BigQuery storage costs are the fees charged for storing data in Google Cloud's BigQuery data warehouse, calculated per terabyte-month of data stored. These costs apply regardless of whether data is queried, and are billed separately from query processing charges.
BigQuery stores data in a columnar format optimized for analytical queries. You pay for the amount of data stored, measured in terabytes per month. Storage costs start accruing as soon as data is loaded into a dataset, even if it remains unused. Google charges $0.02 per GB per month for standard storage, which translates to $20 per TB per month. This rate applies to data stored in the default regional location.
For example, if you store 500 GB of customer transaction data in BigQuery for one full month, your storage cost is 500 GB × $0.02/GB/month = $10. This fee continues until you delete the data or move it to a lower-cost storage tier. Unlike traditional databases, BigQuery does not charge for read operations against stored data-only for the processing of queries.
BigQuery calculates storage costs based on the actual size of data in its internal storage format. Data is compressed automatically, so the physical storage footprint is typically smaller than the original file size. For example, a 10 GB CSV file may compress to 3 GB in BigQuery, reducing storage costs accordingly.
Google uses a monthly billing cycle. At the end of each calendar month, BigQuery totals the amount of data stored across all datasets in your project and multiplies it by the current rate. You can view this breakdown in the Google Cloud Console under the Billing section, specifically in the BigQuery tab.
Storage costs are independent of query activity. A dataset containing 10 TB of data incurs $200 in monthly storage fees, even if no queries are run. Conversely, running 10,000 queries on a 100 GB dataset adds only query processing charges-not additional storage fees.
BigQuery offers multiple storage tiers, each with different pricing and access characteristics. The primary tiers are standard storage and coldline storage.
| Storage Tier | Price per GB/month | Access Latency | Best For |
|---|---|---|---|
| Standard | $0.02 | Milliseconds | Frequent access, analytics, reporting |
| Coldline | $0.01 | Minutes | Archival, infrequently accessed data |
Move data to Coldline storage to reduce costs by 50% if you don't need immediate access. However, retrieving data from Coldline takes longer-typically minutes rather than milliseconds-and you cannot run queries directly on Coldline-stored data. To query, you must first copy the data to standard storage.
When using partitioned tables, storage costs scale with the number of partitions and their sizes. For example, a daily partitioned table with 365 partitions, each holding 10 GB of data, results in 3.65 TB of total storage. That's $73 in monthly storage fees at $20/TB/month. Partitioning doesn't reduce storage cost-it organizes data for efficient querying-but it does affect how much you're charged.
Many users assume that deleting data stops storage charges immediately. This is incorrect. Charges continue until the data is physically removed from the system. Even after deletion, Google retains metadata for up to 30 days during which storage fees may still apply. Always verify deletion through the Cloud Console or API.
Another misconception is that storage costs rise linearly with data volume. While they do increase proportionally, compression reduces the effective cost. A 100 GB JSON file may compress to 25 GB in BigQuery, lowering storage fees by 75%. Compression rates vary by data type-text-heavy data compresses better than binary or encrypted data.
A non-obvious fact: BigQuery charges for storage based on the size of the data after schema enforcement and type conversion. If you load a CSV with mixed data types and BigQuery converts strings to timestamps or integers, the resulting schema may require more space than the original file. This can lead to unexpected increases in storage costs.
An e-commerce company loads 200 GB of daily sales data into BigQuery. Over a 30-day month, this totals 6 TB of stored data. At $20 per TB per month, the storage cost is $120.
The team runs 500 queries per week to analyze customer behavior. Each query processes 100 GB of data. BigQuery charges $5 per TB processed for standard queries. Over four weeks, 500 queries × 4 weeks = 2,000 queries. Assuming each processes 100 GB, total query processing is 200 TB. At $5/TB, query costs are $1,000.
Total monthly cost: $120 (storage) + $1,000 (queries) = $1,120. By enabling query caching and using partition filters, the team reduces query volume by 40%, saving $400 in processing fees. They also move older data to Coldline storage, cutting storage costs by $60.
Use partitioned and clustered tables to reduce query processing costs. Clustered tables improve performance by organizing data by frequently filtered columns-such as date or region-reducing the amount of data scanned per query. This lowers both query time and processing charges.
Set automatic deletion policies for temporary or historical data. Use BigQuery's time-based expiration feature to auto-delete tables after 90 days. This prevents long-term accumulation of unused data that inflates storage bills.
Monitor usage via the BigQuery Cost Explorer in the Google Cloud Console. Set budget alerts to notify you when spending exceeds thresholds. For example, set a $500 monthly cap and receive email alerts when you hit 80% of that limit.
I've personally managed a BigQuery project for a SaaS analytics platform that grew from 10 TB to 100 TB in six months. Without monitoring, storage costs spiked from $200 to $2,000/month. After implementing daily data retention rules and archiving old data to Coldline, we reduced monthly storage costs by 60% while maintaining query performance.
Users frequently mix up storage costs with query processing charges. Storage is billed per GB per month. Query processing is billed per TB of data scanned, not per query. A single query scanning 100 GB of data costs $0.50 (at $5/TB), regardless of how many rows are returned.
Also, BigQuery does not charge for data loading. Ingesting data via streaming, batch loading, or external sources like Cloud Storage incurs no extra fee beyond storage and query costs. This differs from some data warehouses that charge for ingestion.
Finally, storage costs are not affected by the number of users or concurrent sessions. Whether one analyst or 100 employees query the same dataset, the storage fee remains unchanged. Only query processing fees scale with usage.