Summary

• A sustainable digital future that meets the ecological, social, and economic needs of society must be based on humanity-oriented values. To achieve this, it must be built on sustainable digital infrastructure. We aim to create a space for dialogue and informed debate where society, policymakers, and the digital sector can discuss and envision what this digital infrastructure should look like — towards a sustainable, open, and transparent digital infrastructure.

• Society is undergoing a dual transformation towards sustainability and digitalization. While the rapid expansion of the digital sector and cloud computing has brought significant benefits (more connectivity, innovation, and efficiency), the concentration of power within the digital sector in the hands of a few corporations undermines our values.

• While digitalization offers opportunities to address many of the world's pressing problems, the imbalance in control and power within the digital infrastructure undermines a human-centered digitalization and fosters a mentality of "digitalization at all costs." The market is highly concentrated, making it difficult for regional or national competitors in digital infrastructure to gain market share. Dominated by a few oligopolies like AWS, Google Cloud, and Azure, it is practically impossible for smaller regional providers to enter.

• The current cloud landscape is heavily dominated by non-European companies, leading to a dependency on global players. In this way, European companies essentially finance foreign innovation through an indirect "cloud tax" paid by using foreign companies' digital infrastructure. This threatens our ability to embed our values and societal priorities into the infrastructure that drives the digitalization of our economy.

• We promote the concept of regionality, where a region is defined as all cities, municipalities, and communities willing to cooperate on a shared digital infrastructure. This approach ensures the digital infrastructure is tailored to local needs and supports local economic cycles.

How have few digital service companies dominated the digital infrastructure market?

The dominance of Amazon Web Services (AWS), Google Cloud Platform (GCP), and Microsoft Azure is no accident. AWS was launched in 2006 as one of the first major cloud service platforms, allowing them to set industry standards and build a broad customer base. Google Cloud Platform leveraged its expertise in data centers and handling large data volumes. Microsoft Azure built on Microsoft's reputation and expertise in enterprise software and was introduced in 2010.

To offer digital resources on demand (hourly rental of IT equipment), large-scale, long-term investments are required. Digital conglomerates like Microsoft, Amazon, or Google are vertically integrated companies, each having cash-flow-generating core business activities to finance long-term investments in digital infrastructure. For Google, it is advertising, for Microsoft software licenses, and for Amazon an e-commerce marketplace.

This vertical integration allows them to deploy capital to purchase IT equipment in large quantities, reduce procurement costs, and rent it out as an operating expense to customers. This minute-based rental carries significant risk of unused units — a risk other providers cannot absorb without a sufficiently large balance sheet.

These rental agreements are demand-driven and allow customers to rent on an hourly or even minute basis. The risk of unused units arises when rental units are not consumed and thus generate no revenue. Providers must continue to bear the costs of maintaining these units without revenues during idle times. Only providers with significant financial resources can cushion this risk through economies of scale, while smaller providers with less financial flexibility struggle to sustain operations when units remain unused for extended periods.

With their willingness to risk IT equipment and thus enormous amounts of capital, companies like AWS or Microsoft Azure can offer a service (hourly processing, storage, and transfer = digital resources on demand) that competitors can hardly counter. In 2024 it was expected that IT spending would rise by 6.8% to a staggering 5 trillion dollars.

With such market-shaping power, these global digital conglomerates have essentially created markets for on-demand digital resources that are highly profitable. By assuming the capital risk and offering "scalable infrastructure on demand" to everyone, these companies have created a growth engine that requires billions of dollars in capital to compete. A central part of this strategy includes extensive programs that provide startups with free digital resources. These incentives offer significant free credits, making it easier for new companies to build and scale their digital products on the provided infrastructure, thus creating dependency on these services.

Any competitor would need to acquire data center capacity and IT equipment to replicate the established providers' scaling possibilities and then offer it for free to acquire customers. Without an existing cash flow, a competitor finds it impossible to compete on this playing field.

