Introduction

Information economists have typically viewed the cloud computing market as inherently monopolistic: conventional wisdom suggests that economies of scale inevitably lead to market concentration. Our research challenges this perspective by examining an innovative marketplace for cloud services called Akash, which bills itself as "Airbnb for servers." Through empirical analysis of this platform, we uncover evidence that competition may be more viable in cloud computing than previously thought - both between cloud providers and within cloud platforms.

Our findings

Our research yielded several surprising insights that challenge established views about cloud market structure:

Price competitiveness is possible without scale: Despite its small size, Akash offers services at prices competitive with major cloud providers. CPU resources are 86.1% cheaper than median cloud provider prices, RAM is 65.2% cheaper, and even specialized GPU resources used for AI workloads are 13.2-39.2% less expensive. This finding contradicts the assumption that massive scale is necessary to deliver cost-effective cloud services.

Akash is internally competitive: Competitive prices are one thing. But, if a single provider dominates Akash’s marketplace, then all it’s managed to do is reproduce an oligopolistic business. On the contrary, the Akash marketplace demonstrates healthy competition among service providers, with a Herfindahl-Hirschman Index (HHI) of 1,268 - well below the U.S. Department of Justice's 1,800 threshold for market concentration. This finding suggests that cloud services don't inherently tend toward internal monopoly.

Scale economies have a qualitative nature. Different cloud services exhibit varying relationships with scale. While Akash offers competitive pricing for compute and memory resources, storage remains significantly more expensive (99.8% higher than major providers). This cost structure reflects fundamental differences in how scale advantages manifest: while compute workloads can be effectively distributed across heterogeneous hardware, storage systems benefit from concentrated physical infrastructure that enables optimizations in data placement, redundancy, and cold storage technologies. This finding indicates that economies of scale may be more nuanced and service-specific than previously assumed.

Implications for standards

The cloud computing market's tendency toward concentration may stem from technical barriers as much as economies of scale. Internet standards could help create conditions for competition by defining common interfaces for cloud resources—making computational workloads portable across providers, as shipping containers standardized freight transport. The Akash platform demonstrates this potential: its protocols make heterogeneous computing resources appear fungible to developers, enabling a competitive marketplace for cloud services.

Major providers benefit from proprietary interfaces that create switching costs, while smaller providers lack the resources to drive standards development alone. Standards development may need to proceed through bottom-up collaboration among smaller providers or policy frameworks that mandate baseline interoperability requirements. 

Research availability & presentation plans

We are preparing this work for submission to academic journals focusing on Internet economics and market structure. We presented preliminary findings at the Internet Research Task Force Gathering (IETF 124) and plan to share additional results at upcoming academic conferences in 2025.

Next steps

This research opens several promising avenues for further investigation:

1. Implications for Internet standards. Akash demonstrates how protocols can make heterogeneous resources appear fungible—potentially a model for broader standardization. However, realizing this potential requires solving technical challenges (maintaining meaningful guarantees across different architectures) and governance challenges (creating standards despite incumbent resistance). What is the role of standards bodies like the IETF in fostering cloud interoperability? Future work will investigate this question.
2. Market segmentation analysis: Our findings suggest the "cloud market" may comprise multiple distinct markets with different competitive dynamics. Future research should focus on developing empirical methods to identify and analyze these market segments.
3. Regulatory implications: If some cloud services are more amenable to competition than others, this has important implications for antitrust policy. We plan to explore frameworks for identifying which cloud services might benefit from structural remedies versus utility-style regulation.

Several significant challenges remain for future research. First, while our findings suggest competition is possible in certain cloud market segments, we need more granular empirical work to understand where market boundaries are—and why competition can emerge (or fails to emerge) in any given one. Second, the long-term sustainability of alternative business models like Akash's marketplace approach requires further study, particularly their resilience across market cycles and varying demand conditions.

The relationship between market structure and innovation in cloud services also deserves deeper investigation. While our research demonstrates that competition can exist without massive scale, questions remain about how different market structures affect technological advancement, service reliability, and infrastructure investment. These questions become particularly salient as cloud services increasingly underpin critical technological infrastructure.