AWS Builder Center
What Blockchain Snapshots on AWS changes for node operators

What Blockchain Snapshots on AWS changes for node operators

AWS now hosts free Ethereum and Solana node snapshots in Amazon S3. Why sync takes so long, what the new snapshots change, and the limits to know.

Cloud Engineer
On October 14, 2026, AWS announced Blockchain Snapshots on AWS. It publishes pre-synced Ethereum and Solana chain data in Amazon S3, so you can start a node from a snapshot instead of syncing it from the peer-to-peer network.
Snapshots for these chains already existed before this launch. Blockchain foundations and teams publish them. So the useful question is what changes when AWS hosts them.
This post answers that question. It first explains why bootstrapping a node is slow, then covers what the launch provides and where it helps. Everything here comes from the announcement, the project documentation, the public catalog, and the client documentation linked at the end. I did not download a snapshot myself, so there are no timing measurements in this post.

Why a new node takes so long to become useful

A new blockchain node cannot answer requests or validate blocks until it has the current chain state and has checked it. Getting there is called syncing.

Ethereum

The ethereum.org documentation says a full sync, which re-executes every block from genesis, "can take days to weeks." Snap sync is the default on Ethereum mainnet and is much faster, because it starts from a recent checkpoint. The Geth documentation still lists a few catches:
• Downloading the state is the slowest part of snap sync • The healing phase has to outpace the growth of the chain, or the node never catches up • Healing speed depends on the hardware
In practice, the time to a usable node depends on the hardware, among other factors, and it is hard to predict.

Solana

The Agave client builds its starting state by downloading a snapshot from known validators at startup. Anza's operations guide offers a flag, --minimal-snapshot-download-speed, to switch to another validator when the download is too slow. The speed of the source matters here too.

Snapshots were already available

The project documentation states that its snapshot data is mirrored from third-party blockchain foundations and teams. What it changes is the delivery. As the catalog below shows, a single Ethereum mainnet snapshot is around 1 TB. Pulling one from a provider outside AWS means moving that much data over your network link and then unpacking it.


What Blockchain Snapshots on AWS provides

Available snapshots

On October 15, 2026, the us-east-1 catalog listed five snapshots:
• ethereum-mainnet-geth is 1.1 TiB, and ethereum-mainnet-nethermind is 1.0 TiB • ethereum-mainnet-reth is 779.5 GiB • ethereum-hoodi-geth, for the Hoodi testnet, is 89.4 GiB • solana-mainnet-agave is 115.3 GiB
Snapshots refresh on the same schedule as the upstream provider. At the time of writing, the Solana snapshot was 36 minutes old and the Ethereum mainnet snapshots were 9 to 11 hours old.




Requirements

• The supported Regions are US East (N. Virginia), Europe (Ireland), and Asia Pacific (Tokyo) • Downloads work only from an Amazon EC2 instance in the same Region, through an Amazon S3 gateway endpoint in your Amazon VPC. A laptop or a browser cannot reach them • The instance profile needs s3:GetObject and s3:ListBucket • The snapshots cost nothing. Normal charges apply for EC2, Amazon EBS, and other resources you use

How you use it

Everything runs through one script, snapshot.sh. Clone the repository and run setup once:
git clone https://github.com/awslabs/blockchain-snapshots.git
cd blockchain-snapshots/dist
./snapshot.sh --setup
Then pass a snapshot ID and an output directory:
./snapshot.sh --snapshot ethereum-mainnet-geth --out /data/ethereum-mainnet-geth
The script picks the newest catalog entry at run time, so you never choose a block number. For Solana, the files stay compressed. You start Agave with --snapshots pointing at the folder and with --no-snapshot-fetch.

What changes for node operators

The short version: the time to get a node ready becomes shorter, and it should also be easier to estimate. Four details make that happen.

Waiting for sync becomes waiting for a download

A node started from a snapshot begins near the head of the chain and only has to sync the remaining gap. The documentation puts it as "Start from current state, not from genesis."
Sync speed depends on healing and hardware. The download size, on the other hand, is listed in the catalog before you start. Because most of the wait moves to download and extraction, the total time should be easier to estimate.

The download stays inside AWS

The snapshots sit in S3 in the same Region as your instance, and you reach them through a gateway endpoint. AWS documentation confirms three things about that path:
• Traffic reaches S3 without an internet gateway or a NAT device • Gateway endpoints for S3 have no additional charge • S3 does not charge for data transferred to AWS services in the same Region
That means a 1 TB download does not depend on your internet link and does not add S3 transfer charges. The Tokyo Region is on the list, so this also works for nodes running in Japan.

Extraction is built for parallel cores

Downloading 1 TB is only half the work, because the archive still has to be unpacked. AWS republishes the large snapshots in a format designed for that step:
• The seekable format decodes independent zstd frames in parallel • The worker count follows CPU and memory limits, up to 96 • On local NVMe, it writes with O_DIRECT, which bypasses the OS page cache
The Tokyo Region catalog marks the three Ethereum mainnet snapshots as "remint," meaning repackaged. The Hoodi and Solana snapshots are plain mirrors. The parallel format matters most for the large Ethereum mainnet data.

It fits into automation

snapshot.sh is designed to run without a person at the keyboard:
• A stable ID always resolves to the latest snapshot • It checks credentials, bucket access, free space, and dependencies before writing anything • --json prints a machine-readable result • It runs in Docker, Amazon ECS, and Kubernetes, and the documentation shows how to use it with AWS Blockchain Node Runners
The announcement names three use cases: adding capacity to an RPC fleet, recovering a failed node, and re-syncing a corrupted database. Each one asks for the same thing, which is another working node, quickly and the same way every time. One command that always fetches the latest snapshot can go straight into a bootstrap script or an auto-recovery workflow.

Limits to know before you rely on it

It is not an AWS service

The documentation says the project "is not an AWS service or SLA." There is no guarantee of availability, freshness, or correctness. A node restored from a snapshot still validates against its peers, and the documentation calls the network "the final arbiter of correctness."

Location and storage constraints

• Only three Regions are supported, and only EC2 instances in the same Region can download • You need local NVMe or an EBS volume, sized for the extracted snapshot rather than the download • For seekable snapshots, the exact extracted size is in the manifest

No published timing numbers

Neither the announcement nor the documentation says how long a restore takes. To know how much faster it is than syncing, you have to test it with your own instance type and storage.

Who benefits

The snapshot data itself is the same third-party data as before. What changed is how you receive it: from S3 in your Region, with no transfer charge, in a format built for parallel extraction, through a single command that always picks the latest version.
The people who gain the most run Ethereum or Solana nodes on AWS and regularly add, replace, or rebuild them. If you do not run nodes, or you run them outside AWS, this launch changes little for you.

References

Blockchain Snapshots on AWS

• Announcing Blockchain Snapshots on AWS for node bootstrapping:  https://aws.amazon.com/about-aws/whats-new/2026/10/blockchain-snapshots-node-bootstrapping/\  • Blockchain Snapshots on AWS documentation:  https://awslabs.github.io/blockchain-snapshots/\  • awslabs/blockchain-snapshots on GitHub:  https://github.com/awslabs/blockchain-snapshots 

Node sync

AWS pricing and networking


Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.
Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article