Juniper Networks launches AI-native networking platform

Juniper Networks has launched a solution specifically designed for GPUaaS and AIaaS providers to enhance AI deployment and streamline operations.

The new AI-Native Networking Platform from Juniper Networks serves the needs of neocloud and traditional service providers deploying GPU-as-a-Service and AI-as-a-Service. This platform aims to accelerate time-to-market while simplifying service management and reducing operating costs, with a focus on optimising data centre networks for multi-tenant environments.

With enterprises increasingly relying on on-demand AI services, providers need to develop infrastructure that enhances performance and makes these services more accessible and competitively priced. The trend also addresses data sovereignty and privacy needs, as organisations with regulatory stipulations demand tailored solutions. Juniper's newly introduced solution aims to address these demands with high-performing, scalable, and secure networks efficiently.

Juniper's AI-optimised solution encompasses QFX Series Switches, PTX Series Routers, and SRX Series Firewalls, alongside management via Juniper Apstra data centre assurance software and Mist AI. The solution is designed to enhance the deployment speed and management ease for AI services such as training and inference using Retrieval Augmented Generation (RAG).

The platform offers key benefits such as streamlined deployment through automation, which reduces costs by up to 85% and decreases deployment times by up to ten times. With the integration of Red Hat OpenShift, further automation benefits are available in Kubernetes environments.

Security is another focal point, with Juniper delivering Zero Trust Architecture and multi-tenancy capabilities to safeguard infrastructure and sensitive data. The SRX 4700 firewall combines efficiency with high-throughput, offering wire speed MACsec support to protect data in motion.

Juniper also ensures providers can confidently deploy AI solutions through validated designs and testing in the Juniper Ops4AI Lab. With support for multi-vendor environments, including NVIDIA and AMD accelerated computing, the platform promotes flexibility and avoids vendor lock-in.

Juniper Networks is marking a significant presence in the AI data centre market with its solutions, commanding almost half of the 800G Ethernet switch market. As AI service demand increases, their AI data centre solutions are gaining traction among major AI data centre customers globally.

Praveen Jain, Senior Vice President and General Manager for Data Center & AI at Juniper Networks, stated, "Managed AI services, such as GPUaaS and AIaaS, have been growing rapidly. To capture the demand, neocloud, traditional SPs and other AI cloud providers need to move fast and deliver exceptional value. Juniper is excited to deliver a purpose-built solution for these providers, expediting their journey towards offering unique and compelling services with the performance, security, automation and open flexibility their customers demand."

Bratin Saha, Chief Product and Technology Officer at Digital Ocean, remarked, "Enterprises of all sizes are aggressively looking to use recent advancements in AI to transform their businesses, but they are getting slowed down by the complexity of managing infrastructure. AI cloud infrastructure services from Digital Ocean speed up time-to-market, lower cost of ownership, and remove the complexity of deploying AI. Scalable, reliable and operationally simple networking from Juniper for front-end and back-end AI networks helps us deliver high fidelity AI services to our customers."

Jeff Hinkle, CEO of ionstream.ai, highlighted, "Selecting the right GPU infrastructure and deciding whether to build, buy or rent it can significantly impact a company's success in the rapidly evolving field of AI and ML. Juniper provides us with easy-to-deploy, secure, AI-optimized networking for our multi-tenant AI infrastructure that for us is on par with traditional InfiniBand but has a far greater human resource pool to support it since it's Ethernet. It is also far easier to reconfigure for rapidly changing AI requirements in multi tenancy so that we can deliver high-performance GPU as a Service quickly however our customers choose to consume it."

David Gucker, COO of Vultr, added, "As enterprises look to expand AI investments in 2025, they want high-performance, secure and cost-effective AI cloud infrastructure from us. Vultr's collaboration with Juniper Networks empowers these enterprises and AI innovators to harness the full potential of accelerated computing. Using Juniper networking solutions, we can deliver AI services to our customers quickly, securely and scale to their highest performance expectations."