Data gravity is the principle that as a dataset grows, the cost and latency of moving it rise, so it becomes cheaper to bring compute and services to the data than to move the data to them. The cost consequence in a multicloud estate is concrete: every time a workload in one cloud reads from a dataset in another, you pay egress on the source side, and every time you copy the dataset to avoid that, you pay for duplicated storage and the pipelines that keep copies in sync. The buyer takeaway is that placement is an economic decision, not a technical afterthought, and the cheapest architecture is usually the one that keeps heavy compute next to the data it reads most.
Below is how data gravity shows up on the bill, the three levers that control it, and a worked example you can act on this week.
How does data gravity show up on the cloud bill?
It rarely appears as a line called data gravity. It hides inside three categories. The first is egress: AWS, Azure, and GCP all meter data leaving a region or the cloud, typically in the range of a few cents per gigabyte that looks trivial until a daily analytics job moves terabytes. OCI is materially cheaper on egress, which is itself a placement signal. The second is duplication: teams copy a dataset into a second cloud to keep compute local, then pay twice for storage plus the transfer to seed and refresh the copy. The third is the architecture tax: services chosen because they sit next to the data, not because they are the best fit, which quietly raises the run rate for years.
Why does moving compute to the data usually win?
Compute is portable and cheap to relocate; petabytes are neither. If an analytics workload needs to read a 40 terabyte dataset every day, running that workload in the same region as the data costs nothing in transfer, while running it one cloud over could move 40 terabytes a day across a metered boundary. At a representative egress rate, that single design choice can swing five figures a month. The discipline is to identify the authoritative dataset, treat its location as fixed, and place the compute that reads it most heavily right beside it. Lighter consumers that need only summaries can live elsewhere and pull aggregates rather than raw data.
What does deliberate placement look like?
Three levers control data gravity cost. Place the authoritative copy where the heaviest readers live and treat other clouds as consumers of derived data, not raw. Push aggregation to the source so you transfer summaries measured in gigabytes rather than raw events measured in terabytes. And price egress into every cross cloud design before it ships, because the cheapest time to avoid a transfer charge is at the whiteboard, not after the first bill. Where a second copy is genuinely required for resilience or locality, size it against the recovery and latency benefit it buys rather than copying by reflex.
A worked example
A European SaaS company ran its event store in one cloud and its analytics stack in another because two teams had chosen independently. A daily pipeline moved roughly 30 terabytes across the boundary, and the cross cloud transfer alone ran into the tens of thousands of dollars a month. Moving the analytics compute into the same region as the event store, and shipping only daily aggregates to the second cloud for the business intelligence layer, cut the transferred volume by more than 90 percent and removed the duplicated raw copy entirely. The change touched placement and pipelines, not the application, so engineering risk stayed low. Figures are verified against billing data and anonymised.
| Pattern | What moves | Cost driver |
|---|---|---|
| Data to compute | Raw dataset, daily | High egress plus duplicated storage |
| Compute to data | Nothing across the boundary | Region local, no transfer charge |
| Aggregate then ship | Summaries only | Small transfer, single source of truth |
Talk it through with us
If your multicloud bill is heavier than the architecture should justify, data gravity is often the reason, and placement is the fix. We take zero provider commissions and answer only to you, across AWS, Azure, GCP, and OCI. Our guarantee is plain: we reduce your cloud spend or we reimburse our service fee, on either a Fixed Fee scoped up front or a no risk Gainshare share of verified savings. Book a strategy call to scope it for your estate, and follow more analysis in The Cloud Spend Navigator.
Frequently asked questions
What is data gravity in cloud computing?
How does data gravity increase cloud costs?
How do you reduce the cost of data gravity?
Put a defensible number on your cloud spend.
No provider in the room, no published price list. Tell us your footprint and we will scope the savings against your billing data — we reduce your cloud spend or we reimburse our service fee.
The Cloud Spend Navigator: what changed in cloud pricing, commitments, and FinOps — no vendor spin.