The Antitrust Case Hidden in Your Cloud Contract

Contracts for cloud computing don’t seem very dramatic. Procurement teams frequently sign them late at night in an effort to meet deadlines. They arrive as thick PDFs that are packed with rate tables and compliance language. However, regulators across three continents are starting to view those documents more as possible evidence than as standard service agreements. The suspicion is straightforward: the fine print may be engaged in quiet competition.

For the past two years, the Federal Trade Commission has been collecting data in Washington regarding minimum-spend commitments, data-transfer fees, and licensing bundles that consumers claim make switching providers exorbitantly costly. It’s possible that no one clause alone seems coercive. Critics counter that when combined, they produce gravity. “Building a house where the doors technically open but the cost of walking out keeps rising” is how one startup founder put it.

After a two-year investigation, the Competition and Markets Authority in Britain came to the harsh conclusion that cloud users hardly ever switch providers. Annually, less than 1 percent do so. Executives in London’s financial district discuss “multi-cloud” strategies as though they are already a reality, despite the fact that banks operate risk models in windowless server rooms miles away. In reality, a lot of them are still closely linked to a single platform, and their data architectures are designed in a way that makes moving them difficult, expensive, and dangerous.

According to the CMA, Google is far behind Amazon Web Services and Microsoft Azure, which together account for between 30 and 40 percent of cloud spending in the UK. The dominance is reinforced, according to regulators, by licensing restrictions, exit fees, and incompatible technical standards. The businesses disagree, claiming that the market is competitive and dynamic. As this argument develops, it seems possible that both assertions—that innovation happens quickly but switching is still uncommon—are accurate.

Data egress fees, or the price of removing data from a cloud environment, are at the heart of the dispute. These fees, according to providers, represent actual network investments. Critics liken them to telecom contract termination fees, which are intended more to deter departure than to cover costs. Even when presented as technical requirements, switching costs frequently serve as competitive moats, according to market structure economists.

Another layer is added by long-term spending commitments. Customers who commit to spending large amounts of money over a number of years are rewarded by discount programs. Those savings appear to be wise financial decisions for multinational firms. They may appear to be covert exclusive-dealing incentives for startups with erratic growth, pushing businesses closer to a single supplier.

Software licensing comes next. British regulators have drawn attention to how some enterprise software contracts raise prices when clients use competing clouds for workloads. Microsoft calls the market competitive and contests the description. Notably, Google has embraced criticism, portraying the incident as a watershed for equitable pricing. This industry’s alliances and rivalries change swiftly, demonstrating how corporate narratives can be rearranged by regulatory pressure.

The stakes have increased due to artificial intelligence. Large model scaling for AI startups necessitates massive processing power, which is frequently found in hyperscale clouds. As usage grows, some founders report contractual complexity and pricing opacity. At the same time, Microsoft denies claims that its cloud partnership with OpenAI limited compute supply and raised prices in a lawsuit filed in the United States. It’s still unclear if courts will consider these kinds of alliances to be a contemporary form of exclusivity or pro-competitive cooperation.

Echoes can be found in history. For similar reasons, regulators previously examined software bundling practices and telecom termination fees: they increased the cost of leaving. Similar concerns are raised by modern cloud contracts, albeit with less obvious infrastructure. The battlefield is data portability and interoperability rather than cables and operating systems.

Industry experts contend that complexity is not a purposeful trap but rather a feature of cloud architecture. Programs that waive transfer fees for clients moving away have been implemented by certain providers. But the economics are still there. Regardless of policy changes, it is costly to move petabytes of data, reconfigure systems, and retrain employees.

Last year, it was difficult to ignore how tangible the digital economy still is when you were standing in a server room outside of Frankfurt, with rows of blinking LEDs extending into the distance. Cooling ducts, steel racks, and humming airflow. It feels like the infrastructure is immovable. The contractual inertia that regulators are now questioning is reflected in that physicality.

Competition authorities are considering remedies such as portability standards, transparency requirements, and restrictions on specific fees. Heavy intervention, according to some academics, may backfire and force providers to implement new switching costs. Markets change. Contracts also do this.

Cloud agreements, which were previously regarded as technical procurement details, are undoubtedly evolving into tools for market power analysis. Businesses may continue to emphasize service tiers and price breaks when signing new contracts. Regulators are now openly posing the question, however, “How expensive should it be to leave?” somewhere in between the exit clauses and data-transfer tables.