Google Cloud has scrapped data transfer fees for customers running workloads in parallel across multiple cloud platforms in the European Union and the United Kingdom, a decision that immediately reshapes the economics of multicloud adoption and fires a competitive shot at rivals Microsoft and AWS. The move, announced in late summer 2025, arrives just weeks before the EU Data Act becomes fully applicable on 12 September 2025, and as UK competition authorities push for stricter rules on egress charges and licensing practices that they say lock businesses into dominant hyperscalers.
Regulatory Backdrop: The EU Data Act and UK Competition Probe
The European Union’s Data Act entered into force in January 2024 and is designed to dismantle barriers to data portability and interoperability between cloud services. From 12 September 2025, providers must comply with new transparency obligations and, eventually, a ban on switching charges above the actual cost of data transfer. During a transitional period, providers can still charge egress fees but only at a documented cost level. Meanwhile, the UK Competition and Markets Authority (CMA) has concluded that the domestic cloud market suffers from high concentration, with Microsoft and AWS controlling the lion’s share, and that egress fees and restrictive licensing terms significantly hinder customers from moving workloads or adopting multicloud strategies.
What Google Actually Promised
According to a blog post attributed to Jeanette Manfra, Google Cloud’s Senior Director for Global Risk and Compliance, the company is offering a product called Data Transfer Essentials. It waives data transfer charges entirely for organizations that process data in parallel across two or more cloud platforms within the EU or UK. This goes beyond the Data Act’s requirement to charge only “at cost,” because Google’s offer provides the service at no charge for qualifying multicloud scenarios. The waiver is positioned as an opt-in solution that customers can activate, and it explicitly targets the friction that has long made multicloud architectures prohibitively expensive for many enterprises.
How Microsoft and AWS Are Responding
The three largest cloud providers now offer starkly different approaches:
- Microsoft: In late August 2025, Microsoft published guidance for EU customers, detailing an at-cost data transfer process. Customers must open a support request, provide technical details (subscription ID, autonomous system number, transfer estimates), and ensure they meet strict eligibility conditions—such as having a billing address in the European Economic Area and using supported internet service provider paths. Approved customers receive credits or refunds, not a blanket waiver.
- AWS: Amazon Web Services takes a case-by-case approach. It states that EU customers can request reduced data transfer rates for eligible use cases but does not offer a generalized free transfer product. Customers must contact AWS support to negotiate reductions, leaving the outcome less predictable.
This divergence gives procurement teams leverage in negotiations but also increases complexity, as each provider’s terms carry unique operational requirements.
The Fine Print: What’s Not Covered
Google’s reported waiver applies only to workloads processed “in parallel” by the same organization across two or more clouds. This excludes several common scenarios:
- Content delivery or public internet distribution.
- Transfers between different accounts or organizations.
- Premium network paths such as private interconnects or specialized transit options that fall outside standard internet peering.
Microsoft’s at-cost documentation similarly restricts eligibility to specific ISP paths and requires the data to remain under the same organization’s control. Both providers make clear that not all data movement is created equal. Network topology, peering arrangements, and regional pricing for ingress can still generate significant charges, even when headline egress fees are waived. Technical validation remains essential before declaring a migration “free.”
Market Impact: Competition, Procurement, and Regulatory Optics
Google’s unilateral no-cost stance is a calculated competitive gambit with several immediate effects:
- Pressure on rivals: Microsoft and AWS now face heightened expectations to match or beat Google’s terms for European customers. Any failure to do so risks losing multicloud deals and drawing regulatory scrutiny.
- Procurement leverage: Enterprise buyers can use Google’s public commitment as a benchmark in contract renewals. Smart negotiators will demand written assurances, including service-level agreements and credit mechanisms, rather than relying on press releases.
- Regulatory optics: The move aligns with the European Commission’s push for easier switching and could influence upcoming model contractual clauses. It strengthens the narrative that hyperscalers must go beyond bare compliance.
Yet the waiver does not eliminate all migration costs. Engineering work, re‑architecting applications, temporary dual-running, and licensing complexities remain substantial line items in any multicloud budget.
Practical Steps for IT Teams
Organizations evaluating multicloud strategies should act now:
- Get written commitments: Press announcements are marketing. Ask for contract amendments or support‑ticket‑confirmed credit processes that outline exact eligibility, timelines, and claim procedures.
- Map traffic patterns: Identify exactly where data originates, how it routes, and which transfer methods (e.g., Google Storage Transfer Service, direct peering) will be used. Different tools carry different cost profiles.
- Test performance and security: Run a pilot transfer to validate throughput, latency, and encryption. A fee waiver does not guarantee adequate speed or recovery from failures.
- Model total cost of ownership: Include labor, project management, temporary dual-cloud running, DNS and identity reconfiguration, and any third-party tools. Egress is only one piece of the migration puzzle.
- Look beyond egress fees: The CMA has highlighted licensing practices—such as higher costs for running Windows Server or SQL Server on non‑Azure clouds—as a core lock-in mechanism. Tackle those in parallel with transfer costs.
Risks and Caveats
Several risks demand attention:
- Policy volatility: Providers can modify commercial programs; Microsoft’s documentation explicitly states that policies may change. Aim for irrevocable contract terms where possible.
- Hidden network charges: Even if list egress fees are zero, underlying ISP transit or cross‑border peering fees may still apply, and it is often unclear which party bears them.
- Limited scope: The same‑organization, parallel‑processing restriction excludes many real‑world architectures. Verify that your use case fits.
- Regulatory uncertainty: The Data Act’s enforcement details and the Commission’s model clauses are still being finalized. The difference between “at cost” and “waived” could become a flashpoint in litigation.
What’s Next for Cloud Providers
Removing egress fees for key markets eliminates a long‑standing reputational liability. For Google, it signals customer‑friendliness and regulatory goodwill. For Microsoft and AWS, the calculus is trickier: match the free transfer offer and lose a revenue stream, or hold the line and risk public criticism. Over time, providers may double down on service stickiness through integrated AI tools, managed databases, and licensing incentives that are harder to disentangle than network charges. The battle over lock-in is simply shifting from the network layer to the platform layer.
Conclusion
Google’s decision to waive multicloud data transfer fees in the EU and UK is a landmark moment in hyperscaler competition, one that gives enterprises concrete, short‑term cost relief and new negotiating power. It aligns squarely with the regulatory goals of the EU Data Act and the CMA’s competition recommendations. Yet no single price change can unravel decades of technical and commercial entanglement with a primary provider. Smart organizations will use the current momentum to secure written commitments, validate real‑world transfer performance, and address the full spectrum of lock‑in—from licensing to proprietary services. The next 18 months will reveal whether this shift leads to durable, lower‑friction portability or whether new forms of vendor lock‑in simply replace the old.