Three BigQuery pricing changes between 2023 and 2025 affect cost modeling and the on-demand vs. Editions decision. Cost calculations based on pre-mid-2023 documentation are likely stale.
Flat-Rate Deprecation (July 2023)
Legacy flat-rate pricing — the predecessor to Editions — ended July 5, 2023. No new flat-rate, monthly, or flex slot purchases are possible. The migration to Editions also included a 25% price increase for on-demand compute, from $5/TiB to $6.25/TiB.
For organizations still on legacy flat-rate annual commitments, the transition happens at renewal: existing commitments convert to Enterprise Edition at renewal with equivalent slot capacity. The slot counts transfer directly, but autoscaling behavior differs from legacy flat-rate. Legacy flat-rate had fixed slot counts; Enterprise Edition introduces baseline/autoscaling configuration choices and idle slot sharing that didn’t exist before.
Pre-2023 pricing benchmarks are stale. The 25% on-demand price increase changes the breakeven calculation in favor of Editions for borderline workloads.
What to do if you’re still on legacy flat-rate: Plan your transition before renewal rather than being forced into it. Use the slot usage monitoring queries to understand your actual consumption pattern, then design an Enterprise Edition configuration that matches your needs rather than accepting a direct conversion. The conversion gives you the same slot count; optimal configuration might look different.
200 TiB Daily On-Demand Quota (September 2025)
As of September 1, 2025, on-demand projects without custom quotas are hard-limited to 200 TiB of query processing daily. Projects exceeding this limit have queries rejected until the quota resets.
For most organizations, 200 TiB/day is more than sufficient — that’s 6 PiB monthly, a substantial workload. But high-volume operations with inefficient queries can hit the ceiling. A workload scanning 10 TiB per ETL run, running 20+ times daily, reaches 200 TiB before end of business.
The quota change has two implications:
High-volume on-demand users need to review their daily query volumes. Use the hour-of-day query pattern analysis to understand your peak daily consumption:
SELECT DATE(creation_time) AS date, ROUND(SUM(total_bytes_billed) / POW(1024, 4), 2) AS tib_processedFROM `region-us`.INFORMATION_SCHEMA.JOBS_BY_PROJECTWHERE creation_time >= TIMESTAMP_SUB(CURRENT_TIMESTAMP(), INTERVAL 30 DAY) AND job_type = 'QUERY' AND state = 'DONE'GROUP BY dateORDER BY date DESC;If you’re regularly approaching 100+ TiB daily, the 200 TiB ceiling is a real constraint. You can request quota increases through Google Cloud support, but the process takes time and isn’t guaranteed.
Editions users are unaffected. Slot-based pricing has no per-TiB limits — the constraint is slot capacity, not data volume. Heavy on-demand users approaching the 200 TiB ceiling have an additional reason to evaluate Editions.
Cloud Storage Fees for External Table Queries (February 2025)
Starting February 2025, Google began charging previously unmetered Cloud Storage fees when querying external BigQuery tables (also called federated queries or BigLake tables). Two new cost components apply:
Storage retrieval fees: Nearline storage ($0.01/GB), Coldline storage ($0.02/GB), and Archive storage ($0.05/GB) now incur retrieval fees when queried. Standard storage remains free to read. If your external tables sit in cold tiers to save storage costs, those same tables are now more expensive to query.
Inter-region network transfer fees: When your BigQuery dataset is in a different region than the Cloud Storage bucket it queries, network transfer fees now apply. Queries that previously looked free now carry data egress costs.
The combined effect: external table queries against archived or cold-tier data may be significantly more expensive than historical patterns suggested. The cost models that drove your decision to store data in Coldline ($0.004/GB/month) versus Standard ($0.020/GB/month) storage need to be recalculated with query retrieval costs included.
What to audit:
- Identify which external tables sit in Nearline/Coldline/Archive storage classes
- Estimate monthly retrieval volumes (bytes read × retrieval fee)
- Check whether your BigQuery region matches your GCS bucket region
- Compare: (lower storage cost) vs. (lower storage cost + new retrieval fees + transfer fees)
For frequently queried data, retrieval fees may eliminate the storage savings. For rarely queried archival data, storage savings likely still outweigh periodic retrieval fees. The crossover depends on read frequency and should be calculated explicitly.
What Hasn’t Changed
Regional on-demand pricing remains $6.25/TiB across US, EU, and most Asia-Pacific regions (at the post-July-2023 price). BigQuery Omni pricing for AWS S3 and Azure Blob Storage queries remains $9.125/TiB. The core mechanics of columnar billing, the 10 MB minimum threshold, caching behavior, and the free tier haven’t changed.
The Editions tier pricing and commitment discount structure also remains stable: Standard at $0.04/slot-hour, Enterprise at $0.06/slot-hour (with 20%/40% commitment discounts), Enterprise Plus at $0.10/slot-hour (with matching discounts).
The BigQuery release notes track ongoing pricing changes. Google typically announces changes with 3–6 months of notice, though the February 2025 external table fees were not widely anticipated.