The traditional cloud computing model has long been dominated by hyperscalers like AWS, Microsoft, and Google. However, a new trend is emerging where organizations with excess compute, power, or networking capacity—such as enterprises, AI infrastructure operators, telecoms, and colocation providers—may begin selling access to their resources on a temporary basis. This shift could reshape the cloud market into a dynamic exchange, much like an energy or commodity market, rather than a segmented industry. Below, we explore key questions about this evolving landscape and its implications for buyers, sellers, and the broader ecosystem.

1. What is the emerging trend in cloud computing described in the article?

The article highlights a move away from the assumption that only hyperscalers can deliver elastic infrastructure at scale. Organizations with excess capacity—such as data center operators, telecom companies, or large enterprises—are now exploring the possibility of packaging and selling their unused compute, power, and networking resources. This could create a secondary market where capacity is traded dynamically, similar to how electricity is bought and sold on a grid. The trend suggests that cloud computing may become less about proprietary platforms and more about who has available resources at a given moment, enabling a broader range of players to act like temporary cloud providers. For more on the benefits of this model, see question 3.

2. Why would organizations with excess capacity want to sell compute resources?

Organizations with underutilized infrastructure—such as idle GPUs, partially filled clusters, or stranded power and cooling assets—have a financial incentive to monetize that excess. Instead of letting resources sit idle, they can generate revenue by selling access to third parties. Additionally, selling excess capacity can improve operational efficiency by reducing waste and lowering the total cost of ownership for their own data centers. For some, it may also be a strategic move to diversify into a new service offering, build partnerships, or gain a foothold in the cloud market without the heavy investment of building a fully-fledged cloud platform. However, as discussed in question 5, becoming a capacity seller comes with significant operational challenges.

3. What are the three main benefits of this capacity market model?

The article outlines three key benefits: Price, Efficiency, and Optionality. First, pricing can be significantly lower than hyperscaler rates because non-hyperscale providers have different cost structures and margin expectations. They can offer capacity at a discount, especially for unused resources. Second, efficiency improves by utilizing existing infrastructure rather than building new data centers, saving time, capital, and energy—a direct path to sustainability. Third, optionality gives buyers more choices and leverage. Enterprises can avoid vendor lock-in by sourcing compute from multiple capacity providers, increasing negotiating power and architectural flexibility. These benefits are especially relevant for cost-sensitive workloads like AI training and high-performance computing.

4. How could this model improve cost efficiency for enterprises?

Enterprises under pressure to control AI and infrastructure costs stand to gain the most from a capacity market. Non-hyperscale sellers with excess resources often do not carry the same overhead as major cloud providers—lower marketing costs, fewer support tiers, and no need to maintain a full service catalog. This allows them to offer compute at rates materially below hyperscaler list prices. For bursty workloads, temporary projects, or experimental AI models, companies can scale up without committing to expensive reserved instances or long-term contracts. Moreover, the improved utilization of existing capacity reduces the need for new capital expenditure across the industry, which can indirectly lower costs for everyone. As explored in question 7, this also aligns with sustainability goals.

5. What operational challenges do non-hyperscale providers face when acting as cloud providers?

While owning infrastructure is a start, delivering cloud-like services involves more than just spare capacity. Non-hyperscale providers must develop robust provisioning systems, billing platforms, security frameworks, and support teams. They also need to manage multitenancy effectively to ensure isolation and performance consistency across different customers. Reliability is another concern: a datacenter designed for internal use may not have the redundancy or SLAs expected by external clients. Additionally, legal and compliance issues—such as data residency, privacy, and contract terms—require significant effort. Without solving these operational challenges, capacity sellers risk damaging their reputation or incurring unexpected costs. However, as the market matures, we may see intermediary platforms that handle these complexities, as hinted in the original article.

6. How does this shift affect the current cloud market dynamics and vendor lock-in?

The emergence of a capacity exchange could significantly reduce hyperscaler lock-in. Currently, enterprises moving to the cloud often find themselves tied to a single provider due to proprietary APIs, data transfer fees, and integrated services. A capacity market offers an alternative: businesses can source raw compute, storage, or network capacity from multiple sellers, using open standards or middleware to interoperate. This increases bargaining power and allows companies to mix and match providers based on price and availability. While hyperscalers will still dominate for fully managed services and advanced AI tools, the capacity market creates a floor for pricing and innovation. Buyers gain leverage to negotiate better deals with their primary cloud vendors or to shift non-critical workloads to cheaper alternatives.

7. What role does sustainability play in this new capacity exchange?

Sustainability is a central driver behind the capacity market model. By repurposing existing infrastructure—rather than building new data centers—the industry can reduce energy consumption, e-waste, and carbon emissions. Spare capacity often comes with already committed power and cooling, so using it more efficiently directly lowers the overall environmental footprint. This aligns with corporate ESG goals and regulatory pressures to decarbonize IT operations. Moreover, capacity exchanges can prioritize sellers with green energy sources, creating market incentives for cleaner compute. The article notes that we have long talked about sustainability in cloud; this model offers a practical path by making better use of what already exists. For more on efficiency benefits, see question 3.