To further accelerate their growth and retain customers, AWS began the practice of "bundling" — combining open-source software with their own digital resources. This strategy binds customers who build their applications with the bundled services. They also began using charges for outbound data transfer (egress fees) to make it more difficult and costlier for customers to migrate their data — and thus their processing and storage capacity consumption — to other cloud providers. Additionally, they offer new customers free credits, creating a form of lock-in. The purpose of the free credits is to attract startups and small businesses, but once the credits are exhausted, customers are often already so deeply integrated into the provider's services that switching to another provider becomes difficult.

Another commonly used tactic to retain customers is sustainable discounts. These ongoing financial incentives make the provider more attractive and prevent customers from switching to competitors. These tactics make it financially challenging for customers to break away from the services of major providers like AWS, GCP, or Microsoft Azure.

Why is this a problem?

There is a widespread misconception that the Internet, essential for digital service companies, is a public good. In reality, the Internet is essentially composed of private networks owned by companies. The same goes for the "public cloud." Although the term "public" suggests a shared, public resource, the IT infrastructure provided by companies like AWS, Microsoft, or Google is privately owned and controlled.

While private investment is good for the economy, public oversight is essential in certain areas such as infrastructure to ensure it serves the broader public good and not the interests of a few corporations. "Governments can by design play a powerful role in creating an economy focused on people and the planet and can often do so more effectively and efficiently than the private sector," argues Nobel laureate Joseph Stiglitz.

However, labeling it "public cloud" strategically creates the illusion of a public good that "can be trusted." In reality, critical IT infrastructure worldwide is consolidated in the hands of a few unregulated, privately-owned US corporations with astounding economic power.

The dominance of large cloud providers like AWS, GCP, and Azure creates significant barriers to entry and stifles competition and innovation. Unlike other markets, the cloud infrastructure market is heavily consolidated and closed off. Independent operators who manage and sell IT equipment, such as web hosting or infrastructure-as-a-service providers, cannot compete due to the large providers' restrictive practices and economies of scale. These advantages include sustainable discounts, significant free credits, and high exit fees designed to bind customers.

Furthermore, to compete in this market, one must reach the scale of the large cloud providers to offset the risk of the hourly rental product. This requires building infrastructure large enough to provide necessary scaling and flexibility while simultaneously generating enough demand to achieve an average minimum utilization rate of 20–30% to break even. Building this infrastructure requires enormous amounts of capital that are essentially speculative in nature, followed by investments in sales and marketing resources to capture demand.

In the case of vertically integrated digital conglomerates, they have an even greater advantage. If the infrastructure is below the required utilization rate, they can offset this by using it for internal services, with other business areas essentially subsidizing the infrastructure business. An example of this is Microsoft, as the entire Office product suite was moved to the Azure cloud. Where companies previously ran their email servers on-premises, they were strongly encouraged to use the cloud-based alternative. These incentives included discontinuing support for local Exchange servers, offering discounts to customers, and introducing subscription-based pricing models.

This closed market design discourages competition and promotes an unhealthy market environment. These major providers spend significant sums on lobbying to prevent regulation that could open up the market. This allows them to erect protective barriers around their services — akin to building a house with walls and a moat. The technology sector is the largest lobbying sector in the EU and spends 113 million euros per year on digital lobbying to prevent further regulation. In the US, Amazon has successfully lobbied against consumer data protection regulations and other legislative measures that would improve market competition.

The need for regional clouds

To counter this, we envision a regulated, open market where anyone who owns or operates data centers with IT equipment can sell digital resources regionally or nationally — ensuring price transparency and fair competition. This vision, a guiding principle, is based on European values and promotes sustainable digital infrastructure. The infrastructure prioritizes practices like using refurbished IT equipment, waste heat recovery, and integrating renewable energy, allowing smaller players to compete based on their environmental performance rather than size alone. Local ownership of infrastructure by companies, communities, or data center operators is not only more sustainable but also ensures that economic benefits — including tax revenue and job creation — remain in the community.

