TL;DR
Pro and Team plans now use one billing metric for data throughput: data synced. This replaces the existing sync operations and data processing metrics. We’re also increasing the overage rate for peak concurrent connections from $15 to $30 per 1,000. Overall, these changes should lower usage costs for everyone.
Pricing challenges
Since launching PowerSync Cloud pricing back in 2023, we've gathered enough data to analyze our customers’ actual usage patterns, associated invoices, and our related infrastructure costs. Three things stood out:
- Sync operation billing spikes weren’t proportional to costs: Customers often had big spikes in billing for sync operations, but our infrastructure costs didn’t spike in the same way. When there were excessive (and unexpected) spikes, we worked with customers directly to offset costs, but this was time-consuming for everyone involved.
- Not all sync operations carry the same amount of data: %%MOVE%% and %%CLEAR%% operations typically contain very little data, but they were being billed the same as other operations. If developers didn’t defragment regularly, the number of billed sync operations could be an order-of-magnitude higher than the number of rows actually synced. Defragmenting is still a good idea, but the pricing impact was not sufficiently warranted.
- Data processing and peak concurrent connections were undervalued: The overage rates for these metrics didn’t fully cover our infrastructure costs
What is changing
Here’s a refresher of how the billing metrics (mentioned above) are defined:
- A sync operation is when a single row is synced from the PowerSync Service to a client.
- Data processing is calculated as the sum of:
- Data replicated from your source database(s) to PowerSync Service instances
- Data synced from PowerSync Service instances to clients.
- Peak concurrent connections is the peak number of client connections active at the same time across all apps in your organization.
Data synced is now the only metric we use to measure data throughput for billing. Sync operations and data processing metrics have been removed.
Additionally, the overage rate for peak concurrent connections will increase from $15 to $30 per 1,000 concurrent connections.
Pro and Team plans are still usage-based: the base fee includes a certain baseline usage allowance, and usage overages beyond that have an additional cost.
Here’s a summary of what’s changing:
Previous model
New model
Our pricing page has been updated, and new customers are already on the new model.
How we reasoned about it
Our goal was to reduce the impact of sync operation spikes on customers, while making sure that our relevant infrastructure costs are covered.
In practice, every sync operation already contributes to data processing, which can be calculated as: the number of operations synced × the average size of a row. So when you sync a row, it contributes to both the number of sync operations and to data processing.
By removing sync operations as a separate billing metric, your usage is now mainly measured in data terms (GB). This is more intuitive for developers to reason about, keeps billing simpler, and ties in better with our actual cost driver — the infrastructure needed to move data.
To make this sustainable, we had to look into increasing the $ rate for data synced (in our current model this forms part of the data processing line item). Historically, our infrastructure costs for replicating data have been lower than for syncing data. Increasing the rate for data processing as a whole meant our customers would pay disproportionately more for replicating data. We therefore decided to drop data replicated as a billing line item, and only bill data synced. We set the data synced rate to $1.00/GB, which is closer to our real infrastructure costs.
Our infrastructure costs for handling concurrent connections are currently over $50 per 1,000 connections. The previous $15 rate was far below that. We’re raising it to $30 — still well under our actual costs — to strike a balance between covering more of the cost and avoiding too steep a jump for existing customers. Over time we hope we can lower costs for this metric by optimizing our infrastructure further.
Impact on existing customers
We looked at each of our existing Pro and Team plan customers’ historic usage and invoices, and simulated their future billing using the new model. The outcome: 99% of customers will be paying less under the new pricing model, assuming similar usage.
We are working with existing customers to migrate them to the new pricing model and are aiming to have everyone migrated by 31 October 2025.
If you have any questions, feedback or concerns about this update, please get in touch.
Overall, we hope this will simplify billing and reduce billing surprises around sync operations.