xAI becomes a neocloud: what changes for B2B SMBs

The history of a silent pivot

xAI was founded in 2023 with a stated goal: to develop advanced artificial intelligence as an alternative to OpenAI. In a short time, it launched Grok, its own language model integrated into the X platform. However, the company's trajectory has undergone an unexpected acceleration on the infrastructure front.

During 2025, xAI began the construction of Colossus , one of the largest GPU clusters in the world, located in Memphis, Tennessee. The facility quickly reached a capacity of approximately 100,000 Nvidia H100 GPUs. Furthermore, expansion plans have been announced that would bring the total to over 200,000 units in the short term.

According to reports by TechCrunch in May 2026 , xAI's real business might be precisely this: selling computational capacity to third parties, positioning itself as a neocloud. Therefore, the Grok model becomes almost a showcase product, while the data centers represent the structural revenue source.

Neocloud: a term worth understanding

The term neocloud indicates cloud infrastructure providers born specifically for AI-intensive workloads. They differ from traditional hyperscalers in hardware specialization, provisioning speed, and GPU consumption-oriented pricing.

Among the established neoclouds are CoreWeave, Lambda Labs, and Together AI. These operators have captured significant market share precisely because AWS, Azure, and Google Cloud struggle to meet GPU demand quickly. Therefore, xAI's entry into this segment is not an isolated move: it fits into a structural trend already documented by Gartner in its cloud computing forecasts.

Specifically, Gartner estimates that by 2027, over 30% of enterprise cloud spending will go towards specialized AI workloads. Consequently, the demand for GPU-first infrastructure is set to grow steadily. xAI is positioned to capture this demand with a physical asset already built.

Winners and losers of this scenario

Every market pivot produces redistributions of value. In this case, the subjects involved are multiple.

Who gains ground:

• SMEs with scalable AI needs : a new infrastructure operator means more competition and, tendentially, lower prices for compute.
• Integration partners : agencies and system integrators that know how to build solutions on heterogeneous infrastructures gain consulting value.
• Nvidia : any expansion of AI data centers translates into GPU orders. The Californian manufacturer's position remains dominant.

Who risks falling behind:

• Second-tier neoclouds : smaller operators, lacking a recognizable brand, struggle to compete with xAI's media visibility.
• Hyperscalers in specific segments : AWS and Azure maintain the ecosystem advantage, but on pure GPU compute they could face price pressure.
• Pure-play AI companies without their own infrastructure : relying on third-party providers for compute becomes a strategic risk in an increasingly vertically integrated market.

Furthermore, it's worth considering the impact on xAI's brand perception. Indeed, transforming into an infrastructure provider changes its competitive positioning: no longer just a rival to OpenAI, but a direct competitor to Microsoft Azure and Google Cloud in a specific segment.

Reading SHM Studio: infrastructure as a strategic lever

We at SHM Studio We interpret xAI's pivot as a sign of AI market maturity, not just industry news. When a company founded to research models decides to invest billions in concrete and silicon, it means the competition is shifting to infrastructure.

This has a direct consequence for Italian B2B SMEs. Until recently, accessing AI-grade computational capacity required enterprise contracts with large hyperscalers, often inaccessible due to budget and bureaucratic complexity. Today, the landscape is different. Therefore, even a medium-sized manufacturing company or a specialized retailer can evaluate scalable AI solutions without necessarily relying solely on dominant players.

However, the multiplication of providers introduces new complexities in choice. Not all neoclouds offer the same guarantees of uptime, data security, and GDPR compliance. For this reason, selecting AI infrastructure requires a structured evaluation, not an improvised decision based on price.

Our AI services include precisely this type of strategic orientation: analysis of computational needs, selection of the most suitable provider, and integration with existing business processes. Similarly, our activities of Digital marketing rely more and more on AI infrastructures to optimize campaigns and content in real-time.

The still open worksite: risks and unknowns

It would be incorrect to present xAI's pivot as a consolidated certainty. There are open variables that deserve attention.

First off, governance. xAI is controlled by Elon Musk, a figure known for rapid and unpredictable changes of direction. Relying on a cloud infrastructure tied to a single founder introduces a non-negligible concentration risk. In fact, the events of Twitter/X have shown how quickly strategic priorities can change.

Furthermore, the energy issue is relevant. AI data centers consume enormous amounts of electricity. The Memphis cluster has already raised local concerns related to environmental impact and electricity grid availability. As documented by Wired in its deep dive into AI data centers, energy sustainability has become a critical factor for the scalability of these infrastructures.

Finally, the European regulatory landscape remains a hurdle. Any Italian SME wanting to rely on xAI infrastructure would need to verify compliance with GDPR and the European AI Act, which has now fully come into effect. Therefore, the technical assessment cannot be separated from the legal one.

Next moves: what Italian SMEs should do now

The AI cloud market is being reshaped. Italian B2B SMEs that want to seize the opportunities of this moment must act methodically, not urgently.

First, it's helpful to map your current and potential AI workloads. How many computational resources are already being used? Which processes could benefit from generative or predictive AI? This mapping is the starting point for any infrastructure assessment. Our activities SEO and AI-assisted copywriting , for example, are already based on a careful selection of the most suitable computational tools for the context.

Secondly, it's wise to diversify providers. Relying on a single hyperscaler or neocloud introduces lock-in risks. A multi-cloud strategy, even for SMEs, offers greater operational resilience. Therefore, monitoring xAI's evolution as a potential provider makes sense, but not as the exclusive choice.

Thirdly, it's worth investing in internal evaluation skills. You don't need a team of cloud engineers. It's enough to have someone, internal or external, capable of reading service contracts, evaluating SLAs, and comparing real costs. Our web solutions and the google ads campaigns integrated with AI require exactly this type of applied expertise.

Finally, for those who wish to delve deeper into the implications of the AI Act for their sector, the European AI regulatory framework published by the European Commission is essential reading. Likewise, the analyses of Harvard Business Review on AI and machine learning offer strategic perspectives useful for the management of SMEs.

For those who would like to discuss these topics with us, our team is available through the contact page . Furthermore, on our Blog we regularly publish updated analyses on the evolution of the AI and cloud market for the Italian context. The LinkedIn campaigns remain, in this scenario, a privileged tool for B2B SMEs that want to position themselves as credible players in the AI market.