We imagine a transparent market that would reveal any over-concentration of resources and enable governments to adjust regulations as needed to maintain a healthy competitive environment. Currently, the lack of transparency and withholding of critical market information — such as infrastructure utilization rates — by the oligopoly of digital conglomerates hinders regulatory efforts. Transparency is essential to identify flawed market designs and promote an open, democratic process where new and adapted rules can be debated and established based on factual evidence. Without it, the market remains broken and closed, limiting opportunities for innovation and fair competition.

A free and transparent market allows regional and national providers of digital infrastructure to enter and compete. They can offer IT infrastructure and digital services tailored to local needs and utilize regional resources and integrate into existing land and energy infrastructure.

Designing such a market for IT infrastructure and digital resources is becoming technically increasingly feasible through rapidly maturing orchestration technologies like Kubernetes. These technologies enable the unbundling of services and resources and allow the dynamic sourcing of digital resources from various providers based on pre-defined rules. Technologies like orchestration platforms can provide the groundwork to break up the bundling practice of digital conglomerates. They enable customers to use separate services that can be powered by digital resources from any provider — without restrictive lock-ins — thus enhancing competitiveness and supporting the local economy. They promote transparency and cost efficiency, ensuring local businesses do not overpay for unused digital resources.

Examples of regional infrastructure providers with sustainable business practices

Many regional digital infrastructure providers adhere to more sustainable business practices and rely on refurbished equipment and green energy.

For instance, NorthC Datacenters utilizes green hydrogen and 100% green energy and recovers residual heat for surrounding buildings, resulting in significant energy savings. Moreover, their heat recovery systems efficiently repurpose excess heat for heating local facilities.

Another example is Hetzner, which uses renewable energy sources like hydropower in Germany and wind and hydropower in Finland to operate its data centers. The company prioritizes energy-efficient hardware and assembles its own equipment to maximize efficiency.

A regional cloud or IT infrastructure provider is a company supplying digital resources and services that power IT infrastructure and businesses locally or regionally, utilizing regional or local resources and digital infrastructure.

Utilizing the surplus heat from data centers is crucial for a more sustainable digital infrastructure. For example, EcoDataCenter in Sweden uses waste heat for industrial drying of wood pallets.

Finally, OVH, one of the European challengers in competing for global IT infrastructure (but at its core a French company), deploys a proprietary water cooling technology that eliminates the need for air conditioning systems. They design servers for component refurbishment and refurbish buildings for data centers.

Overcoming barriers for regional cloud providers

Several factors hinder the success of regional digital infrastructure providers. These include market design, resource constraints, and anticompetitive practices by global digital conglomerates. Regional providers face significant challenges and often lack the financial means required for upfront investments in facilities or IT equipment. They also lack access to capital markets and debt capital to the extent of large corporations.

However, building on existing local resources and infrastructure can mitigate some of these costs. A single data center operator or IT service company alone cannot provide a regional cloud platform without significant costs. The collaboration of different local actors — such as IT service companies, hardware distributors, and fiber optic network operators — on a profit-sharing model rather than upfront investment could accelerate the launch of regional cloud platforms.

Access to IT equipment remains a bottleneck, as most hardware is not locally produced. Including a regional refurbishment company or an international manufacturer in the cooperation group can provide the necessary refurbished equipment.

Local IT service companies that already enjoy the trust of local businesses can play a key role in operating, monitoring, and supporting a regional cloud platform. These companies can collaborate based on the outlined guiding principle (vision) and form regional markets together.

Global digital conglomerates often offer significant incentives, such as €100,000 (or more) in free digital resources for new customers, making it difficult for local providers to compete, even with better prices and services. Governments can support regional cloud providers by creating regulations and support programs that balance these incentives and target companies in digital transformation or startups. This approach allows companies to choose local providers while still receiving the same free digital resources through the indirect support of regional governments, strengthening the local IT and digital ecosystem. These measures align with the EU's efforts to protect the European market and preserve regional values.

This article is part of our work on the project "Cloud of Amsterdam", financed by the Province of North Holland.