The fast adoption of generative AI and agentic AI is spurring demand for AI infrastructure and the necessity for a wealthy ecosystem that may ship the entire AI lifecycle. On this setting, a brand new breed of AI cloud suppliers—also known as neoclouds—has emerged as agile, technically refined gamers constructing extremely specialised AI cloud providers to serve new markets and wishes. At Cisco, we’re devoted to working with the complete vary of ecosystem contributors. This consists of the normal hyperscalers, in addition to these newer neocloud suppliers and foundational mannequin builders. It additionally consists of communications service suppliers, colocation suppliers, and sovereign cloud suppliers that increase AI cloud providers by delivering the AI internet hosting, safe ubiquitous connectivity, and trusted infrastructure and information providers enterprises want.
The AI infrastructure market: The surging frontier
The demand for AI infrastructure is rising at an exponential price. Whole international information middle IT gear spend over the 5 years between 2025 and 2030 is anticipated to succeed in as a lot as 4.7 trillion {dollars}.1 This encompasses servers, accelerators (XPUs), networking, optics, and storage. At present, it’s estimated that hyperscalers are answerable for over 60% of this infrastructure funding, whereas neoclouds are answerable for about 17%, which is anticipated to develop to over 30% over the following ten years.2 The opposite ecosystem members and enterprises themselves have gotten answerable for an rising proportion of the AI infrastructure buildout as inferencing and agentic AI, sovereign cloud, and edge AI grow to be extra mainstream. In accordance with McKinsey, demand for AI information middle capability will account for 70% of complete information middle demand by 2030 and is the primary driver of a rising infrastructure deficit.3 Neoclouds are quickly stepping as much as fill the hole and are making large strikes to drive new enterprise fashions and know-how improvements that target the readiness, resilience, differentiation, and profitability of their AI cloud service choices.
Three approaches to AI cloud providers
Neocloud suppliers will not be a monolith. They make the most of numerous enterprise fashions to ship AI infrastructure providers that cater to completely different use instances and verticals. Consumption fashions are altering quickly, from bare-metal choices to full-stack {hardware} and software program along with “as a service” choices. We are able to categorize supply and consumption fashions into a number of approaches that mirror numerous enterprise necessities for scalability, flexibility, price, efficiency, and information sovereignty. Particular person suppliers are selecting the approaches that greatest go well with their enterprise priorities and buyer wants.
Three main approaches have emerged:
Devoted cloud AI IaaS: On this mannequin, enterprises decide to utilizing devoted AI infrastructure inside a neocloud’s information middle for a hard and fast time period. The advantages of this method embrace assured capability, higher efficiency isolation, and infrequently important price financial savings in comparison with on-demand pricing. It’s greatest fitted to predictable and long-running AI workloads, similar to large-scale mannequin coaching or steady inference duties, and for organizations with constant AI useful resource necessities or people who want unique entry to compute assets.
Public AI cloud providers: With this mannequin, neoclouds allow enterprise clients to entry shared swimming pools of AI-optimized computing assets, similar to GPUs and tensor processing models (TPUs). Prospects solely pay for the assets they devour with out long-term commitments. This “pay-as-you-go” method gives most flexibility and scalability, making it preferrred for experimental workloads, growth, testing, or bursty AI duties the place demand just isn’t constant.
Hybrid and edge AI IaaS: This mannequin extends deploying and managing AI infrastructure throughout a mixture of public cloud environments, on-premises or colocation information facilities, and edge places. It brings AI compute nearer to the place information is generated or consumed, addressing wants like low-latency processing, information sovereignty, and regulatory compliance. This mannequin is especially helpful for purchasers that require real-time resolution making, similar to industrial web of issues (IoT), autonomous automobiles, or localized analytics.
Seven causes neoclouds can depend on Cisco for AI infrastructure
Whatever the service mannequin adopted, neocloud suppliers require a strong, scalable, and safe basis. Many are discovering that Cisco gives the partnering focus, complete portfolio, and imaginative and prescient essential to succeed. Listed here are seven key benefits to partnering with Cisco:
Complete, in-house AI infrastructure portfolio: Cisco gives the broadest homegrown AI infrastructure portfolio each inside and between information facilities. This consists of built-in full-stack safe AI manufacturing facility options, AI-optimized compute, AI networking, optics, and information middle interconnect options, in addition to safety and observability for AI. Advantages for neoclouds embrace architectural flexibility, streamlined procurement, constant operations and help, and a extremely sturdy provide chain.
Architectural and operational flexibility: With Cisco, neocloud suppliers have the pliability to decide on the reference structure and working setting that greatest meets their technical necessities. Whether or not they need to adhere to the NVIDIA Enterprise Reference Structure (ERA), the NVIDIA Cloud Accomplice (NCP) reference structure, Cisco Cloud Reference Structure, or just need to construct their very own—we offer the mandatory flexibility and validated designs to speed up deployment.
Business-leading AI networking materials: Cisco AI Networking delivers scalable front-end and back-end materials inside information facilities, along with information middle interconnect and scale-across networks between information facilities. Neoclouds can select between the confirmed reliability of Cisco NX-OS for operational consistency, or SONiC for open networking requirements.
Expanded partnership with NVIDIA: Our expanded partnership with NVIDIA allows NVIDIA Spectrum-X know-how integration on Cisco Silicon One–primarily based switches and Cisco N9100 Collection Switches that incorporate NVIDIA Spectrum-X Ethernet swap silicon—all unified below the Nexus working mannequin for seamless integration with Cisco Nexus Dashboard. Nexus is already well-known inside many neoclouds who’ve established it of their front-end computing networks.
Highly effective and numerous companion ecosystem: Cisco companions with a broad ecosystem to ship the perfect options and worth. This consists of compute and accelerator companions like NVIDIA, AMD, and Intel; storage companions; software program companions; liquid cooling companions; system integrators; and plenty of others.
Finish-to-end safety and observability: Safety and reliability are prime considerations for enterprises deploying AI. Cisco Safe AI Manufacturing facility with NVIDIA supplies a full-stack resolution with safety built-in at each layer. With Splunk, Cisco supplies Observability for AI for deep visibility into each layer of the setting, together with GPU infrastructure, GenAI, LLM outputs, and AI purposes. Cisco gives unmatched safety and observability for AI deployments, serving to neoclouds defend their infrastructure and information.
Unified silicon structure with Cisco Silicon One: Cisco Silicon One is a groundbreaking switching and routing silicon structure. This single structure is the muse used throughout a broad vary of networks, from back-end and front-end AI networks to scale throughout, information middle interconnect, wide-area networks, and edge. It delivers superior congestion management and cargo balancing for AI materials and is programmable to help rising requirements.
Constructing the way forward for AI cloud providers with Cisco
Cisco is uniquely positioned to help neocloud suppliers’ AI infrastructure wants throughout all their service fashions. By offering cutting-edge know-how—from Cisco Silicon One and NVIDIA Spectrum-X silicon-based switching to Cisco UCS compute to built-in safety and observability stacks—Cisco allows neoclouds to ship scalable AI providers with pace, reliability, and effectivity.
Because the market continues to evolve, Cisco supplies the experience, go-to-market help, and options to assist neoclouds speed up time to service, maximize the effectivity of GPU clusters, and in the end safe a foundational position within the burgeoning AI ecosystem.
1. The price of compute: A $7 trillion race to scale information facilities, McKinsey & Firm, April 2025.
2. AI Infrastructure Survey, McKinsey, June 2025. (Inner Cisco survey)
3. AI energy: Increasing information middle capability to satisfy rising demand, McKinsey & Firm, October 2024.
