On 11 January 2024, Google Cloud announced that it would eliminate the data-transfer charges customers incur when they leave the platform entirely — the “exit egress” fees that have long been one of the most-criticized features of hyperscaler pricing. Google explicitly tied the move to the European Union’s Data Act, the regulation aimed in part at removing barriers to switching cloud providers. The change is narrow in scope but large in symbolism, and it raises an obvious question for everyone watching the cloud market: is this the beginning of the end for transfer-based lock-in, or a one-off gesture that the other major providers will decline to match?

What Egress Fees Are and Why They Matter

Cloud providers generally charge little or nothing to move data into their networks, but charge for data transferred out — egress. The pricing asymmetry is deliberate and has a clear strategic logic: cheap ingress encourages customers to bring their data in, while metered egress makes it costly to take that data elsewhere. For data-intensive workloads, accumulated egress charges can represent a meaningful fraction of total cloud spend.

The lock-in effect is most acute precisely when a customer wants to leave. An organization that has built data pipelines around a provider’s object storage may find that the bill to move petabytes of data out — to a competitor or back on-premises — is large enough to deter the move on its own, independent of any technical migration difficulty. The data is, in effect, held in place by the cost of moving it. Critics have long argued that this turns egress pricing into a barrier to competition rather than a simple cost-recovery mechanism.

What Google Actually Changed

It is important to be precise about the scope of Google’s announcement, because the detail determines how much it actually matters.

The change waives egress charges specifically for customers who are leaving Google Cloud entirely — migrating their data out as part of a full exit from the platform. It is an exit-migration waiver, not a general elimination of egress fees. Customers still pay normal egress charges for ordinary day-to-day data transfer, for multi-cloud architectures that routinely move data between providers, and for serving data to end users. The waiver applies to the specific act of departing.

There is also a process dimension: the waiver is not automatic at the click of a button but involves a request and validation process to confirm that the customer is genuinely migrating off the platform. The principle, though, is the substantive point. Google has conceded — and is the first hyperscaler to concede in this form — that the cost of leaving should not itself be a barrier to leaving.

The Regulatory Driver

Google’s explicit citation of the EU Data Act is the key to understanding the move. The Data Act, which the European Union has adopted and which entered into force in January 2024, includes provisions specifically intended to make switching cloud providers easier — addressing both the technical interoperability barriers and the commercial barriers, of which egress charges are the most visible example. The Act envisions a phased reduction and eventual removal of switching-related charges over its implementation period.

Read in that light, Google’s announcement is partly anticipatory compliance: getting ahead of a regulatory requirement, and earning the goodwill of being first, rather than being compelled later. This is a familiar pattern — regulation reshaping market behavior not only through enforcement but through the incentive it creates for providers to lead rather than follow. The EU’s regulatory reach, given the size of its market, means that a change made to satisfy European rules often becomes a global product change because maintaining separate pricing regimes is impractical.

Will the Others Follow?

This is the question the announcement leaves hanging, and as of early February the honest answer is that we do not yet know. There are reasons to expect the other major providers to face similar pressure, and reasons they might resist.

The pressures pushing toward matching the move:

  • The same regulation applies to all. The EU Data Act’s switching provisions are not specific to Google. Every provider serving the European market faces the same eventual requirement to remove switching barriers, which makes a similar exit-egress concession a plausible compliance path for each of them.
  • Competitive and reputational dynamics. Once one major provider has framed exit-egress charges as a barrier it is willing to remove, the others risk being cast as the ones still holding customers’ data hostage. That framing has competitive cost.
  • Customer expectation. Large enterprise customers, now aware that exit without an egress penalty is possible, will raise it in negotiations regardless of whether the others announce a formal policy.

The reasons for hesitation:

  • Egress is meaningful revenue. For providers with very large data-storage businesses, egress charges are not trivial, and there is no commercial eagerness to forgo them ahead of a hard requirement.
  • Scope games. A provider can satisfy the letter of a switching-cost rule with a narrowly scoped exit waiver — much as Google’s is narrow — while preserving egress revenue on the far larger volume of ordinary transfer. The substance of any follow-on move will be in its scope, not its headline.

What It Means for Cloud Customers

Whatever the other providers decide, the development is useful to infrastructure teams in a concrete way.

  • Re-examine assumed lock-in. If exit-egress costs were a factor deterring a migration off Google Cloud, that calculation has changed. Organizations that shelved a move because the data-transfer bill looked prohibitive should revisit the analysis.
  • Raise it in negotiations everywhere. Even before formal policy changes elsewhere, the existence of Google’s waiver is a negotiating reference point. Large customers can and should ask other providers what exit terms they will offer.
  • Distinguish exit egress from operating egress. The waiver addresses the cost of leaving, not the cost of running a multi-cloud or data-heavy architecture day to day. Lock-in via API and managed-service coupling is untouched by any egress change and remains the deeper portability question.
  • Watch the scope of any follow-on. If and when other providers respond, read the fine print. An exit-migration waiver and a general egress reduction are very different concessions.

The elimination of exit-egress fees by a single hyperscaler does not, on its own, dismantle cloud lock-in — much of which lives in proprietary APIs and managed services that no egress policy touches. But it removes one of the most concrete and most criticized barriers, and it does so under explicit regulatory pressure that applies equally to Google’s competitors. Whether this proves to be the first move in a market-wide shift or an isolated gesture is the question the coming months will answer.

Further Reading

Back to Blog