- Picking the wrong AWS region can double S3 storage costs
- Inter‑region transfers can cost more than the storage itself
- Hidden premium features add up fast
- AI workloads carry steep mark‑ups compared to general‑purpose S3
- MSPs can protect margins by optimising region/storage class selection, minimising outbound transfers, and auditing advanced feature usage regularly
AWS S3 pricing is famously (well, infamously) confusing. Cloud engineers ask the same questions on repeat:
“Am I paying too much in my region? Would moving storage or changing tiers save costs?”
The answer: probably.
We dug into AWS’s own pricing dataset (~8,000 rows of OnDemand USD data since 2023) and surfaced what really drives your bill. Spoiler Alert: it’s not just where your data lives—it’s how it moves, and what exotic features you switch on without realizing.
All prices below are current as of August 2025.
Regions: Location, Location, Location
Not all AWS regions are priced equally. Store a terabyte in one place, and it’ll cost double what you’d pay elsewhere.
Cheapest regions: ~$0.02/GB/Mo
Most expensive regions: ~$0.04/GB/Mo
That’s a 100% swing just for picking the wrong map pin! Region choice alone can cut your storage bill in half.
Looking beyond regions, AWS’s storage classes differ even more dramatically. Premium options like High Performance and Vectors cost an order of magnitude more than baseline General Purpose, while cold storage tiers like Archive are a fraction of the price. The heatmap below shows how each class prices out across regions—darker blues represent higher cost.
The story is consistent worldwide: the wrong storage class will cost you more than the wrong region.
High Performance and Vectors storage classes likely represent the premium tier, commanding higher prices due to specialised capabilities and faster access speeds.
Archive and Cold storage options (including Archive Instant Retrieval) are positioned as cost-effective solutions for long-term data retention where frequent access is not required.
General Purpose and Intelligent-Tiering appear to offer balanced pricing for standard business workloads, providing good value for regular data access patterns.
When people talk about S3 costs, they usually mean “storage.” And yes, storage is the biggest driver—averaging about 2 cents per GB per month. But AWS’s own pricing data shows there’s a second pillar: fees. These include things like retrieval charges and minimum storage commitments. Finally, requests and data transfer show up as small but meaningful slices. In practice, storage is the headline, but fees and requests are the hidden extras that often catch teams off guard.
Moving your data can cost more than storing it
Storage grabs headlines, but S3’s hidden toll comes when you move data. AWS’s pricing makes a clear statement: getting data into AWS is free or cheap, getting it out is expensive.
Inter-region outbound transfers are the priciest by far—moving data from one AWS region to another can double your costs.
Even intra-region outbound transfers are more expensive than inbound.
Inbound transfer into AWS is essentially subsidized—it’s the outbound flows that sting.
For global apps, transfer can cost more than the storage itself. Architects need to factor not just where data lives, but how often it moves.
InterRegion Outbound is by far the most expensive transfer type (sending data out of a region to another region).
IntraRegion Outbound (within the same region) is also costly, though slightly lower.
Inbound transfers (into AWS, or between regions) are much cheaper. AWS Inbound is close to zero—reinforcing AWS’s strategy: cheap to get data in, expensive to get it out.
Not only is outbound expensive, but where you send it matters—some corridors cost up to 3x more than others. Looking at continent-to-continent flows, the story is clear: inter-region transfers are not equal. Moving data into the Middle East costs the most—averaging $0.07 per GB, compared to $0.02 for local transfers.
Even flows between Europe, Asia, and South America hover around $0.06 per GB. For global workloads, cross-region data movement can add up faster than the storage bill itself.
Not every inter-region transfer comes with a heavy bill. In fact, some AWS corridors are surprisingly cheap—as low as $0.01 per GB. Transfers like Jakarta → London, Oregon → Taipei, or Bahrain → Calgary barely register compared to the $0.07/GB routes into the Middle East.
That’s a 7x difference between the cheapest and most expensive paths. For workloads that replicate data across continents, it means smart region pairing can cut transfer costs to a fraction of what you’d pay on less favorable routes.
Beyond the familiar storage and transfer charges, AWS’s SKU catalog hides a set of premium features that carry price tags far above baseline S3. If you don’t know they’re there, these costs creep into your bill unnoticed.
Batch Operations: $0.25 per job
Batch Operations let you perform large-scale edits across millions of S3 objects. Useful, yes—but expensive. Across regions, Batch Operations clock in at $0.25 per job. That’s 10x higher than the per-GB cost of standard storage. A modest workload of 1,000 jobs a month translates into $250 just in Batch charges, easily eclipsing storage itself.
Vectors: AI storage at a premium
As AWS pivots into AI workloads, it’s attaching a heavy premium. Vector storage (Vectors-Put-Bytes) costs $0.20–0.22 per GB, while Vectors-TimedStorage sits at $0.06–0.07/GB/Mo. Compare that to S3 General Purpose at ~$0.02 per GB-Mo, and you’re looking at a 3–10x markup for storing embeddings or vectorized data. For teams experimenting with retrieval-augmented generation (RAG) or semantic search, these costs add up fast.
Exotic Storage Classes (XZ)
Even within timed storage, certain exotic classes (TimedStorage-XZ-ByteHrs) price at $0.11–0.12 per GB-Mo: about 5x more than S3 Standard. These classes are niche and often provisioned automatically as part of specialized workloads. If you don’t pay attention, you can find yourself paying a 400% premium for what looks like “just another S3 bucket.”
Multi-Region Access Points (MRAP)
Convenience has a cost—and that cost can add up. Multi-Region Access Points, designed to route requests intelligently across AWS regions, carry transfer rates of $0.06–$0.07 per GB. That’s 3–7x more than a standard regional transfer. For globally distributed applications, MRAP can become the hidden tax that doubles your data movement bill.
Final thoughts
AWS isn’t just charging you for what you store. It’s charging extra for where you put it, how you move it, and which features you turn on.
The big cost levers:
- Pick the right region → up to 50% savings.
- Use the right storage class → up to 10x swing.
- Minimize outbound transfers → avoid the silent killer.
- Watch advanced features → Batch Ops, Vectors, MRAP can nuke your budget.
Cloud costs aren’t about waiting for AWS to lower prices—they’re about taking control and playing the AWS pricing game smarter. What measures will you put in place to lower your S3 bill?
Your next read: Could UK competition regulator (CMA) weaken Microsoft and AWS domination of Cloud services in Europe?
