrum Ethernet into the Minipack3N switch and FBOSS, we can extend our open networking approach while unlocking the efficiency and predictability needed to train ever-larger models and bring generative AI applications to billions of people.” Spectrum-X Switches and SuperNICs NVIDIA’s Spectrum-X Ethernet platform - which includes both the Spectrum-X switches and SuperNICs - is the first Ethernet solution purpose-built for AI workloads. It enables hyperscalers…